Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Dennis on December 05, 2005, 07:46:06 PM
-
Hi all,
My name is Dennis and this is my first post to this forum, although I've been reading it for quite some time. :-)
I am a long time Amiga user and I've owned an A500 and an A1200 with BPPC/Bvision. My current Amiga is an A3000 with Xsurf, Cybervision/3D and a recently diseased Cyberstorm MKIII. :-(
I am posting now because this forum inspired me to begin the biggest hobby project I have ever done.
Some time ago (january this year) there was a disccussion on this forum about implementing the Amiga custom chips in an FPGA. The more I though about this idea, the more it fascinated me.
So, I bought a Spartan-3 FGPA development board, learned Verilog (after finding out that VHDL was not my cup of tea) and started working on Minimig. Minimig stands for (very originally :-) ) mini Amiga. My aim with Minimig is to built a complete OCS A500 (with some extra grunt and features like 4Mbyte ram and fast 68000 processor) on a circuit board about the size of a floppy drive. Loading of programs will be done by means of a MMC flash card, which holds the .ADF images of the floppies like a sort of hardware UAE!
I have been working on this for almost a year now and so far I have the OCS Agnus, Paula, OCS Denise and both the CIA's running in the FPGA. The only thing that is missing is sound and keyboard support. All other parts of the chipset are functioning.
It currently runs Worbench and some games like New Year Lemmings quite well although there are still some bugs in the copper and disk controller that prevent other programs to function correctly. I am building this thing mainly for my own challenge, but I was wondering if there would be interest in the community for this kind of machine?
I have uploaded some photo's and screenshots to the assorted hardware forum but they don't show up yet. As soon as they do, I will post some links to them.
Cheers,
Dennis
-
Cool!!!! I definitely will keep an eye on it.
Where are the photos??
-
sounds extreemly cool. i am very interested. i tried stuff with fpgas before but didnt get very far, i suck at hardware stuff. what fpga are you using?
-
FPGAs are cool.
Miggy's on FPGAs are cool too.
This sounds veerrrry interesting.
Are you PAL or NTSC (or both, or either)?
And yes, I find the idea of such a machine quite interesting.
And since it's gonna get asked sooner or later ...
Would AGA be addable, or would that be a complete rework?
-
I use a Xilinx Spartan-3 400K gate FPGA. All in all the custom chips (even OCS) are quite complex and at the moment (without the sound) occupy about 60% of the FPGA. The system I am implementing is PAL, but it could easily be converted to NTSC.
AGA: well, to be honest I haven't really thought about that yet. My first goal is to get an OCS system going 99.9% correctly.
Dennis
-
YES! I am interested!!!
However for me to actually want to use it,
it would have to:
1) Actually work
2) Have all AGA features
Yes I agree you should get ECS working first. That is a good plan. Its just that AGA came out 13 years ago so that is as retro as I am seriously interested in. :-)
p.s. Getting the sound to work should be ez.
I can explain to you the Amiga sound system in plain English if you like?
-
Dennis wrote:
I was wondering if there would be interest in the community for this kind of machine?
Is the Pope Catholic?
Do bears sh*t in the woods?
Of course there will be interest! :-D Put me down as another interested person! :-)
- Ali
-
@ Chaoslord,
Well, it works! It's a shame the pictures don't show up in the album so I can show it. (How long do they take to show up after uploading?)
As for the sound, I also think that would be relatively easy. I own the 3 Amiga hardware reference manuals which tell 90% of the stuff. The other 10% I will have to find out myself. (Just like with the Blitter, that thing took me 2 months to figure out completely :crazy: )
-
Sounds like a fun project to tinker around with. Maybe it could be useful for those who want an Amiga on a PCI card.
Have you thought about possibly having a PCI bus for expansion cards or host computer interfacing? And with a video slot it might be possible to connect a Video Toaster but run the software in an AmigaOne's PPC chip, or have Scala work nicely... I look forward to seeing where this may go.
-
Dennis wrote:
Well, it works! (...)
Man, here, when such a thing happens, we DRINK CHAMPAGNE !
:pint: :pint: :pint: :pint:
-
Nice, at least some "new" amiga hardware.
-
Welcome!
Sounds like a great project. Make sure you keep us all updated.
-
Dennis wrote:
Loading of programs will be done by means of a MMC flash card, which holds the .ADF images of the floppies like a sort of hardware UAE!
That would be great! This is what people are currently missing from the C64DTV.
Flash memory is in fact a great idea - reminds me a bit of the MMC64 (this is how my C64 gets access to a 1GB MMC :-) )
-
WOW, I hope you realise how pleased the Amiga community will be to hear about your project. Please keep us up to date with your progress!
If this could be redesigned as a PCI card and drivers written for OS4 on the AmigaOne, then I would just soil myself on the spot :lol: Even so, the chance to own a brand new classic Amiga system is very very exciting!
How are you planning to maufacture and distribute this? (assuming it is more than just a personal project. I can assure people will pay good money for your product).
Great work Dennis!
--
moto
--EDIT
I have just done some quick research to find out what FPGAs are, and from what I understand you actually program the arrangement of the logic gates in the chip. So it's not like a PIC, where you could write code to emulate the Amiga's chips, and compile it to run on a PIC processor chip. You are actually recreating the logic arrangement of the custom chips, right? How is this possible?!?! I thought the plans for the chips were long gone?
-
Finally!
It's almost a year that I have been planning to do the same but I am not very good at hardware development so I have tried to convince other people to do this kind of stuff, including Amiga Inc., the guys who make the GP2X and others but no answers...
So my idea was very similar: a small portable Amiga compatible device able to either just run all of the games available (there are some very great games for miggy out there so keep it as a retro console) or a bit more sophisticated like a PSP (maybe running workbench).
I would keep it simple and compact (no PCI expansions) nor being itself a PCI card but a battery powered (with also a charger) miggy that you can carry with you.
Qualities:
Small
100% compatible with Amiga software (in ADF format or other)
Running just games or able to boot up Workbench and then run games/software(different flavours i.e. 1.3, 2.0, 3.0)
AGA compatible
6 or 8 switch buttons
Possibility of attaching a keyboard
USB port maybe?
LCD screen integrated
Dream machine:
TV OUT
Integrated mouse pad
Integrated mini keyboard
Recognized on another computer as another drive to move things across (software)
Integrated 56k modem
Since I am a designer, check some of my Amiga stuff at www.loriano.pwp.blueyonder.co.uk
I can help you out with designing the case, let me know.
Great work :-D
Don't give up :-)
-
Sounds good! This will be a true amiga since a long time!
How is this possible?!?! I thought the plans for the chips were long gone?
If UAE is possible, this is possible. He just reversed engineered it for behavioural point of view. It takes a genious to do it correctly, but it seems the community has just been enriched by such a person! Great work!
-
Doppie1200 wrote:
Sounds good! This will be a true amiga since a long time!
How is this possible?!?! I thought the plans for the chips were long gone?
If UAE is possible, this is possible. He just reversed engineered it for behavioural point of view.
So he's guessing? ;-)
Here's to Dennis!
:pint: :pint: :pint: :pint: :pint:
And here's one for luck, hopefully it will be released to the public! I don't care what it costs, I WILL buy one.
:pint:
--
moto
-
Dennis: I think you project has quite good commercial value too.. I'm sure quite many would like to have classix amy if price (+ game copatibility) is OK..
-
Trust me, price will not be an issue. Come on, I've just spent over £500 building an A1200 system!!! I'm sure the cost of manufacturing and distributing Dennis' Minimig wouldn't even come close to that, allowing a healthy profit margin and still sell at a price Amiga enthusiasts would welcome!
--
moto
-
I am quite sure Dennis Ellsworth will be contacted by a toy company soon. :roll:
Better get in touch with somebody like Jens Schönfeld to get it done properly.
-
motorollin wrote:
So he's guessing? ;-)
I think in a way that is true; you make an assumption on how you think that could have been implemented and then you prove your assumption. At least that is the way how I got things done not nearly as complex as this.
So now I'm guessing how he did it ;)
But indeed :pint: bottoms up! Lets all get drunk!
-
That is absolutely amazing. How did he know where to start :-o
--
moto
-
This sounds fantastic!
A year ago you never touched an FPGA and now you have a working system running workbench.... and you APOLOGISE for not yet having sound! BLAAAAAAAAAAAAAAAAAHHHHHHH! :idea:
Do you realise Mick Tinker was trying to get the BoXeR working for YEARS and the holdup was the custom chip implementation in FPGAs.
Two questions - Well about 200 but we don't have all day :lol:
Can you licence 68020 core for FPGA easily and cheaply or are you running with a separate 68K chip?
Apart from the dev board can you give a brief rundown in (how the hell!) you learned verilog and reverse engineering so quick. I'd be grateful.
Good luck with the project!
bty - to reiterate others here - people would rip your arm off for this stuff on PCI for OS4 and morphos. :pissed:
-
A lot of the work I did was indeed some form of calculated guessing. Days of reading through the hardware reference manual and trying to find the out the logic behind things. Also, because Winuae and Winfellow already exist, I have the luxury of being able to look how they did it when I'm stuck. (which I did on some occasions)
However, Minimig is far from finished yet, a lot of custom track loaders choke on the disk controller and the copper is not perfect. But I have gotten this far and I'm determined to finish this! :-)
-
Go for it don't stop now! I and many others would pay good money for this product.
Thanks again Dennis!
--
moto
-
Hey Dennis
I think your first post must count as one of the most, if not the most, awesome and impressive first postings I have ever seen in any Amiga forum! :-P
-
ME WANT ME WANT!!!!
out of interest, how much would something like this cost?
-
This sounds fantastic, and so many possibilities (the current crop of "retro computer in a joystick", custom chips on a PCI card, a complete new Amiga etc etc)
The only problem I can see with making it commercial (not that you were suggesting this, but it certainly could do well), is licensing the Kickstart ROM image(s). Amiga Inc haven't been too responsive or helpful of late, as far as I understand.
Good luck with the project, very interested to see the progress. Perhaps you can take it along to a show or even a user group meeting near you, if there are any?
Chris
-
Dennis wrote:
Hi all,
My name is Dennis and this is my first post to this forum, although I've been reading it for quite some time. :-)
I am a long time Amiga user and I've owned an A500 and an A1200 with BPPC/Bvision. My current Amiga is an A3000 with Xsurf, Cybervision/3D and a recently diseased Cyberstorm MKIII. :-(
I am posting now because this forum inspired me to begin the biggest hobby project I have ever done.
Some time ago (january this year) there was a disccussion on this forum about implementing the Amiga custom chips in an FPGA. The more I though about this idea, the more it fascinated me.
So, I bought a Spartan-3 FGPA development board, learned Verilog (after finding out that VHDL was not my cup of tea) and started working on Minimig. Minimig stands for (very originally :-) ) mini Amiga. My aim with Minimig is to built a complete OCS A500 (with some extra grunt and features like 4Mbyte ram and fast 68000 processor) on a circuit board about the size of a floppy drive. Loading of programs will be done by means of a MMC flash card, which holds the .ADF images of the floppies like a sort of hardware UAE!
I have been working on this for almost a year now and so far I have the OCS Agnus, Paula, OCS Denise and both the CIA's running in the FPGA. The only thing that is missing is sound and keyboard support. All other parts of the chipset are functioning.
It currently runs Worbench and some games like New Year Lemmings quite well although there are still some bugs in the copper and disk controller that prevent other programs to function correctly. I am building this thing mainly for my own challenge, but I was wondering if there would be interest in the community for this kind of machine?
I have uploaded some photo's and screenshots to the assorted hardware forum but they don't show up yet. As soon as they do, I will post some links to them.
Cheers,
Dennis
Hi CU Amiga :roll:
-
I'm interested also. In fact, I registered just to register my interest! Hopefully, this will be the great beginnings of Amiga 2, a new-age homebrew and commercial platform.
-
Soon after reading this thread my morale and faith in the Amiga community went down like the Titanic in the Atlantic Ocean. I simply cannot beleive almost everyone so far beleived this guy's far out tale!
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
I'm writing this message neither to debunk this absurd lie, nor to slug it out with this Dennis bloke. But simply to vent my disappoinment that so many Amiga.org members have fallen for this. No wonder Elbrox can spin tales of fabled Sharks and Dragons, no wonder old unreliable boards change hands for almost a grand! If this is what's left of the Amiga community then we're all well and truly *****
I do understand this post might be offensive, and will definitly show me in a bad light. But for Christ's sake, this guy deserved to be shot down in an apocalyptic flamewar the moment he posted his absurd claim! Instead I see people rejoicing and heralding the arrival of the Amiga 2. How can you be so NAIVE!!
-
Jethro_Tull wrote:
Soon after reading this thread my morale and faith in the Amiga community went down like the Titanic in the Atlantic Ocean. I simply cannot beleive almost everyone so far beleived this guy's far out tale!
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
I'm writing this message neither to debunk this absurd lie, nor to slug it out with this Dennis bloke. But simply to vent my disappoinment that so many Amiga.org members have fallen for this. No wonder Elbrox can spin tales of fabled Sharks and Dragons, no wonder old unreliable boards change hands for almost a grand! If this is what's left of the Amiga community then we're all well and truly *****
I do understand this post might be offensive, and will definitly do me no favours. But christ's sake, this guy deserved to be shot down in an apocaplyptic flamewar the moment he posted his absurd claim!
Hey! Not everyone.
Give some credit where it's due :-P
/Edit
Ah you wrote almost...
Anyway I am prepared to look a bit silly if I am wrong.
-
Jethro_Tull wrote:
I simply cannot beleive almost everyone so far beleived this guy's far out tale!
Well I think Gadgetmaster believes this, and my money would also go in the same direction.
-
Colin_Camper wrote:
Well I think Gadgetmaster believes this, and my money would also go in the same direction.
Don't you mean I believe otherwise ;-)
[Edit]
Just to clarify. I'm a non believer
Hey Dennis ! Prove me wrong. I'm a skeptic. :-P
-
@Colin-Camper
A fool and his money are easily parted. Read what 'GadgetMaster' just contributed to this thread!
-
>>Anyway I am prepared to look a bit silly if I am wrong.
My humble pie is always ready for the eating too! However, given Dennis's post which makes frivolous work of such an undertaking, I think my pie will rest safely in the fridge!
-
Jethro_Tull wrote:
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
Well, it might just be that that core is not optimized.
Jeri Elsworths C64 core fits inside an Altera 100k gate chip. A C64 is a lot more complex than the VIC20.
-
@Ogun
>>Well, it might just be that that core is not optimized.
Sure! There might also be the remote and slightly bizarre possibility that this Dennis guy is telling a truckload of lies.......
Also, the Altera core you mentioned contains a highly optimised processor core emulating a C64, not a circuit replica, which is more akin to running UAE then having the real thing.
Anyways, lets' not give undeserved attention to Dennis, the hallucinating bloke. When a trustworthy member of this community sees the FPGA Amiga in action, then I'll beleive this tale, otherwise, let's not fuel this attention seekers' fire any longer, shall we?
-
AWESOME! :-D
I hope it's not a joke!
I will certainly buy one, and for sure, price is not an issue (i have spent billions of dollars in Amiga hardware so far :-p).
But, is it supposed to connect to a TV screen or RGB monitor or does it have a scandoubler integrated?
-
I have to agree. I don't see it as being possible, at least not on one FPGA. Most of the chipset designs are lost also and this is a claim to also have circut compatibility?
I'm an x engineer and this type of stuff is far from easy when you do have chip schematics to pull from. It's not impossible, Jeri E did it with C64 however the chipset/circuts weren't nearly as complex and schematics were availble. She looked up the old chips in engineering books.
I'd wonder why ppl insist on stiring up a hornets nest.
Good luck to those who want to try, I'm waiting/wanting new hardware or replacements.
-
I agree that it is a hoax.
-
Jethro_Tull wrote:
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
I really know nothing about FPGA's, but a different page (http://home.freeuk.com/fpgaarcade/atari_amiga.htm) on that same site you linked to has a project to recreate the Amiga and Atari ST on FPGA's and he says:-
The Atari chips are very simple, the Amiga AGA chipset is more complex, but will still easily fit in a small FPGA. The only problem is the amount of time needed to get it right ...
So someone else also seems to think it is possible also (AGA chipset on one FPGA).
-
Hi Lando, I never said implementing the Amiga on FPGAs was not possible. Indeed the BoXer motherboard was just that! However , the page I pointed to was merely to expose Dennis' lies. If you've read MY link, you would have seen that even a relativly simple Vic-20 needed more gates then Danny boy's MiniMig implementation!
The page author himself, who undoubtedly has far more proven experience then our Dennis here, hasn't got very far in his attempts. In the same page, he states that barely got through connecting a 68000 to a Gate Array! And just completed the code for a simple Yamaha sound chip, something much simpler and more documented then the Amiga's Paola.
My originial point was simply an appeal to be realistic, and not to be so gullable!! If you're happy to beleive to every charlatan and liar that comes here then suite yourself! :lol:
-
Agreed, it is a hoax, though who wouldn't want to be proven wrong? If it took the Lorraine team several years to design and work the bugs out, what chance does a newbie lone wolf in Holland have in 12 months? Not much.....
-
@Argus, I would be more then happy to be proven wrong! Indeed, I'd buy one myself! But, as you've rightly concluded, the wording and absurd technical details in Dennis' posts clearly point to a lie.
I've noticed that Troika are online right now, maybe they'll negotiate a deal with Dennis for a PPC Amiga with classic chipset!! Vapour, Vapour everywhere, this place looks more and more like a sauna. ACK, Troika, Dragon and now MiniMig! Bring it on guys!!! :lol:
-
@Jethro
Yeah, lots o' smoke goin' round these parts. I do think that Adam Kowalczyk's (spell?) seems a bit more realistic and he has at least 'some' track record with the AmigaOne design. As far as his delivery date has slipped, that's understandable as it's tied to the OS4 release; unless he has designed some kind of wait-state for the trapdoor accelerator to allow it to boot using the 1200's native 68ec020. I wish he would reveal his hardware partner (Far East?) however, as it's again hard to believe one however ambitious lone designer is going to deliver a new accelerator to the expectant masses. I do wish him luck though.
-
Right on target Argus! 'Expectant' being the operative word here! However, with a circus act which includes a couple struggling 'lone wolves', a tiny reseller who can hardly get his act togather and a loonatic with dubious business practices, we're well and truly screwed over here.
'Expectant' we will remain me thinks......
-
Edited by Admin (Argo): Personal Attack
-
Dennis, it sounds TGtBT.
I have no doubt it is/will be possible to put the A500's -custom- logic on a FPGA but until I see proof, I'm tending to guess you're taking the piss.
Even if someone managed to get a mini A500 together, you can be sure AInc (or whoever owns the IP now) would fumble its marketing and make sure it was never a viable commercial venture. :-P
-
Your images are approved and in the gallery for viewing.
http://www.amiga.org/gallery/index.php?c=16
-
According to Mr McEwen and Mr Moss in 2000, it is too difficult and too expensive to reverse engineer and create new version of AGA chipset, because the chipset design was lost somewhere in Bermuda (triangle?).
Mr Tinker failed attempt to create FPGA based AGA emulator:
http://www.amigahistory.co.uk/boxer2.html
Failed Smaky 68K PCI card from Epsitec (Swiss):
http://web.archive.org/web/20000608021841/lamicounter.epfl.ch/Smaky/sm400.html
If a team manage to create AGA emulator on FPGA, perhaps it's better to use Xilinx FPGA Virtex-II Pro or Virtex 4 FX with IBM PowerPC 405 core.
There is a news that AmigaOS 4 will run on top of IBM PowerPC 405.
Possible legal problem: license from Amiga Inc.
If Mr Dennis is an expert on FPGA and Amiga hardware architecture and design, perhaps he can circumvent the license/IP problem by creating Amiga accelerator boards based on the Xilinx Virtex-II Pro.
IMO it is possible to create a single accelerator boards for multiple Amiga models using FPGA programming: ie one program for A1200, second program for A4000, third for A2000, fourth for stand alone mode etc.
Perhaps Netventures (the owner of Amiga Inc) from Netherlands will contact Mr Dennis, if the project is real and the design is 100% compatible with Classic Amiga OS/applications.
> Creating 68K PCI card
What about using ColdFire CPU, Gbit Ethernet, Firewire, USB-2.0, PCIe with external PCIe connectors?
-
Well looking at the images this is either an incredibly elaborate hoax or the real deal.
Given that he's using a real 68K 350K doesn't sound unreasonable at all for an OCS chipset reimplementation. There's a 68000 core on opencores that fits comfortably in a 200K gate device and as advanced for it's time as it was, the OCS chipset isn't that complicated. Comparing the dev time for the original system and this clone is ridiculous. Debugging a design in an FPGA is a heck of a lot faster than plugging ICs in to breadboards and the relatively hich clock rates and gate counts of FPGAs open up options that weren't available to the original designers.
-
Jethro_Tull wrote:
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
The Vic20 implementation referenced include the 6502 CPU. I didn't read Dennis's post to imply that he was implementing the 68000 CPU on the FPGA (which would be an overly ambitious project on its own)
Apparently "Seeing pictures of the early Amiga, it's almost impossible to imagine that the piles of wires and boards could eventually be reduced to something the size of an A500. The first Agnus was three lots of eight bread boards, each with 250 chips, and this was repeated for the other two custom chips which were nicknamed Daphne and Portia in those days and metamorphosed into Denise and Paula."
Now, that's 6k chips for Agnus. 6k for each of the other 2, so 18k chips. As to how many gates per chip, ... that'd be hard to say but this was but it's within the realms of possibility.
-
Your images are approved and in the gallery for viewing.
Edited link (http://www.amiga.org/gallery/index.php?c=16)
There, now don't all you naysayers feel silly?
I have to say, that was a pretty callous welcome a lot of you gave to Dennis.
You want to adopt a wait and see attitude? That's fine. I am. But to outright attack Dennis' credibility and intelligence, that's just cruel. Come on, guys, we're better than that!
-
Yeah I have to admit I am a little disappointed in the pessimistic approach most Amiga users take. I *can* understand it, but perhaps its simply because I am an optimistic person. :|
-
@Dennis:
Good job !
Avoid reading those comments like "you cannot, it's a hoax" etc.
If you want to achieve something, ignoring silly comments is the way to go.
I also got a couple of "you cannot" during my life, and I got by far, farer than most of the ones who told me I was unable.
From what I see in the fotos, it looks quite interesting.
I was wondering if you were using a separated 68k or not, but now that is solved :-)
Now my queston is... do you have a Kickstart in the MC card, or that dev board has also flash, and you programmed it ?
Do you have a driver to access the MC card ? (can you GPL it ;-)), that can be quite handy !
Which version of the dev tools are you using ? I haven't used the windows version of WebISE, but I hope is not as bad as the linux version !, is nice that they license it for free for linux, when Altera doesn't, but... has to be converted from the windows, in such a bad manner ?, I'm sure can be done better.
May be you can GPL your project (or BSD it or MIT it, or whatever open source licese you like) so every skeptic can have a look for themself, and some good comments and ideas thrown at it...
-
With regards to the Kickstart ROM, I don't see why that's an issue. Kickstart ROM chips are readily available to buy legally. So he could just put a socket on the board and people could supply their own, either by buying one or pulling from a dead Amiga.
--
moto
-
Wow! What a project. This looks very promising. Keep up the good work and of course the updates.
-
:-o
-
Matt_H wrote:
I have to say, that was a pretty callous welcome a lot of you gave to Dennis.
You want to adopt a wait and see attitude? That's fine. I am. But to outright attack Dennis' credibility and intelligence, that's just cruel. Come on, guys, we're better than that!
If he is a regular reader of this forum as he has stated then he will be familiar with this type of talk. Anyway, if you go back and read my posts I don't think I was rude. Just overly skeptic. I actually thought it was CU Amiga in one of his posting fantasies again.
I always err on the side of caution. Asking for proof is not sacrilege. I already said I am prepared to look silly if I am wrong but after all the crap our community has had to endure over the years I am sure even a genius called Dennis would understand why we take things like this with a pinch of salt.
If he is genuine then he has a clear winner on his hands and a few skeptics like me wont deter him.
There are more important things for him to be worried about anyway. Like licesensing sharks.
-
Dennis,
Incredible work, congratulations!
-
@Jethro Tull
If you really want to know what is wrong with the Amiga "community" then place yourself in front of a mirror and look directly at it.
It's fine to be skeptical, I am too to a certain degree, but you've majorly crossed the line here. Not only have to rabidly attacked this guy with no back-up, you've also mananged to insult everyone else in the process.
Nice one!
I'm incredibly dissapointed that one of the members here could be so amazingly rude.
The guy wasn't trying to sell anything, or become the saviour of Amiga, he just wanted to share his project with us. Even if it comes to nothing in the end, who cares, it's interesting to hear about. And if people want to give him the benefit of the doubt, then that is their right. (and yours to disagree with that, but not to insult)
It's on-line attitudes like this that have ruined everything as much (if not more so) as vapour over the last few years. If you want to call bullcrap then do so, but do it with some manners, some respect and some first-hand evidence. If others don't agree, then that is thier loss.
Going for the throats of those who take a different view, or don't understand as well as you do has never acheived anything positive.
@ everyone else
(Sorry for the OT rant)
-
Wow.........................................
Wow........................................
......and holy s*!t!!!!
That is awesome.
Are you planning on GPL'ing the cores?
If you are, are there any other FGPA people here to help with the development effort (ie many hands make light work)?
OT: If you could make a core of the SID chip in the C64, you could sell a few of those too :)
BTW why are there so many IP disputes over a technology that realistically left the main stream 10 years ago? What capital is left in it for the IP owners? Mobile phone games run on ARM cores or similar, not 68k using Amiga kickstarts. I can't think of other commercial apps for the kickstart codebase in the modern world. Embedded apps don't use paulas argus and denise. And 68k don't run at 2.8 GHz
If the owners of the kickstart aren't going to use it for anything, they should release it into the wild.
-
"PCI card and drivers written for OS4 on the AmigaOne"
I would like it! A Great idea having it on the A1.
PCi Card... hmm Make it Dennis if You can ;)
Here is a buyer with a new Visa-card and Paypal.
(Where is my TA?)
-
Hmmm
-
Jethro_Tull wrote:
How can you be so NAIVE!!
1. I like this idea!
2. It does not cost me anything to beleive it for a while.
3. I expect nothing from the so called community, so I'm never disappointed.
4. I am used to vapourware so I have more faith in hobbyism than in a commercial enterprise claiming it is creating something amiga.
Let us be. Lighten up!
-
@Dennis
Is the verilog source going to be made available? As an ASIC engineer writing in VHDL & Verilog I'd be happy to help debug a few things. (Searches for his v3 HRM).
Thalion Webshrine
http://thalion.atari.org
-
I like the idea, too. Especially because making custom chips should be easier than the whole thing (?)
Sooner or later we all will run out of them spare A500 used only to get CIA, Agnus, etc.. :-(
it will be great to make all those butchered A500 work again with something like this.
the future is bright
:-)
-
My pictures are online now!
You can find them at Minimig pictures (http://www.amiga.org/gallery/index.php?c=16)
To those who think this is a hoax, let me tell you some details about how minimig works:
The idea is that minimig has the Amiga core (the custom chipset) implemented in an FPGA. The CPU and SRAM is not implemented in the FGPA. To keep things simple I have merged the Amiga chipram bus and fastram bus into a single synchronous bus running at 7.09379MHz. This makes Minimig a functional replica, not a circuit-level replica. This bus only exists inside the FPGA. To connect the RAM and CPU, 2 bridges exist that translate the internal synchronous bus to the asynchronous busses of the SRAM and CPU. Inside the FPGA, all devices (including the cia's) connect to this synchronous bus. Because the bus run's at twice the speed of the original Amiga chipram bus, I can allocate the odd cycles to the chipset and the even cycles to other devices.
The system does not have a ROM, instead kickstart is loaded from the MMC card into the RAM. This is done by a small bootrom that does exist inside the FPGA. I currently have only 1MB of SRAM of which half is devoted to chipram and the other half to kickstart area. If you look at the picture of the complete system, you can see an additional chip that sits between the FPGA board and the MMC card, this is a PIC18LF252 microcontroller that implements the FAT16 filesystem and "emulates" a floppy drive. This is much easier and more comfortable than using a real floppy drive.
Please understand that this is for real. It is not working perfectly yet and a lot of software stalls while loading (diskcontroller is not OK yet) but it does work. As soon as I have some more software running I will be happy to show this at a meeting or show. (I believe there is one every couple of months in Maarssen, the Netherlands)
Dennis
-
Simple suggestion to settle the "it must be a hoax" issue:
@Dennis
You are in the Netherlands, right? There are plenty of other members from there on these boards. Would you be willing to demonstrate your hardware to one of them perhaps?
@Netherlanders...
Is anybody willing to visit the guy if he agrees?
-edit-
Doh! the man himself beat me to it :-D
-
@Dennis
Sorry about the 'hoax' bit, if it really works then I wish all the best to you. As you may recall, for years an awful lot of us waited for Mick Tinker's Boxer Amiga motherboard. His design had the chance back in 1998-2000 to literally 'save' the Amiga and at the same time deliver new features like >2Mb chipram, no need for expensive custom chips via fpga implementation, active pci slots, dimm memory, cheap ppc expansion, etc. Sadly, it never completely materialized and I think that's where a lot of the skepticism stems from.
-
Way to go Dennis! I'll be following the progress.
I think the extra socket for the reuse of your
exsisting KS Rom is a good ideal to avoid licencing problems.
Plaz
-
Hi Dennis!
I believe you and I´m very impressed about your great work. I would like to help you with your projekt to speed up the development, but I can`t contact you because I haven`t got your email. Please send me a small message to my account.
Thank you!
Greetings
AmigaFPGA
-
Your graceful handeling of the negativity indicates that you are one cool dude.
Any plans on submitting your stuff on opencores.org ?
-
Dennis,
how do you interface your PIC-based 'floppy drive' with the rest of your system? I am interested to know basically because I've been working on similar floppy emulation project since I have got my A500 this summer (will post details separately).
-
@Dennis
Impressive stuff! :-o I've always wondered how one could clone the OCS chipset in FPGA and I still don't get it how you did it, but if you got New Year Lemmings running, that's all the proof we need :lol: Keep up the good work en zet 'em op! :-D ;-)
-
Myabe one idea would be to make a video of the unit disconnected, the video you connecting the unit to monitor etc and switching it on?
-
smartrod wrote:
Myabe one idea would be to make a video of the unit disconnected, the video you connecting the unit to monitor etc and switching it on?
That would be good.
-
Hi
Well done Dennis. :pint: :banana:
That is very impressive. I love to have one (if you decide to market them). What about a nano-itx board? Good luck with getting it compleated. I'd say this is probably the most exciting thing to happen to the classic hardware in a long time.
Robert
-
Keep up the good work!!!!! I definitely will buy one of these if they are available.
Putting it into a pci card would be a attractive idea too...
-
@Netherlanders...
Is anybody willing to visit the guy if he agrees?
Maarssen is about 20 minutes away by train, so practically it's no big deal. However, I'd feel quite embarrassed to visit to 'verify' things. For one thing, what good is my word (or that of Speel, Odin, Seer, and Dopie1200) against sceptics, nay-sayers and other cretins who made up their mind the instant they saw the announcement? Much more importantly, Dennis is for all intents and purposes an enthusiastic hobbyist without any intention of 'saving the Amiga', or to make some money out of his project. There is no need to 'verify' things. I wish him lots of success in completing this thing: if I had the skill to work with FPGAs I'd probably offer my services.
@Dennis: heel veel succes met je project gewenst. Houd je er verder alsjeblieft rekening mee dat als je besluit om de boel te commercialiseren je moet oppassen met patenten die op de Copper en Blitter rusten? Sowieso is het hele gebeuren rond het intellectuele eigendom van Amiga, Inc. een waar mijnenveld---eigenlijk weet niemand meer wie nou de baas is, en waarvan. En negeer het stel idioten dat zei dat het een broodje aapverhaal was.
My apologies for the brief intermission in Dutch. Normal English will now be resumed.
-
Also, the Altera core you mentioned contains a highly optimised processor core emulating a C64, not a circuit replica, which is more akin to running UAE then having the real thing.
Incorrect, the Altera core he mentioned reimplements and improves on the C64 design. It's hardware, it's not a software emulation. I am not foolish enough to say whether or not this guy's work is the real deal with the current information that we have but he doesn't claim circuit compatibility.
-
Cymric wrote:
@Netherlanders...
Is anybody willing to visit the guy if he agrees?
Maarssen is about 20 minutes away by train, so practically it's no big deal. However, I'd feel quite embarrassed to visit to 'verify' things. For one thing, what good is my word (or that of Speel, Odin, Seer, and Dopie1200) against sceptics, nay-sayers and other cretins who made up their mind the instant they saw the announcement? Much more importantly, Dennis is for all intents and purposes an enthusiastic hobbyist without any intention of 'saving the Amiga', or to make some money out of his project. There is no need to 'verify' things. I wish him lots of success in completing this thing: if I had the skill to work with FPGAs I'd probably offer my services.
I understand. Of course you don't just have to visit to verify it, you do share a common interest and could have a beer or whatever too :-D
As one of the more level headed and objective members of the site I am sure most people here, myself included would trust your word or Speels or Odins. It just seems quite sad (though understandable given what the community has been through as a whole) that we are generally so jaded that having someone inspect a quick demonstration is required to satisfy people's opinion.
Personally I don't know enough about FPGA development to know how feasible the claim is so I have not commented. I have talked about the idea with people in the past but that's all it was, just musing. I don't recall it ever being considered impossible in theory.
It is a moot point anyway if Dennis plans to demonstrate it at a show. I'm sure he also understands the mixed reaction his claims have generated if he has lurked here as long as he says :-)
-
Karlos wrote:
Personally I don't know enough about FPGA development to know how feasible the claim is so I have not commented. I have talked about the idea with people in the past but that's all it was, just musing. I don't recall it ever being considered impossible in theory.
A quick check, shows that a cheap Spartan3 board should be quite able to carry a basic OCS compatible chip.
-
Dennis wrote:
Hi all,
.....
I have been working on this for almost a year now and so far I have the OCS Agnus, Paula, OCS Denise and both the CIA's running in the FPGA. The only thing that is missing is sound and keyboard support. All other parts of the chipset are functioning.
It currently runs Worbench and some games like New Year Lemmings quite well although there are still some bugs in the copper and disk controller that prevent other programs to function correctly. I am building this thing mainly for my own challenge, but I was wondering if there would be interest in the community for this kind of machine?
Cheers,
Dennis
Excelent News! :-D :-D
I'm realy looking forward to see your progress on this project. Best of luck.
//Niclas
-
Regarding what I suggesed earlier, just in case it wasn't clear, I should point out that it was only *if* the guy wanted possible a way to silence the critics. In his shoes I'd be more than happy to ignore them ;-)
-
All I can say is absolutley amazing. If you need any help with this project I am more than willing. As to what I could help with I dont know but If you were to compile a list of things that needed help and posted it I'm sure the help would arrive.
As for those that completely discount any fresh ideas that come thru, I thought this was an Amiga community!!! So for those that want to throw mud go and play in somebodies back yard. Those that want to keep a communty going all juump on board
-
@Dennis
What are the Three LED's for on the CPU board?
What does 3C D4 Mean?
-Edit- That's some damn fine solder work on these custom boards...
-
According to Mr McEwen and Mr Moss in 2000, it is too difficult and too expensive to reverse engineer and create new version of AGA chipset, because the chipset design was lost somewhere in Bermuda (triangle?).
If you set out to build such a thing for the purpose of making profit without a customer already lined up, then it's probably a ludicrous commercial mistake. But if you're doing it as a hobby and don't care if you make profit or not, and are interested primarily in the educational and/or fun experience of it all, then the given reasons no longer apply.
I once thought Jeri was crazy for her CommodoreOne project, and she had a group of C64 enthusiasts behind her. But look what happened, some toy company had an idea that her project fit nicely into, and she got a pretty decent market out of the thing. (not so coincidentally, most people I know think I'm crazy for still being interested in Amiga and anything I'm involved in here)
It'd have made commercial sense to build it after having such a toy product customer on board, but (IMHO) she got lucky and things worked out pretty well for her. There's some chance that the same could happen for an Amiga in an FPGA to plug Lemmings, Worms, Pinball Fantasies and some other classic Amiga games direct into a TV, but it's a very risky commercial venture to get into such a thing without having a marketing team working on a final product.
As an educational or fun experience, especially if it might look good on your resume in the future, then go for it. Ignore the nay-sayers. because if we listened to these nay-sayers all the time there wouldn't be an Amiga market at all today, as it simply doesn't make commercial sense at all anymore. I myself just spend close to US$3000 on a single piece of equipment that I hope to learn some stuff from. It could be used to help with debug of software for Amiga stuff, but it'd be rediculous to get such a thing hoping for good returns from any software that might be developed using this thing. It's experience and something on my resume I'm looking for from this expense, and for those reasons it might actually be worth it. Same for Jeri or Dennis, even if he doesn't get rich directly from selling any OCS/ECS in an FPGA gizmo. Knowing how to do such a thing can be worth quite a bit.
-
FluffyMcDeath wrote:
Jethro_Tull wrote:
The complete Amiga chipset taking up 60% of a 400K gatearray. A humble Vic-20 takes up way more then that! htp://home.freeuk.com/fpgaarcade/vic20_main.htm
"... Agnus was three lots of eight bread boards, each with 250 chips, ..."
Now, that's 6k chips for Agnus.
Think I was a little too worst case there. I think I'd like to reconsider my chip count since I don't really think that I could get 250 chips on a breadboard even if they were one legged chips. So I think they meant that a set of 8 had 250 chips which makes for a mere 750 chips for Angus rendering the whole chipset would be 2.25k chips.
I revise my estimate from do-able to very do-able.
-
I think I'd like to reconsider my chip count since I don't really think that I could get 250 chips on a breadboard even if they were one legged chips. So I think they meant that a set of 8 had 250 chips which makes for a mere 750 chips for Angus rendering the whole chipset would be 2.25k chips.
Hewlett Packard had, some years ago, some 30x30 cm boards with 256 (I counted them) i.c.s each. They were normal DIL chips, that was a 10 layer board, quite impressive stuff. So yes 250 ics fits a board. If I find it I'll post the part number. (And may be a photo !)
-
What are the Three LED's for on the CPU board?
The leds on the CPU board connect to the status pins of the MC68SEC000 (FC0,FC1,FC2). They aided in early debugging so I could see if the processor was running in supervisor or user mode, doing data operations etc.
What does 3C D4 Mean?
3C and D4 are the current contents of the lower 8 bits of the DDFSTRT and DDFSTOP registers. (some weird stuff goes on there which I didn't understand, so at first I was often fetching one word too many or less)
Dennis
-
Either 4 or 8 gates on each I would guess. Not that many gates in the entire chipset then.
-
When I first read Dennis' post I must admit to being sceptical (though very excited).
However, he had said he had posted pictures, so I decided to wait to see them before commenting.
I wish a few others had taken this approach, rather than insulting virtually everyone that commented on this thread.
Best of luck with your project Dennis. Please please please keep us updated!
Have you considered a small webpage to post progress?
-
I think the possibilities for this are great, even if it remains OCS only (plenty of Amiga users get along without AGA as it is). Plus, Dennis already mentioned that by using SRAM the memory bandwidth was doubled. If the chipset isn't stealing cycles from the CPU and the CPU can access memory every 2 cycles (instead of 1 of 4) then it's already faster than AGA with just a 16-bit wide data bus.
I think it would be nice to include this hardware on a new AT(X) motherboard along with PCI slots and a ZIF socket for 040/060 chips.
Thanks to Dennis for working on this project.
-
This is great... if I recall the pictures of the Boxer prototypes, there were 3 Altera chips on them... and that was about 5 years back now... By now, all of that should fit in one chip. :-)
I recognize that Spartan III proto board in the photos... I came across it when I was shopping for one.
Good luck Denis
-
Maybe Dennis might offer a community buy-out like Blender (www.blender.org) did.
That is, set a figure that Dennis would be happy to sell the Cores for, and ask the community to raise that money, then once raised, the Cores become open sourced.
(There is a website that acts as a clearing house for this transaction, can't remember the name off hand)
That way the community can all pull together and develop a new Opensourced, Amiga hardware platform and Dennis can re-coup some of his efforts.
Just an idea to allow the most access (and hence most likeihood of seeing this project come to fruition for hobbyists) without some acknowledgement of Dennis' huge effort.
-
Dingo_aus wrote:
Maybe Dennis might offer a community buy-out like Blender (www.blender.org) did.
That being the same community which flamed and attacked him a little earlier in this thread?
It is just that kind of attitude which scared off DCE (to name just one) from ever doing anything amiga ever again, let alone release any of their designs.
Dennis made this as a hobby project, he should keep it like that. If there is enough (commercial!) interested he should sell it only to an established company. Individual Computers springs to mind.
-
ajlwalker wrote:
I wish a few others had taken this approach, rather than insulting virtually everyone that commented on this thread.
Well, I commented on the thread and I didn't mind people being sceptical or even calling me naive (or worse!).
Stuff like this is what keeps me in the Amiga community and coming back for more - where else would you find stuff like this going on..except on the rollercoaster we call Amiga! :lol:
I love it! :-D
-
Hewlett Packard had, some years ago, some 30x30 cm boards with 256 (I counted them) i.c.s each. They were normal DIL chips, that was a 10 layer board, quite impressive stuff. So yes 250 ics fits a board. If I find it I'll post the part number.
But that's not a breadboard (http://www.physics.rutgers.edu/gifs/breadboard.gif)is it?
-
FluffyMcDeath wrote:
...even if they were one legged chips.
I don't recall showing you the (ejector seat edition) eprom chip from my apollo 1240 card when you visited... Guess I must have :lol:
-
@Dennis
Awesome work with the FPGA design. I'm extremely interested in seeing you getting this up and running. The piccies look very convicing and the fact you've got any software running at all is amazing!
Your grace in handling the nay-sayers is also to be commended.
I'll be keenly watching this thread for your progress...
Something interesting and only slightly related... I just saw a new product that is being manufactured by DCE in Germany is using the same XLINX Spartan FPGA chips! ( Offsite Link (http://www.tomshardware.com/storage/20051205/a_new_dram_hard_drive_on_the_block-01.html)) Didn't DCE make Amiga hardware in the old days?!
Spooky :-D
-
Remember to keep us updated on this extremely intresting project!
-
If the chipset isn't stealing cycles from the CPU and the CPU can access memory every 2 cycles (instead of 1 of 4) then it's already faster than AGA with just a 16-bit wide data bus.
Well, I think that is the main goal is not to beat AGA chipset, but is only to give a second life for amiga. If it could only be possible than i would prefer to have PPC emulating m68k and AGA for classic stuff and some newer hardware for all new stuff. :roll:
-
That being the same community which flamed and attacked him a little earlier in this thread?
It is just that kind of attitude which scared off DCE (to name just one) from ever doing anything amiga ever again, let alone release any of their designs.
Dennis made this as a hobby project, he should keep it like that. If there is enough (commercial!) interested he should sell it only to an established company. Individual Computers springs to mind.
well spotted Tahoe! (apart the DCE bit that has nothing to do with the users)
@Dennis
your doing a great job (after looking at you pictures, i beleive you even more).
Do what fits better for YOU . . . the so called "Amiga community" would never happy anyway.
I can't speak for all but i highly apreciate your skills, whatever will be the target.
Keep it up, Dennis and my sincere compliments for your project :-)
EDIT- ah . . someone on another site, pointed to this page
http://home.freeuk.com/fpgaarcade/atari_amiga.htm
saying that this is your project as well.
Is it true? thanks.
-
@Dennis
Again, good luck with the project; the photographs are very encouraging
@all
really, I don't understand the outrage at understandable skepticism out there...especially before the photos were posted...(anyone remember iWin?)....makes me feel like amiga.org is the computer equivalent of O'Reilly Factor. 'what! you're not with us...shaddup!" Certainly hardware manufacturers, even hobbyists, have to be thicker-skinned than little schoolchildren scared off by the slightest criticism.
So just relax all, make some tea an have a biscuit or two and dream of fpga's with hidden Amigas inside by next Christmas.
As for DCE, I doubt that all those who never saw their faulty ppc cards again have much sympathy for the plight of that company (who arguably have done more damage than any army of bomb-throwers on amiga.org could ever do).
-
by motorollin on 2005/12/6 3:17:12
With regards to the Kickstart ROM, I don't see why that's an issue. Kickstart ROM chips are readily available to buy legally. So he could just put a socket on the board and people could supply their own, either by buying one or pulling from a dead Amiga.
Or write a kickstart clone. There are a couple of bounties for such a beast over on TeamAROS (http://www.thenostromo.com/teamaros/). For an OS, the cheapest possible way to reproduce this (commercially) would be to use AROS but that would require digging into the AROS-68K problems to resurrect it from inactivity.
Dammy
-
dammy wrote:
by motorollin on 2005/12/6 3:17:12
With regards to the Kickstart ROM, I don't see why that's an issue. Kickstart ROM chips are readily available to buy legally. So he could just put a socket on the board and people could supply their own, either by buying one or pulling from a dead Amiga.
Or write a kickstart clone. There are a couple of bounties for such a beast over on TeamAROS (http://www.thenostromo.com/teamaros/). For an OS, the cheapest possible way to reproduce this (commercially) would be to use AROS but that would require digging into the AROS-68K problems to resurrect it from inactivity.
Dammy
Should be quite a bit easier now since Bernd's AFA...
-
um, no it is mine (mikej) at www.fpgaarcade.com
similar work going on there, but I have been working on the Atari chipset (simply as I have detailed info on them), and as someone noted the Yamaha sound chip. That took a while to model the complex analogue interaction of the chanels, which is used by all the digisound trackers.
Moving to Sweden has also taken a lot of time, but I hope to be back on track soon. I have been measuring a lot of signals on the chips (amiga & atari as you can see on the website) as I want it to be absolutely cycle spot on and work if required with real components.
There has been some comments about the size of Dennis' design, but I can see no reason why it would not easily fit in a SP3 xilinx - the vic20 is bigger because of the 6502 core, and the virtex1 I used for that is significantly smaller than the newer chips.
Dennis looks to be doing a good job and I look forward to seeing the code. Perhaps it can be mated with my 68k core when it is complete.
Cheers,
Mike
support@fpgaarcade.com
-
This could be a great thing for a retro gaming device (akin to joystick C64 etc). Or even in some form of SBC (single board computer).
-
If you look at the picture of the complete system, you can see an additional chip that sits between the FPGA board and the MMC card, this is a PIC18LF252 microcontroller that implements the FAT16 filesystem and "emulates" a floppy drive. This is much easier and more comfortable than using a real floppy drive.
I'm confused now. Earlier you said it uses ADFs. I take it the MMC doesn't work as a huge floppy, so how does Minimig decide which ADF to read? Or does it mount them all as DF0: - DFx:?
-
@mikej
thanks for explanation.
Good luck for your project as well :-)
-
I'm confused now. Earlier you said it uses ADFs. I take it the MMC doesn't work as a huge floppy, so how does Minimig decide which ADF to read? Or does it mount them all as DF0: - DFx:?
The MMC card is FAT16 formatted and holds lots of ADF images. The PIC18LF252 selects one of them and mounts them as DF0:. You control the PIC18LF252 to "mount" an ADF image, much like in tnt23's project.
For now my "user-interface" to the PIC18LF252 is very rude and works via a serial port connection, but in the end I want to make a simple onscreen display that is overlayed on the Minimig's video output. You can then use the joystick/mouse or some simple buttons to "insert" an ADF image.
-
Dennis wrote:
The MMC card is FAT16 formatted and holds lots of ADF images. The PIC18LF252 selects one of them and mounts them as DF0:. You control the PIC18LF252 to "mount" an ADF image, much like in tnt23's project.
Nifty.
-
@Dennis
Thanks for the explanation.
-
Argus wrote:
really, I don't understand the outrage at understandable skepticism out there...especially before the photos were posted...(anyone remember iWin?)....makes me feel like amiga.org is the computer equivalent of O'Reilly Factor. 'what! you're not with us...shaddup!" Certainly hardware manufacturers, even hobbyists, have to be thicker-skinned than little schoolchildren scared off by the slightest criticism.
I don't think it was the skepticism that got people upset as much as it was the inflammatory way in which certain individual(s) expressed their skepticism.
@Dennis
So do you have any plans on what you're going to do when you finish this project? Sell Minimigs? Release the source? Keep it to yourself?
-
This makes me sad to realise how lame computer industry is :( I guess we could see AAAA 32bit graphic chips in 1991, only if management did their job.
:madashell:
Great work Dennis
-
@Dennis
YES!! YES!! YES!! YES!! YES!! YES!! I bow to you oh great one! Interested doesn't even BEGIN to describe it! I've only been wishing for this since I first came to this board. If you build it, I will buy.
@All
You know, I've followed many many news articles about things in the computer and electronics world over the years... I've discovered once simple truth that has been proven EVERY time... The fastest way to get proven DEAD WRONG is to tell a hacker something cannot be done. "This encryption is unbreakable..." Broken. "This OS will only work on our hardware..." Cracked. "You can't read MP3s off of this device, only write them..." Hacked. Etc. Etc. Etc...
EVERYTHING is possible... Some things are just more likely than others.
-
I am building this thing mainly for my own challenge, but I was wondering if there would be interest in the community for this kind of machine?
I believe the interest is huge. =) Keep going!!! ;-)
And keep community in a course, please!
-
Argh this thread is getting huge!
My suggestion of a community based buy out is to serve two purposes:
1) The community gets the open cores to play with, modify develop and improve (hence the end result is likely better because many people have improved it and contributed many more man-hours than possible for one person).
2) Dennis gets something for his work.
Out of the all the alternative uses of Dennis' work, I'd think this is the most "Win-Win" for everyone. If this goes closed source and ends up in a Joystick that you plug into your TV to play Flimbo's quest or Afterburner etc, what can the community really do with that?
(Plus we couldn't see how cool Dennis' work really is!)
-
Hello Dennis,
first, congratulations for your work, this is awesome !
I am also working on an Amiga clone myself but my aim is different : I want the fastest possible chipset (112 MHz).
I have solved the problem with paula this way : every cycle (3.54 MHz), you accumulate the 14-bit value (sample x volume) in a 21-bit accumulator (actually, two values since you have to mix two channels). Every 64 cycles, you output the accumulator value to the DAC (f = 55,420 Hz) and you clear the accumulator.
Good luck
Frederic
-
Hi Frederic,
I followed your thread "amiga dans un fpga" on fr.comp.sys.amiga. I didn't know you were still working on it. How are you going? Last thing you mentioned was that you had the intterupt controller working and was starting with the sound part of Paula. Are you working with an external DAC or are you planning a one-bit delta-sigma modulator? My plan is to use a delta-sigma modulator to keep the hardware simple but as I found out through some reading on the subject and some simulations, getting it to sound good is not that easy :crazy:
Anyway, succes with your project!
-
@Dennis
My experience of sound generation is that 1-bit SD sounds crap unless you are dealing with extremely high speed of conversion (typically at least 32x more than the original sample frequency). Otherwise you can get too much slew in the waveform.
-edit-
Of course of you are talking about multi MHz rate 1-bit SD, forget everything I just said...
-
Hello Dennis,
I am doing fine. I put the project on hold for few months but now, I have more spare time.
first I wanted to use 4 1-bit DAC inside the FPGA, the mixing being done in analog. But, I have found a board with a cyclone II FPGA that integrates everything : SDRAM, 2x10bit video DAC and a 24-bit sound DAC. So, I will go for an I2S interface but I do not know what is the behaviour of the DAC when it does not operate at the recommended frequency (wrong filtering ?).
I think you should be fine if you do the delta-sigma @ 7 Mhz.
Tell us when you have the audio working.
Regards,
Frederic
-
FrenchShark wrote:
I think you should be fine if you do the delta-sigma @ 7 Mhz.
Tell us when you have the audio working.
I have found a pulse rate of around 2Mhz is great for a reproducting a CD quality audio signal... Which I'm sure is more than enough for reproduce the aged Amiga audio... One problem is that the the DACs in Paula are not exactly linear... if you want that "True Amiga" audio, I would suggest not using something as HQ as a 1bit SD DAC...
-
If this project is REAL and succesful, perhaps someone can create some sort of terminator /I/O + Kickstart ROM board for ColdFire/PowerPC accelerators (dragon, CDTV project).
Using this I/O board, the accelerators can be converted into stand alone, fast AmigaOS machine.
http://www.cdtv.org.uk/coldfire/
http://elbox.com/news_04_12_17.html#UPDATE
-
+ Kickstart ROM board
I'd definitely favour the AROS Kickstart Replacement bounties instead. Besides all the other advantages, doing so would benefit more than one project.
Kickstart ROM Replacement (Phase I) (http://www.thenostromo.com/teamaros/bounty_details_23.html)
Kickstart ROM Replacement (Phase II) (http://www.thenostromo.com/teamaros/bounty_details_24.html)
(Once I'll be employed again, I'll donate there as well, of course.)
-
Senex wrote:
(Once I'll be employed again, I'll donate there as well, of course.)
I thought you worked for AHT, does that mean that that endeavour has ended?
I don't mean to pry, tell me to sod off if you wish :-)
Good luck with the job hunting :-)
-
@uncharted
does that mean that that endeavour has ended?
I left the AHT management already right after CeBIT this spring. Mainly because my profession is something completely different than being a businessman. Thus the AHT endeavour has always just been secondary to me besides working on my doctoral thesis in molecular biology. Also, originally I just joined because of OS4, and as you know, these plans failed due to AI.
Good luck with the job hunting
Thank you! :-)
-
Argus:
I don't understand the outrage at understandable skepticism out there...
Oh, you don't? Let me refresh your memories.
Jethro:
This guy deserved to be shot down in an apocalyptic flamewar the moment he posted his absurd claim!
Argus:
Agreed, it is a hoax... lots o' smoke goin' round these parts.
Don't think you're going to get away with it so easily.
-
Congratulations on this cool project! Best of luck to you I will follow this interesting project in the future.
Personally I have been waiting for something like this happening to classic Amiga hardware.
I would be willing to pay for it when it's finished. I have always dreamt of a small, relatively compatible Amiga system capable of 3.0/3.1.
-
Please keep us posted on this project. Updates, news, etc, please post them here.
-
Hi Everyone,
What a great forum and thread congratulations. Thank you Dennis for showing us your Minimig Project.
I only wanted to suggest an idea. I have seen on the net people have 'depackaged' IC's such as those from smart cards and photographed them under a microscope. Could it be possible to do so with Amiga Custom chips and then write some sort of image recognition software to help plot the logic gates, individual components and circuits into a schematic?
http://austech.info/showthread.php?t=72548&page=3&highlight=skin+fox
http://www.cl.cam.ac.uk/~sps32/mcu_lock.html
http://www.cl.cam.ac.uk/users/rja14/tamper.html
Hope the OCS chips don't have this heat sink grid
http://www.xenatera.com/bunnie/proj/anatak/xboxmod.html
Johnno
-
Hasn't anyone ever heard of truth tables?
We know the pinouts of the chips, they're well documented. Build an interface card for each chip... Place all possible input combinations on the inputs, read all results from the outputs, compare it all with known facts about the chips. Program the FPGAs accordingly.
-
@Dr_Righteous
With complex chips like amiga custom chips the actual state of the output might depend on *previous* inputs aswell, not just combination of current inputs. Actually, the actual output state can depend on unlimited number of input states.
Thus, this can't work for complex chips.
-
@Piru
Ahh crap, didn't think about that. Excellent point. Nitric acid it is then... :-o
-
Hi,
Thanks for all the support and suggestions, knowing that there is so much interest really helps to keep me motivated. I will be posting updates to this thread every time I have made some progress and here is the first:
The copper bug is fixed, the bug was caused by a broken wait state after the WAIT instruction. This caused the copper to immediately process the instuction after a wait for the end of line 255 ($FFDF,$FFFE), this WAIT instruction actually ends before the end of line 255 ($df instead of $E2) and depends on some kind of delay to take the copperlist to the beginning of line 256.
I have also found the cause to why a lot of trackloaders choke on my disk controller; the disk controller is too slow, triggering a timeout. I am now working on making it faster.
-
I have no idea what any of the means but I'm glad you fixed it :-)
So... you've fixed the copper bug, you're working on the trackloader, that just leaves sound, right?
--
moto
-
Hi Dennis!
I believe it is absolutely amazing what you are creating! And I admire you because of your cool approach on the project. And that applies also in case you cannot fully complete this project. Very unlikely though!
Thanks for sharing this!
Leo
-
This is a really interesting project, I look forward to seeing its outcome. Sounds like you've put a lot of work into the project, and learnt a load too!
Do you implement a built in scandoubler for a VGA output? That'd be pretty neat!
Is that a 20MHz 68000? I remember that 28MHz Supra accelerator for the A500 ...
Once you've got it all working, what are your plans? I'd be interested in a stand-alone motherboard with the 68000, the FPGA, a few MB of memory and the flash memory card interface for disk images (especially if the PIC controller could do multiple disk images at the same time - imagine Monkey Island 2 running on DF0: thru DF10: ) TV out (SCART?) would be good too. Input via USB keyboard and USB mouse ...
-
I was very sceptical at first (I understood you also FPGA'ed the 68k), but since I've been very impressed by what you've reported here. :-o
Do you have any plans what to do with it once it's completed? I'd love to see an open Amiga FPGA core that might even be improved and extended quite a bit (AGA, maybe even some AAA-like features?). This would be an incredible addition to the AROS project. Count me in if you're looking at the buyout option. :banana:
-
@Dennis
It might not be that your controller is too slow, perhaps your "chipset" is calling it up too fast... Expecting it to respond in normal time when your timing is what, 4x? Just a thought.
-
@Dennis
You say it boots up workbench? Is there any chance you could put up a picture with the output from Showconfig and/or SysInfo?
I'm just really curious ;-)
-
I came in late on this, and I am not in any way enough of a computer expert to understand most of the techie issues being discussed, but I would like to point out that of the three people living in this house, two of them have harrassed me rather a lot to get my old A2000 OS 1.3 up and running so they can play some of the older games that my OS 3.1 machines do not like. Now, I just came back from the mall... ye gods.... and I noticed that lately there is a HUGE interest in retro machines, like a version of the Atari with 40 built in games, old arcade boxes etc. So maybe someone might think about making a minimig of an earlier OS and primarily for the running of all those cool pre-AGA games that made it survive long enough to become a legend. Anyway, that's my 2 cents worth. Talk amongst yourselves.
-
@Will-i-am:
You should really check out WHDLoad (http://www.whdload.de/). It enables you to run all those old cranky games directly from the harddrive, disregarding if your Amiga is accelerated or has a newer os.
/Patrik
-
I'd be interested in a stand-alone motherboard with the 68000, the FPGA, a few MB of memory and the flash memory card interface for disk images (especially if the PIC controller could do multiple disk images at the same time - imagine Monkey Island 2 running on DF0: thru DF10: ) TV out (SCART?) would be good too. Input via USB keyboard and USB mouse ...
That is what Minimig essentially is going to be. For the video I will be using a SVGA style connector that can output video switchable between 15KHz and 31KHz. (just like an A3000) Using an adapter cable in 15KHz mode, Minimig can then be connected to the SCART input of a regular TV. Mouse and keyboard will be connected through PS/2.
It might not be that your controller is too slow, perhaps your "chipset" is calling it up too fast... Expecting it to respond in normal time when your timing is what, 4x? Just a thought.
I have now completed the "speedup" of the disk controller and a lot more games now load and run correctly (North & South, PP-Hammer, Turrican) but other games still fail on a timeout (Parasol Stars, Rick Dangerous). I am now pointing my attention to the CIA's which may indeed run too fast in some modes....
You say it boots up workbench? Is there any chance you could put up a picture with the output from Showconfig and/or SysInfo?
I will try to do that (I hope it will run :-) ) as I am a little bit curious myself too.
Dennis
-
The CIAs indeed NEED to be at the specified frequency since a large number of programs use that for timing purposes. In fact, that was a big problem to the Boxer folks when they were working on their design. You might make the same change that I suggested to them (which they did use by the way): make accesses to the "official" CIA addresses the slow compatible speed, but make accesses to the CIA mirror addresses full speed. That allows old programs which use the CIAs for timing to work correctly, while allowing new programs to access the CIAs without the slowdown. It's too bad the Boxer never made it to release - I was down for one of the first. There were a variety of really nice upgrades to the Boxer compared to regular Amigas.
-
The CIAs indeed NEED to be at the specified frequency since a large number of programs use that for timing purposes. In fact, that was a big problem to the Boxer folks when they were working on their design. You might make the same change that I suggested to them (which they did use by the way): make accesses to the "official" CIA addresses the slow compatible speed, but make accesses to the CIA mirror addresses full speed. That allows old programs which use the CIAs for timing to work correctly, while allowing new programs to access the CIAs without the slowdown. It's too bad the Boxer never made it to release - I was down for one of the first. There were a variety of really nice upgrades to the Boxer compared to regular Amigas.
My CIA's connect to the bus at full speed but they do have an seperate 'E' clock running at 709KHz. This way, timers A and B should run at correct speed. Do you suggest that I need to insert some waitstates when accessing the CIA's ? The games and apps that do run seem to run at correct speed.
BTW, I have put up a screenshot of Minimig running Sysinfo.
It can be found here (http://www.amiga.org/gallery/index.php?n=1236=13)
Dennis
-
@Dennis
Reading some CIA reg in a loop is used for small delays (yes it's ugly as hell, and surely against all portable coding practices, but it IS used nevertheless). The accesses must be equally slow as on the real thing, or things will break.
-
Piru wrote:
@Dennis
Reading some CIA reg in a loop is used for small delays (yes it's ugly as hell, and surely against all portable coding practices, but it IS used nevertheless). The accesses must be equally slow as on the real thing, or things will break.
Has anyone noticed any differences in CIA write timings between Amigas?
I did some tests last week that showed my A1200 took approx 2300ns compared to ~1300ns on my A4000..pretty sure I read something similar the other day but as usual I can't find the link to it now!!
-
@Dennis
Dude, that Sysinfo shot is some impressive sh*t! Runs well enough to fool Sysinfo, that's pretty amazing!
-
Hello!!!
I'm interested in your porject, I like it very much.
I would like know where did you find a specifications for hardware??
I would like make something similar as student project with my promotor but he asked me about hard spec but I told him that I have only Aga spec. Could you help me bring some materials about A1200?
-
It all sounds quite exciting. I would definately be interested in buying either mini-Amiga or PCI-Amiga or whatever form it will be in. But it'll be some time before it's finished I guess.
Anyway, something you might find interesting. The copper bug you had reminded me about a thread on EAN (English Amiga Board) about undocumented Amiga Hardware behaviour. Toni Willen (who wrote the AGA support for UAE and worked on the diskdrive/blitter/sprite/cd32 emulation) posted about some weird behaviour of the copper,blitter and some hardware registers.
For example, he mentioned the copper SKIP doesn't really SKIP but prevents a directly following WRITE from writing.
(meaning that when the next instruction is a WAIT or another SKIP, it won't be skipped)
I though you might find it usefull to read:
Undocumented Amiga hardware stuff (http://eab.abime.net/showthread.php?t=19676)
[dutch] p.s.: leuk om te zien dat hier ook veel Nederlanders zitten :-D [/dutch]
-
Dennis wrote:
My CIA's connect to the bus at full speed but they do have an seperate 'E' clock running at 709KHz. This way, timers A and B should run at correct speed. Do you suggest that I need to insert some waitstates when accessing the CIA's ?
Yes, that is exactly what you have to do. Reading or writing always synchronizes the 68000 to the ECLK. That never changed all the way to the last Amiga and software writers knew that. Even I used that fact for timing in a device driver I wrote (for my PSX controller adapter, PSXPort).
As I mentioned in the last post, accesses to the official CIA addresses should synchronize to the ECLK, but accesses to mirror addresses could be run at full speed. That would keep the backward compatibility while allowing new programs to operate more efficiently.
Another thing to think about - once you have this design going well, you might try hooking a ColdFire MPU to it instead of the 68000. Instead of yet another PPC-based Amiga, it would be great to have a 200+ MHz 68K-based Amiga. Just something to think about for later. :idea:
-
Yes, that is exactly what you have to do. Reading or writing always synchronizes the 68000 to the ECLK. That never changed all the way to the last Amiga and software writers knew that. Even I used that fact for timing in a device driver I wrote (for my PSX controller adapter, PSXPort).
I will try that, Thanx! :-)
Dennis
-
@Piru
I know there is a difference between ECS/OCS CIAs and AGA CIAs. The ECS/OCS ones run at 709 KHz and the AGA ones run at 1.418 MHz. I am surprised that between two AGA machines, you have timing difference.
-
Dude, that Sysinfo shot is some impressive sh*t! Runs well enough to fool Sysinfo, that's pretty amazing!
I was thinking the same thing! If Dennis had been around back in the Commodore days, the A5000 wouldn't be the forgotten myth it is today! Wow. Go Dennis, you're giving me thoughts about taking on that Aros K.S. bounty.
Plaz
-
I take my hat off to you - superb achievement. I wish I could do something to help, but alas I don't have the knowledge...
-
If Dennis had been around back in the Commodore days,
... he wouldn't be able to make this board since FPGA's weren't available then.
-
Plaz wrote:
Wow. Go Dennis, you're giving me thoughts about taking on that Aros K.S. bounty.
Take the bounty! Take it!
-
JLF65 wrote:
Another thing to think about - once you have this design going well, you might try hooking a ColdFire MPU to it instead of the 68000. Instead of yet another PPC-based Amiga, it would be great to have a 200+ MHz 68K-based Amiga. Just something to think about for later. :idea:
Wouldn't it be more interesting to replace the Chip to CPU bus with a PCI bus...? Then we could connect this Chip to any CPU that takes our fancy...
-
Interesting project :)
Thumbs up for Dennis !
-
Dennis,
I'd be interested to know:
1) how small the final product might be
2) would it include interfaces for IO like KB, Mouse etc using "standard PC" components
3) will there be any other options other than memory card such as IDE or FDD (again PC standards) and is the type of memory card fixed?
and I guess most people will be wondering "WHEN" and "HOW MUCH" ....
Great work, I got all excited when first reading thread then disheartened by all the "Nay-sayers" and then all excited again. :-o
-
FrenchShark wrote:
@Piru
I know there is a difference between ECS/OCS CIAs and AGA CIAs. The ECS/OCS ones run at 709 KHz and the AGA ones run at 1.418 MHz. I am surprised that between two AGA machines, you have timing difference.
No, there is no difference in the ECLK from OCS to ECS to AGA. Now the TOD register on one CIA hooks to the HSYNC and the TOD on the other CIA hooks to the VSYNC, so if you ran your AGA machine in DBLPAL/DBLNTSC like most folks, you'd see different TOD timing compared to an OCS/ECS system running on PAL/NTSC. The chips themselves still use the same ECLK frequency.
-
FrenchShark wrote:
I am surprised that between two AGA machines, you have timing difference.
From what I remember about the other guy noticing a difference in timing, he said he tested an A500 and an A3000 ( with very similar ~2300/1300ns times compared to my tests), so I guess it comes down to the different Gary chips used..although why they stick different wait states is beyond me, where's Dave Haynie when you need him?
-
@xeron
... he wouldn't be able to make this board since FPGA's weren't available then.
You miss my point. The comment was a complement to Dennis' talent and initiative. "If his talent was available to commodore..." FYI, FPGA's have been around since 1985. Of course they were not as powerful in 1994 as they are today and probably wouldn't have been used in A5000 developement.
But Dennis would have found a solution :-).
Plaz.
-
Rock on, baby! THIS IS INTERESTING! Now we might actually get that ATX-amiga MB...
-
Plaz wrote:
You miss my point. The comment was a complement to Dennis' talent and initiative. "If his talent was available to commodore..."
Its also seems a bit rude to the Commodore engineers. You were saying that Dennis would have saved the day. Do you think people like Dave Haynie, or Jay Miner didn't know how to develop hardware?
I know it wasn't your intention, but thats what your statement implies. Commodore wasn't short of R&D talent, it was short of decent management and marketing.
(I am in no way belittling Dennis' work here, btw, he is doing an excellent job).
-
it's not rude. Commodore was short of both (also technical). So the amiga 500 and the amiga 4000 wasn't so different after 8 years has passed since the original amiga 1000 project (date back in 1984.
Haynie was quite a good guy, but I think you need more than one man to develop a computer...
Commodore was a mass market computer producer. They used to produce low cost computers for playing games mainly. The Amiga project never completely fit in this kind of environment. It's the way a company thinks and acts, and you can change these things. In fact the best people in the R&D department understood this and left...
And Commodore died not only because of the management but also because they were unable technically (which means also implies economic issues) to stay updated when the rest of the world was progressing...
-
chsedge wrote:
And Commodore died not only because of the management but also because they were unable technically (which means also implies economic issues) to stay updated when the rest of the world was progressing...
Commodore had huge sums of money... unfortunatelly the board of directors had to cancel most R&D projects to pay for their toys.
-
Its also seems a bit rude to the Commodore engineers. You were saying that Dennis would have saved the day. Do you think people like Dave Haynie, or Jay Miner didn't know how to develop hardware?
Your are correct that it was not my intention or implication to slap other great folks. Dave is a great guy with awesome talent (and that goes for Bil Herd and others too). I payed Dennis a compliment while using a hypothetical and humorous reference. Of course humor is relative and mine doesn't always apply to everyone.
Plaz
-
Then we could connect this Chip to any CPU that takes our fancy...
Maybe just an alternate type of CPU socket would do. For instance, I could plug in an 040 instead. I like the ideal of Coldfire, but I think there's going to be code compatibility problems with that.
Plaz
-
Please, lets talk about this project and its positive effects and capabilities instead of old stuff about decisions at Commodore that happened 13 years ago.
I watch this thread every day, its nice when some advance is made.. keep up the good work!
-
According to Freescale, 68K IP cores is already obsolete and new design should be based on ColdFire IP cores.
What about combining Minimig design with 68K IP Cores for creating a high speed (GHz) System On Chip (SOC) ASIC?
Is this idea possible with the old 68K IP cores?
What about compatibility with old classic Amiga applications? (Time/clock dependent)
According to Freescale, GDA is one of the commercial IC design companies familiar with 68K cores. Can they design the new SOC?
GDA Technologies Inc. (GDA) provides synthesizable IP cores and value-added design services for the Networking, Embedded, and Consumer Electronics markets. GDA's expertise includes System, Board, FPGA and ASIC design, with significant expertise with many computing architectures. GDA has 180 engineers. In addition to development, GDA can provide turn-key services from concept to production.
Developer: GDA Technologies, Inc.
Platforms: i.MX, 68K, Other
OSes: Linux, OSE, VXWorks, Other
-
Plaz wrote:
Then we could connect this Chip to any CPU that takes our fancy...
Maybe just an alternate type of CPU socket would do. For instance, I could plug in an 040 instead. I like the ideal of Coldfire, but I think there's going to be code compatibility problems with that.
The problem is tying this "custom chip" to a specific processor. By using a standard PCI bus to connect it to the CPU, the Chip becomes more than just a toy, but a useful tool.
-
@ konrad
I'm interested in your porject, I like it very much.
I would like know where did you find a specifications for hardware??
I would like make something similar as student project with my promotor but he asked me about hard spec but I told him that I have only Aga spec. Could you help me bring some materials about A1200?
I own the 3 Amiga reference manuals (hardware, libraries and devices, includes and autodocs) from which I got the most information. These books are hard to get second-hand but they are also included in HTML and Amigaguide format on the AmigaOS 3.5 developer CD V2.1. Maybe you can still buy this CD somewhere. To my knowledge there has never been a hardware reference manual revised for AGA. Besides the reference manuals there is the source code of Winuae and Fellow which you can learn a lot from.
But remember that all this information usually tells only 90% of the stuff, you will have to find out the rest yourself, especially when it comes to AGA.
Good Luck with your project!
-
THX a lot!!!!
I hope I will have Amiga Hardware Refernces Third Edition and start project as soon as it's possible. I will try to find AmigaOS 3.5 developer CD V2.1
Your project is super!!!
-
Searching google for amiga hardware reference yields some results for the actual text of the reference. Obtaining it this way is somewhat of a grey area legally so I won't post any links, but it's not terribly hard to find if you know what to look for.
-
So the amiga 500 and the amiga 4000 wasn't so different after 8 years has passed since the original amiga 1000 project (date back in 1984.
Well, this wasn' thte engineers' fault. They were working on the A3000+ which would have been the first AGA machine. Unfortunately, managment changed and the new leader canceled the good projects started under the former managment that would have made the previous guys look good. Or something. But the engineers did have something much better than the A4000 in the works. The managment change scrapped that and led to hacking together the 4000 which was less computer than the A3000+ would have been. The A3000+ was to have a DSP for soft-modem usage and great audio capabilities, quite a difference from the Paula chip that stuck around in all Amigas Commodore actually sold.
Plus, the engineers working on AAA had left Zorro behind in favor of PCI, and Hombre was going to use HP's PA-RISC CPU technology instead of 68K. Managment cancelled those projects by hammering the company as deeply into the dirt as they possibly could.
-
What about combining Minimig design with 68K IP Cores for creating a high speed (GHz) System On Chip (SOC) ASIC?
Interesting idea, but Dennis can't afford that. ASICs are EXPENSIVE. Having them work at GHz speeds is REALLY EXPENSIVE.
Jeri's CommodoreOne got into an ASIC because someone found a market to sell the things in the hundreds of thousands level of quantity. And it's nowhere near GHz speeds, it might be around 33MHz or so if I remember the Stanford video correctly. (please slap me if I'm wrong)
If you want to make such a very high-speed chip, you need to have a LOT of money to make it happen, many millions of US$ to invest. If you know someone that's superrich and can't figure out how to spend it all, maybe there's a chance of a wild and crazy investor. I'd be happily extremely suprised if the Amiga/FPGA concept makes it into an ASIC of any speed, simply because of the costs involved. But if someone comes along with a good offer and serious plan to make something like the DTV joystick for Amiga games, that'd be extremely cool. But also exremely suprising, just because of the extreme money involved to make ASICs.
-
Well actually Commodore sold 50 or so A3000+ to developers, but the rest of the article is correct. Newtek (actually Alan & Stuart) after Denver Devcon were interested in getting Lightwave rendering using the DSP, which given its time frame would have been a huge increase in rendering speed over the 040's that were just coming out. Unfortunately the DSP subsystem got abandoned and that was that. You should also have talked about the A1000+ which would have been a really sweet little machine, much cooler then the 1200 that eventually took its place :).
-Tig
-
I was re-reading the original post and I had to laugh at Dennis' total understatement of the decade....
Dennis wrote:
but I was wondering if there would be interest in the community for this kind of machine?
Like, if he held this project up in an Amiga meet anywhere he would probably have is fingers, hand and arm bitten off!
-
Any more developments on this subject? Perhaps hs should set up a blog or something (me suggesting a blog -- gasp) so that we can track his progress <3
-
we can keep track right here, in this dedicated thread. blogs are ghey.
-
I'd love to help out on the emulated floppy drive via PIC, although I love AVRs for the fact there is AVR-GCC.
Any chance Dennis you might separate it out and allow others to help?
-
I have now solved the last floppy bug. Apparently, some programs try to find a certain sector by rereading the same track several times. My floppy 'emulator' did not correctly handle this. This is now fixed.
I am now busy cleaning up the code and solving some small issues with the sprite color selection and the blitter nasty mode. Once that is done I can start on the audio part of Paula.
Perhaps hs should set up a blog or something (me suggesting a blog -- gasp) so that we can track his progress
I am not such of a regular poster or blogger, but I will post any progress in this thread.
Any chance Dennis you might separate it out and allow others to help?
Maybe in the future, first everything has to work properly.
-
I think it would be better to post your progress as new threads or news items. This thread is getting very long and difficult to follow now.
Just a thought :-) Congrats again and keep up the good work! If you need any help or even donations I'm sure the good people of a.org will be able to come to the rescue! I would definitely donate if it would help your research.
--
moto
-
Dennis wrote:
I have now solved the last floppy bug. Apparently, some programs try to find a certain sector by rereading the same track several times. My floppy 'emulator' did not correctly handle this. This is now fixed.
One of the original methods of copy protection for floppies was to have multiple sectors with the same sector ID. Depending on when you started reading the disk, you'd get different data from the "same" sector. This was trivial to copy on the Amiga, so it was quickly dropped in favor of more extreme methods of protection.
Psygnosis had some really tough protections all combined together. They'd use long tracks with custom syncs (the long tracks served a dual purpose - to make it hard to copy, and to hold more data on a disk), short tracks, strong and weak bit protection, and specifically located track start positions (with reference to the index point).
-
quote:
Same for Jeri or Dennis, even if he doesn't get rich directly from selling any OCS/ECS in an FPGA gizmo. Knowing how to do such a thing can be worth quite a bit<<<<<
Yes, and contact us Dennis, sure some would be interested in a PCI card for Amy.
-
quote:
----------------------------------------------------------
My CIA's connect to the bus at full speed but they do have an seperate 'E' clock running at 709KHz. This way, timers A and B should run at correct speed. Do you suggest that I need to insert some waitstates when accessing the CIA's ? The games and apps that do run seem to run at correct speed
------------------------------------------------------
ahhh CIA's you have to know the history to deal with the CIA's
I can help. contact Troika at: admin@troikang.com
-
Once again, Dennis - congratulations on your work so far and good luck with the Paula/sound.
I expect you already checked out this processor on Open core - what do you think of it?
K68 core (http://www.opencores.org/projects.cgi/web/k68/overview)
It would be great if you could end up running everything in one FPGA (besides the ROMs).
Also - can you recommend any 'practical' verilog/synthesis books or resources?
Cheers! Oh and Merry Christmas! :-)
-
I expect you already checked out this processor on Open core - what do you think of it?
I have checked it out and considered it for minimig, but then I found out the core is not complete yet. As I was expecting a lot of development problems anyway, I though it was better to take no chances on the cpu.
Also - can you recommend any 'practical' verilog/synthesis books or resources?
I bought the book A Verilog HDL Primer by J. Bhasker (http://www.amazon.com/gp/product/096503917X/qid=1135191614/sr=8-2/ref=sr_8_xs_ap_i1_xgl14/104-2452059-6339100?n=507846&s=books&v=glance)
It is probably not the best book around, but it has teached me the basics. I also bought (from the same author) Verilog HDL Synthesis (http://www.amazon.com/gp/product/0965039153/qid=1135192023/sr=1-1/ref=sr_1_1/104-2452059-6339100?s=books&v=glance&n=283155) which tells how to apply verilog in hardware synthesis. This second book is much better and very practical, but you need some basic verilog knowledge to understand it.
-
Dennis, how feasible would it be to offload some functions of this replica classic Amiga processor to some other modern processor, such as a Via C3 or something else, in order to speed up execution time and throughput? I know there might be issues with data in CPU registers having to be thrown about, etc, but would it work?
-
I though it was better to take no chances on the cpu.
Absolutely. While reverse engineering, its worth using all the "certainties" you can - like the CPU and the ROMs. However when you have the custom chips 'nailed', it could be worth another look.
Thanks for the book suggestions - these books are very pricey and it's hard to judge Amazon reviews - I'd rather take your word for it! :-)
-
I bought the book A Verilog HDL Primer by J. Bhasker
It is probably not the best book around, but it has teached me the basics. I also bought (from the same author) Verilog HDL Synthesis which tells how to apply verilog in hardware synthesis. This second book is much better and very practical, but you need some basic verilog knowledge to understand it.
Verilog HDL by Samir Palnitkar is also a good intro book. In its 2nd edition now I think, I have the first and really like it.
I've got the Verilog HDL Synthesis book as well but haven't had time to get into it yet, hopefully someday. I've mostly been a chip layout guy and have tried learning verilog on my own, hopefully soon I'll get an opportunity to get into it at work.
-
I have posted a little summary story of this project here:
http://www.bytecellar.com/archives/000063.php
FYI.
blakespot
-
Oops, it looks like the CPU is not in the FPGA, when I thought it was. Good work on the article.
-
am not such of a regular poster or blogger, but I will post any progress in this thread.
Maybe someone could help you with your site, where you will post news and progress? This would be much more greater that to post progress here =)
just my thoughts :)
I would like to know, if you are planning for a1200 in fpga?
-
An A1200 would totally rock.
-
acidrain wrote:
am not such of a regular poster or blogger, but I will post any progress in this thread.
Maybe someone could help you with your site, where you will post news and progress? This would be much more greater that to post progress here =)
just my thoughts :)
I would like to know, if you are planning for a1200 in fpga?
One thing to keep in mind for anybody willing to put up a site on this project, the slashdot effect. If it gets thrown up on the web, it will most likely end up on the main drag at slashdot, digg, boingboing, hackaday, etc. This could be expensive.
-
Well Dennis.
Your threads at 200 posts already.
Guess that means you really did say something interesting!
:-)
Keep up the good work and I hope it all comes together for you and there'll be even more excitment in the new year.
-
what I would like to see is miniA1200 without prosessor, just with a1200 compatible accellator port.
Would be nice to see how fpga would speed up AGA.
-
More updates please! ;)
You must have worked on it some more since your last update.
-
More updates please! ;)
I have spent a lot of time trying to find some bugs in the blitter and sprite engine. The blitter is now fixed (untill offcourse, I find another bug) by hacking the blitter nasty mode a bit.... The sprite bug is harder to find, it manifests itself in the score board of Parasol Stars, the background color should be dark blue instead of green. I haven't found the culprit yet and as it is not that important at the moment, I have started on the audio.
That is going along nicely and is already producing very recognizable music. For the output I am now using a simple first-order sigma-delta modulator running at 7.09MHz. To my surprise it doesn't sound bad at all and I think that by adding some dithering and other tricks it will be good enough for the final thing.
Cheers,
Dennis
-
Hello Dennis.
Congratulations and respect!!!!
A friend of mine told me about your work because I am working on a similar project (native amiga).
Unfortunately (for me :-() you are the first one who got the Amiga running :-)
I am using the Altera MAX+ deign-environment and a modified c-one board as development system. I think you are almost done with the audio... It took me days only to understand the audio statemachine. You solved this with an own modulation? - I wanted to use a commercial DAC - but then I may have a problem with the samplerate which is completely flexible on the amiga, but fixed on the DAC.
Do you only want to use an MC68000? - I went back to scratch because I want to attach "any" processor via PCI to my design. I am working quite long on this issue.
Thomas
-
Keep up the really great work on the Amiga chipset in an FPGA !! I just found a link to this thread by accident when I was searching for info on retro Arcade FPGA emulations & retro Atari 2600 FPGA emulations. Since I use Verilog myself and since I am very long time Amiga A1000 owner, I was very impressed by your progress on the Amiga Chipset implementation. As a possible improvement, MMC/SD card can XFER data at 9 to 10Mbits/sec while Compact Flash can transfer at 16mbits/sec via a 16bit CF interface. Use of a Compact Flash controller would allow faster loading of files to the Amiga Spartan3 FPGA. Xilinx has a really neat Compact Flash JTAG configuration loader and Compact Flash controller IC called the SYSTEM ACE that could be applied to your project/Spartan3 FPGA. This would allow you to place your jedec logic files on the compact FLASH for power on reset configuration of the Spartan3 FPGA. If you hook System ACE HOST MPU interface to the Spartan3 FPGA, then you could further use the Xilinx System ACE as a full 16bit Compact Flash controller.
http://www.xilinx.com/bvdocs/publications/ds080.pdf
I am really hoping that your project turns into something really neat like the C one C64 board/project where we can get a Amiga FPGA board to work with. Please keep us updated on your progress.
:-D
-
I am using the Altera MAX+ deign-environment and a modified c-one board as development system. I think you are almost done with the audio... It took me days only to understand the audio statemachine. You solved this with an own modulation? - I wanted to use a commercial DAC - but then I may have a problem with the samplerate which is completely flexible on the amiga, but fixed on the DAC.
Hi Thomas,
It seems like I'm not the only one working on this! Can you tell some more details about your project? Are you working on a "chipset on a PCI card" or are you creating a core for the c-one? And yes, I use an own delta sigma modulator to avoid the same problem you are facing; the conversion of the sample rate. Also, a delta-sigma modulator is very easy to implement in an FGPA. (and cheap too!)
Dennis
PS: There are DAC's who are flexible with their sample-rate requirements, as long as you scale the masterclock accordingly. Also, older DAC's like the TDA1541 or TDA1543 are just plain 16bit converters; they accept any samplerate upto about 176KHz.
-
I agree with jkonstan -- I think CompactFlash would probably be a better idea than MMC, and would also open the door for other devices based on PATA and IDE to be inserted. I think CompactFlash is also more widely used than MMC, and has a really nice form factor (not too big, not too small).
It would also let me use all these CompactFlash cards lying about if it ever goes into production, or whatever ;)
1gb Kingston CF card is £35.95 inc VAT
512mb MMC Kingston Memory Multimedia Card Plus is £20 inc VAT, and MMC cards are in sort supply on many sites (they seem to be dying out)
-
Hi Dennis,
Also, I looked at the pictures that you posted for your Amiga FPGA system. You appear to be using a Xilinx Spartan 3 Eval/starter board. It is a a good Xilinx eval board. I have one of these as well.
http://www.xilinx.com/xlnx/xebiz/designResources/ip_product_details.jsp?key=DO-SPAR3-DK&sGlobalNavPick=&sSecondaryNavPick=
Do you think that can post a pdf to the schematics to your two adapter boards (68K CPU & PIC MMC) that plug into the Xilinx Spartan 3 Eval/starter board expansion connectors so that we can get a closer look at your implemnentation ?
Best Regards,
John
:-)
-
I have finished the basic sound controller and it is now working 100% OK.
:banana: :banana: :banana:
It supports dma and interrupt mode. Attached modes are not included yet, but on the games that i've tested that does not seem to be a problem. (It probably isn't used very often)
I agree with jkonstan -- I think CompactFlash would probably be a better idea than MMC, and would also open the door for other devices based on PATA and IDE to be inserted. I think CompactFlash is also more widely used than MMC, and has a really nice form factor (not too big, not too small).
The MMC card is not directly connected to the FPGA so even if I would replace it by a compactflash card you would not be able to access it as an ATA device. As it is now, the PIC18 microcontroller (to which the MMC flash card is connected) "controls" minimig, it loads the core into the FPGA, loads the kickstart into the rom area and emulates the floppy.
Do you think that can post a pdf to the schematics to your two adapter boards (68K CPU & PIC MMC) that plug into the Xilinx Spartan 3 Eval/starter board expansion connectors so that we can get a closer look at your implemnentation ?
I will try to do that, first I will have to draw a new schematic of the disk controller as i have been modifying the pcb a bit in order to speed up the data transfer.
Basically the diskcontroller is just the PIC18 mcu with the fpga and the mmc card both connected to the spi bus. The clock to the mmc is gated so it can be shut down. This is necessary as the mmc card i've been using still "listens" to the bus even when i de-assert it's CS line. The PIC18 continuously checks the status of the floppy logic inside the fpga, when a track is 'requested' the PIC18 reads the track from the mmc, encodes it to mfm and then writes the track into the fpga. This is done sector by sector. (The PIC18 can hold only a one or two sectors of data into it's internal memory)
-
@Dennis
I have finished the basic sound controller and it is now working 100% OK.
Fantastic! This is great.
Thanks for the regular updates. This work of yours is being watched (and appreciated) by many, many people in the Amiga community. (Maybe not Mick Tinker! :-? )
-
This is really excellent news, indeed.
-
How long till i can buy one? :-D :-D :-D hehe ;-)
-
I have just registered Amiga.org to say something :)
Congrulations Dennis... I hope you also make it AGA compatible...
best regards from Turkey...
-
Dennis, do you pretend to sell your product when done ?
Could you give me (us) a link to your webpage progress status if you're keeping one ?
Sorry, this threat is huge (and huge is the interest of it, too) but I don't have the time to read all of it :)
Tnx for your talent !
-
@Dennis
Nice work. How is the RGB signal generated? Couldn't see any Video DAC on your assy.
@Colin Camper
The K68 core is at it's present state far from beeing usable as a processor running classic Amiga software.
Mainly because it's limited to one operating mode, but there are other issues as well.
With last update back in may'03 i don't think it's much interest for this project either.
@JLF65
IIRC some of the old games did access the CIA's from mirror adresses instead of using the proper ones. I think it was done as some lame way attempting to protect their work against crackers. Same method was also used with the custom chips.
@all
I think we all can agree this is an interesting project undoubtly creating a lot of entusiasm, but keep in mind that ATM, it's just a hobbyist project. Adding AGA, Coldfire, PCI , or other fancy stuff means a LOT of work. Not to mention an attemt to transfer it into a commercial product.
w
-
Nice work. How is the RGB signal generated? Couldn't see any Video DAC on your assy.
My DAC (ahem) consists of 4 resistors for each color R,G and B. You can see them next to the VGA connector.....
Dennis
-
Great work dennis :)
Just opinion please :) how hard would be make to fpga chip aga and other chips with a1200 style accellator port?
How much more powerfull it would be than orginal aga?
-
Well, I guess it works... :-D
Actually, the A1000 Video DAC was made of resistors too. :-)
R-2R ladder made by using precision resistors i think.
w
-
@Dennis
Congratulations, your sound success sounds great! :-)
Can you tell in detail how your MFM encoding is done? I do mine quite straightforward based on C code from Amiga ADF FAQ, but the performance is rather poor. Takes almost 100 seconds to encode all 80 tracks.
-
I think we all can agree this is an interesting project undoubtly creating a lot of entusiasm, but keep in mind that ATM, it's just a hobbyist project. Adding AGA, Coldfire, PCI , or other fancy stuff means a LOT of work. Not to mention an attemt to transfer it into a commercial product.
Yes, true. But I think that the consensus is that AGA is very desirable. I don't know how much more work AGA would be - a fair bit I suppose - but if Dennis came up with AGA and an A1200 trapdoor I/O & Clockport I/O he would have a big hit on his hands. Not least from the people buying PowerVixxens. :-D
-
Dessimat0r wrote:
I agree with jkonstan -- I think CompactFlash would probably be a better idea than MMC, and would also open the door for other devices based on PATA and IDE to be inserted. I think CompactFlash is also more widely used than MMC, and has a really nice form factor (not too big, not too small).
It would also let me use all these CompactFlash cards lying about if it ever goes into production, or whatever ;)
1gb Kingston CF card is £35.95 inc VAT
512mb MMC Kingston Memory Multimedia Card Plus is £20 inc VAT, and MMC cards are in sort supply on many sites (they seem to be dying out)
Don't SDIO cards have a fallback MMC mode?
-
Not sure -- I always thought that SD cards and MMC cards were totally different technology, and only the style of the connector in the card slot determined if it could use SD or MMC, or both.
However, looking on Wikipedia, this does not seem to be the case.
"All SD memory and SDIO cards are required to support the older SPI/MMC mode which supports the slightly slower four-wire serial interface (clock, serial in, serial out, chip select) that is compatible with SPI ports on many microcontrollers."
-
Can you tell in detail how your MFM encoding is done? I do mine quite straightforward based on C code from Amiga ADF FAQ, but the performance is rather poor. Takes almost 100 seconds to encode all 80 tracks.
You don't have to encode the tracks 'exactly'. The decoding routines used by the OS just strip the clock bits so you can get away by just inserting some fixed clock bits on the right places. Petter Schau's Winfellow emulator does this and I use it too. The only problem you may have is that Paula's PLL gets out of lock but even then you could do correct mfm per word and forget the 'inter-word' correction. That should speed up things considerably.
-
@Dennis
Thanks a lot for the tip!
-
Looks like this thread has evolved quite a bit! To my pleasant surprise it also does look like the MiniMig is not a dupe after all! Congratulations to you Dennis, for what is an amazing technical feat! I'm still perplexed how you managed all this in a year, without prior knowledge, but hey, all your posts indicate your IQ as being way above average (to say the least)!
I hope you will take the time and trouble to travel to your next local Amiga club meeting / convention and display your creation to fellow enthusiasts! Maybe something will come out of this which will benefit the whole Amiga community, or maybe not. Either way, it seems like the long lost secret of the Amiga chipset has been finally rediscovered!
Cheers lad! :pint:
-
@Jethro
Some of us have a lot of of techie stuff around our houses, and have too much time on our hands. I myself am doing PCI prototyping projects.
-
What about creating some sort of 68K Emulator/Instruction converter in FPGA and use the PC X86 CPU, not the original 68K or Coldfire?
The Minimig FPGA will be mounted on a PCI Express high speed card (1X => 2.5 Gbps, 16X => 40 Gbps) etc.
-
asian1 wrote:
What about creating some sort of 68K Emulator/Instruction converter in FPGA and use the PC X86 CPU, not the original 68K or Coldfire?
If you are plugging the chip into a CPU as powerful as an x86 having an FPGA to do the 68k emulation would be pointless... though some kind of hardware byte swapper might make sense.
The Minimig FPGA will be mounted on a PCI Express high speed card (1X => 2.5 Gbps, 16X => 40 Gbps) etc.
Now you are talking (Though a normal PCI slot would be more than enough for any Amiga chipset).
-
@bloodline: the coming x86 might very well have byte swapping hardware, since Apple will need that, too.
@Dennis & tnt23: While for Minimig's purpose it might be enough to fake an MFM encoded bitstream - Minimig's Paula is designed not to mind - it would probably raise problems with the original Paula because it violates MFM clock timing.
If I were to design the floppy emulator, I'd put the MFM encoded tracks on the SD card - considering the price of SDs there's not much wrong about it. Encoding the tracks on a fast host system should be fast enough to not be a bother. Maybe you can just provide support for both, ADF and MFM?
-
Zac67 wrote:
@Dennis & tnt23: While for Minimig's purpose it might be enough to fake an MFM encoded bitstream - Minimig's Paula is designed not to mind - it would probably raise problems with the original Paula because it violates MFM clock timing.
It was a problem indeed. Paula chip choked on the illegal MFM stream producing either header or data errors, or, more frequently, both.
I have tried to simplify the MFM encoding routine by leaving out the connectivity part - that is, always coding as if previous bit always were 0 - and that seems to keep Paula happy. 80 tracks are loaded for 24 seconds and encoded for 35 seconds.
If I were to design the floppy emulator, I'd put the MFM encoded tracks on the SD card - considering the price of SDs there's not much wrong about it. Encoding the tracks on a fast host system should be fast enough to not be a bother. Maybe you can just provide support for both, ADF and MFM?
Pre-encoded MFM file would be 1.8M as opposed to original 880K ADF. This will take twice as much time to load. To keep things simple, I'd produce fatter files, like 2.5M, and that'd be three times longer.
-
Zac67 wrote:
@bloodline: the coming x86 might very well have byte swapping hardware, since Apple will need that, too.
I doubt iNTEL will make any special arrangements for apple. Anyway I've used rosetta on my Athlon64 and it works fine without any special hardware.
-
tnt23 wrote:
Pre-encoded MFM file would be 1.8M as opposed to original 880K ADF. This will take twice as much time to load. To keep things simple, I'd produce fatter files, like 2.5M, and that'd be three times longer.
But wouldn't you be able to transfer the data on the fly (i.e. w/o 'loading')?
-
Zac67 wrote:
tnt23 wrote:
Pre-encoded MFM file would be 1.8M as opposed to original 880K ADF. This will take twice as much time to load. To keep things simple, I'd produce fatter files, like 2.5M, and that'd be three times longer.
But wouldn't you be able to transfer the data on the fly (i.e. w/o 'loading')?
I am not at all sure. Even if I go for 16MHz I'd get ~100ms of loading time per cylinder, and that wouldn't allow the virtual drive to be reactive enough (~3ms per track with the real drive).
-
Hello Dennis!
Great to hear that your Audio is working now.
It seems like I'm not the only one working on this! Can you tell some more details about your project? Are you working on a "chipset on a PCI card" or are you creating a core for the c-one?
About two years ago I was working on a PCI Card "Paula Card". I implemented all features of the Paula Chip - and a little more. It was only a prototype with an Altera 1k50, 2mb "chip"-Ram and soldered wires.
The 2MB RAM were accessible under Windows. And I was able to read an Amiga MFM Track into this buffer, and write it back to the disk (DD and HD) with a common disk drive attached to it. I was nearly done with audio, but I never tested it. :-(
I thought that this is not really helpful for me. I wanted to get a stand-alone device with all advantages of the Amiga. Not running under Windows. WinUAE does that in a brillant way.
Because of this I began the "NatAmi"-Project. The "Native Amiga" features flexibly attached to any CPU. I do not want to rebuild the Amiga bit-by-bit. I want to rebuild all its features and its interface, the ChipRegisters.
At the Moment I am working on my DMA system which is the base of the idea. I am doing this on the c-one board because it has PCI, two Altera FPGAs, VIDEO out, ram sockets, etc..
For my testing I built a CPU card. It was a PCI-IO-Card with an Altera Chip and memory (a not fully working spare-part from development). I soldered a 68030 CPU with wirerap lines ontop of it.
I think it is worth to take a picture, it has about 100-150 wires...
But It is working, the CPU can access its own RAM and other locations through its PCI interface. Only Flash ROM (for Kickstart) is still missing.
Now I am trying to get access from the CPU to the c-one board.
And yes, I use an own delta sigma modulator to avoid the same problem you are facing; the conversion of the sample rate. Also, a delta-sigma modulator is very easy to implement in an FGPA. (and cheap too!)
How is the resulting quality? - Are you using external filters?
PS: There are DAC's who are flexible with their sample-rate requirements, as long as you scale the masterclock accordingly. Also, older DAC's like the TDA1541 or TDA1543 are just plain 16bit converters; they accept any samplerate upto about 176KHz.
Hmm... thats a good idea! - My "old" CD-Player has a TDA1541. Do you
know if they are still in production?
Again my regards,
Thomas
-
I hope you'll stick with SD cards. They're everywhere, affordable and some of them are lightning fast (up to ~25MB/sec).
Plus they're more handy than CF cards, take up less space and look nicer ;-)
-
I hope you'll stick with SD cards. They're everywhere, affordable and some of them are lightning fast (up to ~25MB/sec).
Remember what he said about them -- you can only use them in MMC mode, where you won't achieve anywhere near 25Mb/s.
-
@ bluea
How is the resulting quality? - Are you using external filters?
Sound quality is about as good as my A3000. The modulator itself is just a very simple 1st order device but as the Amiga only outputs 8bit samples at a relatively slow sample rate, it is good enough.
I am using an external 2nd order passive filter that directly drives the output.
Hmm... thats a good idea! - My "old" CD-Player has a TDA1541. Do you
You have a CD player with a TDA1541 :-o , guard it with your life! The TDA1541 is one of the best audio DAC's ever made! I have bought a broken(!) second hand Philips CD player just to solder out his DAC. Unfortunately, they are not manufactured any more and are hard to get. The TDA1543 is very similair but easier to get.
Succes with your project!
Dennis
-
Don't forget that many programs (especially trackers) used the ability to modify the volume of one channel with another to achieve 14bit audio.
-
@JLF65
Don't forget that many programs (especially trackers) used the ability to modify the volume of one channel with another to achieve 14bit audio.
Uhhhm. Didn't 14bit audio use one channel with full volume, and 2nd with volume 1? The full volume channel would be the upper 8 bits, and the 6 uppermost bits of the volume-1 channel the lower 6 bits.
-
By the way, I'm not sure I would want Minimig inside a PC -- I would probably want it as a seperate console with keyboard, which I could plug joysticks or mice into as usual, and also have different types of video out (composite, digital, etc)
-
@Dennis
Speaking of floppy emulation, you do support other than 11 blocks per track and 81 and 82 tracks?
Diskspr3.lha (http://www.aminet.net/package.php?package=disk/misc/Diskspr3.lha)
I guess these images would require other than ADF format... Maybe some MFM image like what SPS (http://www.softpres.org/?id=home) use? Or perhaps add direct support for SPS images. That'd surely rock.
-
By the way, I'm not sure I would want Minimig inside a PC -- I would probably want it as a seperate console with keyboard, which I could plug joysticks or mice into as usual, and also have different types of video out (composite, digital, etc)
I agree with you for 100% - this would be a really new amiga =) And I would like to have irda, bluetooth and mmc/cf (not just any emulator of FD) card integrated reader and HD to store all info... =) Just a thought :)
-
AFAIK x86s have had a byte-swapping instruction for ages.
-
Piru wrote:
@JLF65
Don't forget that many programs (especially trackers) used the ability to modify the volume of one channel with another to achieve 14bit audio.
Uhhhm. Didn't 14bit audio use one channel with full volume, and 2nd with volume 1? The full volume channel would be the upper 8 bits, and the 6 uppermost bits of the volume-1 channel the lower 6 bits.
@Piru
As far as I know that's how 14-bit modes work, yes.
I've never seen a genuine practical application of channel coupling; all the trackers etc I ever looked at that don't do software mixing just modified the period/volume registers for each channel directly.
-
Dennis, any further updates on Minimig?
-
Another update :-)
Because the audio is now working, the core of Minimig is more or less complete. What I know need to do is finish all the loose ends; making Minimig load it's FPGA core from the mmc card, adding the onscreen display, adding keyboard interface, etc.
Currently I am working on the scandoubler. Untill now I was using a Commodore 1940 monitor in 15kHz mode. With a scandoubler I will be able to use it in 31kHz mode as well as connecting Minimig to other crt's and/or tft's. I also have added a 15kHz mode in which vsync becomes 3.3V and hsync becomes a 15kHz composite signal. With the proper cable, Minimig can be connected to the RGB scart input of a television set, just like in the old days.
I am working hard to get all this done for the demonstration on 18 februari in Maarssen (Netherlands). Click here (http://commodore-gg.hobby.nl/feb18.htm) and scroll down for more information.
Dennis
-
@Dennis
I am working hard to get all this done for the demonstration on 18 februari in Maarssen (Netherlands). Click here and scroll down for more information.
Good luck sir!
-
Congrats on getting the audio working!!! When can I buy one? :-D And do you have any plans to add AGA?
--
moto
-
Just a thought ... If planned to go commercial, will it be legal ?
-
ptek wrote:
Just a thought ... If planned to go commercial, will it be legal ?
:lol:
-
No kidding ... it reminds me the Bernie's nightamre called Amithlon :-o
-
ptek wrote:
No kidding ... it reminds me the Bernie's nightamre called Amithlon :-o
Oh, sorry I though you were making a joke. No there are no legal issues unless Dennis has used some copyrighted cores in his FPGA...
Also if he wants to sell it (I know I want one), he can't sell it with a kickstart or workbench... Maybe AROS would be a good fit :-D
-
bloodline wrote:
Also if he wants to sell it (I know I want one), he can't sell it with a kickstart or workbench... Maybe AROS would be a good fit :-D
AROS would work, or bundling it with Amiga Forever would also get rid of those issues, or directly licensing the files from AI as Amiga Forever has done.
-Tig
-
Dennis wrote:
Another update :-)
Because the audio is now working, the core of Minimig is more or less complete. What I know need to do is finish all the loose ends; making Minimig load it's FPGA core from the mmc card, adding the onscreen display, adding keyboard interface, etc.
Currently I am working on the scandoubler. Untill now I was using a Commodore 1940 monitor in 15kHz mode. With a scandoubler I will be able to use it in 31kHz mode as well as connecting Minimig to other crt's and/or tft's. I also have added a 15kHz mode in which vsync becomes 3.3V and hsync becomes a 15kHz composite signal. With the proper cable, Minimig can be connected to the RGB scart input of a television set, just like in the old days.
I am working hard to get all this done for the demonstration on 18 februari in Maarssen (Netherlands). Click here (http://commodore-gg.hobby.nl/feb18.htm) and scroll down for more information.
Dennis
You my friend are a GOD!!!
You have done so much progress in so little time that i almost can't keep up just reading about it.
This will go down in history as one of the greatest Amiga "HW hacks" of all time.
Keep it up.
//Niclas
-
Niclas wrote:
"You my friend are a GOD!!!
You have done so much progress in so little time that i almost can't keep up just reading about it.
This will go down in history as one of the greatest Amiga "HW hacks" of all time.
Keep it up."
Indeed. It would be nice if someone could arrange to record the demonstration and upload a video somewhere later.
I'd love to see the demonstration!
-
I'll second that.
-
Any perspective where your Minimig core will lead to? Are you more willing to go commercial or make it open?
What's the approximate price tag and amount of time to build a Minimig?
-
Tigger wrote:
bloodline wrote:
Also if he wants to sell it (I know I want one), he can't sell it with a kickstart or workbench... Maybe AROS would be a good fit :-D
AROS would work, or bundling it with Amiga Forever would also get rid of those issues, or directly licensing the files from AI as Amiga Forever has done.
How could AROS work? AROS can no more run on MiniMig than it can on a real A500 surely?
--
moto
-
Hi Dennis,
excellent news and I wish you the best success in whatever venture you want to go with the miniMig (commercial or eventual possible open source for the hobbyists).
I know you've done your best to make the miniMig OCS chipset as exact as you can, but I was wondering (if you have the time and inclincation) what the results of the following undocumented chipset tests are, so the enthusiasts out there know exactly how compatible with a real OCS A500 the miniMig is? (all are written by Toni Wilen, author of UAE - and fully documented at the English Amiga Board - http://eab.abime.net/showthread.php?t=19676)
"I have found and confirmed many undocumented chipset features when improving emulation and i thought there may be some interest in seeing them documented..
Some of these may be common knowledge.
I'll start with the copper:
1. Writing to "dangerous register" with COPCON bit 1 zeroed stops the copper (See AGA.guide COPCON for more information)
- SKIP doesn't really skip anything, it just prevent's next MOVE to do its write. SKIP followed by WAIT or another SKIP never "skips". (Game Apocalypse, uses SKIPs as nops in copper list)
- skipped MOVE to "dangerous register" (and COPCON zeroed) also stops the copper (Rainbow Island requires this.. it has quite buggy copperlist.)
- Copper is waiting, disable copper DMA, access both COPJMP1 and COPJMP2, enable DMA -> copper stops.
Copper stops = stays stopped until next vblank
More random info later..
EDIT: COPJMP1+COPJMP2 stop condition updated.
2. ECS/AGA and DDFSTOP >= 228 (0xe4) results in full horizontal overscan bitplane DMA. Like someone had set DDFSTRT=0x18 and DDFSTOP>=0xd4.
Does not happen in OCS.
Also some "illegal" DDFSTOP values can cause complete display corruption, syncs get messed up etc... (don't remember the exact values now)
Non-aligned DDFSTRT values cause bitplane delays (BPLCON1) work incorrectly. Basically hidden offset which depends on DDFSTRT is added to shift value. (I have called it "delayoffset". New Zealand Story, James Pond II AGA, Pinball Illusions, Loons Docs disks, Overkill AGA etc..)
3. Weird bitplane effect used in SWIV's scoreboard.
PF2P > 5 (BPLCON2) and BITPLANES == 5 and NOT AGA
- pixel in bitplane 5 = zero: planes 1-4 work normally (any color from 0-15 is possible)
- pixel in bitplane 5 = one: planes 1-4 are disabled, only pixel from plane 5 is shown (color 16 is visible)
4. Mysterious world of blitter line mode..
- D-channel can be disabled. No effect on anything.. (even speed of blit stays the same)
- C must be enabled (nothing is drawn without C)
- B (not tested)
- disabling A stops BLTAPT from updating (result is incorrect lines)
- BLTDMOD is not used in line mode. BLTCMOD is used for both C and D.
And here is the most interesting undocumented feature ever:
- First pixel is written to address pointed by BLTDPT, all following pixels are written to address pointed by BLTCPT! (Demos Cardamon and Cardamom)
comments from PHoton:
This is most interesting, because you normally want 1 pixel of the line to be omitted if you want to fill the poly afterwards... Otherwise you get double pixels plotted at the corner points of the poly, resulting in the infamous horizontal lines emanating from the corners when it's filled.
I did it by subtracting 1 from the blit height in the linedraw routine just before drawing, Slayer did it by XOR'ing an extra pixel so it goes away, but could this mean that by pointing BLTDPT to a trash area, the pixel goes there and the rest of the line goes to BLTCPT and is correct for filling?
If it's true, then a bunch of cycles would be saved in the linedraw routine
a) one register saved and one sub.w (up to 12 cycles on A500)
b) one longword or word write to memory for the BLTDPT. UNLESS BLTDPT is increased after the blit to point to just beyond the last pixel. (up to 12 cycles more)
BLTDPT == BLTCPT after blit ends. (Demo Deformations won't work properly without this)
5. Attached sprite:
Documentation says: odd sprite's SPRxCTL bit 7 must be set.
ECS (probably OCS too, not tested yet): sprite is attached if odd _or_ even (or both) sprite SPRxCTL bit 7 is set.
AGA: only odd sprite SPRxCTL bit 7 enables attachment.
6. More or less useless info..
CLXDAT bit 15 is always set. ("not used" says HRM) Game "Barney and Freddy Mouse" breaks if this bit is not set...
HRM says [sprite to playfield] "collision detection does not change when you select either single- or dual-playfield mode". This is not true. There is small but significant difference..
7. Switching bitplane DMA on during DDFSTRT ~<= hpos ~<= (DDFSTOP - size of fetch unit) can have 3 different results:
- OCS: nothing happens, bitplane DMA fetches don't restart until next line. (the rest of current horizontal line is also filled with background color)
- ECS/AGA: DMA restart normally if DMA was previously disabled when hpos was outside of DDFSTRT and DDFSTOP. (Sixth Sense Investigation's buggy copper list requires this..)
- ECS/AGA: display processing jumps to "DDFSTOP passed"-condition (process modulos after next fetch unit and end current line) immediately if DMA was previously disabled when hpos was between DDFSTRT and DDFSTOP. (Found this by doing tests on my real A1200)
Thank you.
Gizmomelb
-
@ GizmoBelb
1) Minimig behaves as described except for the dma on/dma off bit. The copper stuff is required to even make it to the famous 'hand' screen.
2) "illegal" ddfstop value will corrupt screen but syncs (from agnus to denise/paula) will remain intact. non-aligned ddfstrt values do nothing to the offsets.
3) I don't know what would happen in Minimig in that case. Color selection is pretty complex...
4) Minimig behaves as desribed except for the first pixel part. (This IS true by the way, I've seen that behaviour on my A3000 too and I will implement it in the future) Again, all this stuff is required for a correct boot screen.
5) Minimig behaves as described.
6) Minimig behaves as described.
7) Minimig will start fetching bitplane data as soon as dma is enabled, even if it is halfway a line. If a program requires this, it will corrupt the screen.
Offcourse, all these 'features' can be added later on. But for now, I am focussing on getting everything demonstration friendly. As I still have only 512kB of memory I can't test all programs anyway.
Dennis
-
@ Zac67
Any perspective where your Minimig core will lead to? Are you more willing to go commercial or make it open?
I really don't know yet, I like the open-source idea, especially if it would be combined with AROS. But otoh, I would just hate it if somebody else grabs the code and stuffs it into a joystick (or mouse?) and makes lots of money. If it goes open-source, it belongs to the community and not to a big toy company. All this started out as a hobby project and it still is just that; a hobby project. I realize that there is a lot of interest from the community and I will make it available to the community in one form or another, I just haven't decided yet.
Dennis
-
The best thing to do is to open the cores to the open source community, but hold all the copyrights yourself, so individuals can do what they want, but companies cannot.
-
Dennis wrote:
@ Zac67
Any perspective where your Minimig core will lead to? Are you more willing to go commercial or make it open?
I really don't know yet, I like the open-source idea, especially if it would be combined with AROS. But otoh, I would just hate it if somebody else grabs the code and stuffs it into a joystick (or mouse?) and makes lots of money. If it goes open-source, it belongs to the community and not to a big toy company. All this started out as a hobby project and it still is just that; a hobby project. I realize that there is a lot of interest from the community and I will make it available to the community in one form or another, I just haven't decided yet.
Dennis
Checked the linked site with the picture and must say I am impressed. You are doing this alone? I would love to be able to use your results to create a personal device in the future, be it a joystick device, a mini Amiga clone that is portable, or an Amiga clone laptop. If it is just for personal use and not for commercial release, I wonder if there will be any copyright infringement problems when loaded with a legal copy of the ROM and AmigaOS?
Keep up the great work.
-
Booya! I guess this is a another long-time-lurker's post... Seems like you're getting eventually actually done with what I've just mused by myself in my cave for ages; reverse-engineering this lil' bugger :-D
Anyway, great job Dennis!
If you ever release the core and/or "blueprints" of your job as a whole, I suppose, according to all the previous posts, I'd not be the only one who'd just love to get their hands dirty with it.
-
Hi Dennis,
thankyou for taking the time to reply to my in-depth questions about chip compatibility - very interesting!
Also thanks and congrats on all the hard work getting the OCS chipset working. Hopefully in the future (if you still have the desire and drive) then ECS support will be added (should be relatively 'easy' to add - as it is well documented like the OCS chipset). However I don't expect AGA support as there is little/no technical documentation for it available. Also the main reason why most Amiga fans would be interested in the miniMig would be to play old games - the majority of games would work under OCS/ECS, not AGA. (yes, I realise there will be a vocal minority screaming for AGA as they need it to run some specialised utility etc. but you are not the target audience if the miniMig goes commercial).
Speaking of if your project goes commercial - is there any chance you will be able to 'build in'/make accessible hacks/upgrades into the design, much like Geri did with the C64 DTV?
for example: being able to add an IDE HDD, or having access to the 68000 pinouts and being able to add a homebrew 86pin Zorro slot connector for RAM expansion, or adding a real floppy disk would all be fantastic.
Thanks.
-
My suggestion: split this long thread into parts.
For CPU, perhaps Dennis should create several daughter boards / terminator boards:
68K, ColdFire, PowerPC, PCI (host X86), PCIe (host X86), USB etc.
-
yes, I realise there will be a vocal minority screaming for AGA as they need it to run some specialised utility etc. but you are not the target audience if the miniMig goes commercial
Excuse me! Have you just appointed yourself head of marketing strategy for Dennis Designs B.V. ?? :crazy:
Maybe (Mr. 13 posts!) spend a little time on the forums before you start deciding what people do and do not want. :lol:
Anyway, Dennis has been at pains to state that this is a hobby project - he might just build the one and then burn the plans!
Muahahahahahahaha! :madashell:
-
Colin_Camper wrote:
yes, I realise there will be a vocal minority screaming for AGA as they need it to run some specialised utility etc. but you are not the target audience if the miniMig goes commercial
Excuse me! Have you just appointed yourself head of marketing strategy for Dennis Designs B.V. ?? :crazy:
Not at all, I was merely speculating aloud what would seem to be the obvious reality if the mimiMig were developed commercially. The commercial target audience for an 'Amiga in a joystick' (similar to the Atari Flashback and C64 DTV) would be mid 30's nostalgics and their kids, who want to play a few games. Considering the ratio of OCS/ECS games compared to AGA games, and the availability of technical information regarding the OCS/ECS chipset versus the AGA chipset - then the effort required to research and develop a commercial FPGA (or whatever) AGA Amiga is simply not worth the outlay. Also putting it bluntly, I personally do not think there were any outstanding AGA only games released. All were slightly upgraded ports from ECS games, or blocky lo-res DOOM clones.
If my speculation of commercial reality for the Amiga offends, please ignore me and go back to dreaming about Amiga's comeback.
Colin_Camper wrote:
Maybe (Mr. 13 posts!) spend a little time on the forums before you start deciding what people do and do not want. :lol:
No egos on this forum hey? What does amount of posts I've made have to do with the amount of posts I've read? Or the simple fact that in this 14 page (so far) thread, that around every fifth post is asking if AGA Support is coming, or how much they'd love AGA support? That would seem to indicate what people want.
Colin_Camper wrote:
Anyway, Dennis has been at pains to state that this is a hobby project - he might just build the one and then burn the plans!
Muahahahahahahaha! :madashell:
That is his choice and one possible outcome of his work.
Or he might sell it as a commercial venture, but then the manufacturer decides there isn't enough of a market so it is never made. I hope that doesn't happen.
I wish Dennis the best of luck with whatever he plans to do with the miniMig, however if it is sold commercially then maybe it would be good if he just sold/leased the use of his design for a couple of years so that he retained copyright and could then release the design to the hobbyist market once the product has finished it's commercial shelf life.
-
Also putting it bluntly, I personally do not think there were any outstanding AGA only games released. All were slightly upgraded ports from ECS games, or blocky lo-res DOOM clones.
Weren't ANY outstanding AGA games?
Xtreme Racing is one of the best multiplayer racing games ever made, on any platform.
Alien Breed 3D (the first one, not the awful sequal) may well be very blocky, but its actually a very very good game. Better than Doom, in my not so humble opinion.
Slam Tilt is another example of an awesome AGA game.
I also really enjoyed Bubble & Squeak AGA, and James Pond 3 AGA, but those weren't so popular.. but then I always liked platformers.
T-Zer0 was very underrated. It was very hard, but if you stuck at it, it really was an excellent shoot 'em up, IMHO.
Guardian was also really good. It was in Amiga Powers all time top Amiga games, for example.
What about Deluxe Galaga AGA? or Blitz Bombers AGA? or Napalm? or OnEscapee? or Super Stardust? ...
-
As stated, it's my personal opinion. I was never impressed by any AGA games.. they were just slightly colour enhanced versions of the OCS/ECS game. In a lot of cases it was difficult to tell the difference between the ECS and AGA versions.
Alien Breed 3D 1 was good, but nearly unplayable on a stock A1200/CD32 - you needed a faster CPU for the game to be smooth and enjoyable.
I'm not purposely trying to rile people up, but AGA (games) never excited me with their visuals or quantity - compared to OCS/ECS games. To me AGA was simply not a big step up. And hopefully this is the end of the off topic posts in Dennis' thread.
-
Maybe (Mr. 13 posts!) spend a little time on the forums before you start deciding what people do and do not want.
Maybe (Mr. 123 posts!) you should spend a lot of time posting on Amiga.org before you go around being rude to new members.
Signed,
Mr 4335 posts. ;-)
-
Maybe (Mr. 123 posts!) you should spend a lot of time posting on Amiga.org before you go around being rude to new members.
I thought HE was the one being rude and arrogant - I was just pointing this out in a measured and tactful manner.
You don't jump into the end of a 14 page thread, with 13 posts EVER under your belt, and presume to speak for EVERYONE interested in Dennis' work.
Anyway he reposted with an IMHO so that's good enough for me.
Hey! If you think I was being rude there I would hate to tread on your toe whilst eying up your missus! :lol:
-
I'd eventually like to play Roadkill AGA on it, too, but you can't have everything ;)
I'm just sitting back, seeing how Dennis progresses with this. All I want at the moment is a fully functional OCS / ECS Amiga.
-
Well, Minimig is looking more of a reality all the time. Perhaps it's not TGtBT?
I work in the industrial/product design industry and am currently working on my own products. You have to get good advice when protecting your IP.
Taking Jeri's experiences with big toy companies as an example, I'd be very careful licencing the IP to big business.
I'd find out just what the legal stance is on replicating/reverse engineering a technology such as the A500 without a licencing agreement from the current IP owners of that tech? You may not actually own anything commercially viable?
If it ever gets into production, be sure and insist on a percentage-profit share agreement, just in case it does a c64dtv and outsells expectations, even a tiny percentage will get you a nice little cheque in the mail each year.
All the best.
-
xeron wrote:
Xtreme Racing is one of the best multiplayer racing games ever made, on any platform.
I for one think that XTR is the best argument for having AGA support in minimig :-) Imaging having a portable Amiga you can carry round to your mate's house with 4 joysticks, plug it in to his telly and you're away. Great!
xeron wrote:
T-Zer0 was very underrated. It was very hard, but if you stuck at it, it really was an excellent shoot 'em up, IMHO.
I like T-Zer0 too. Unfortunately, I tend to get frustrated with it, mainly because I'm not actually very good at games like that :roll:
--
moto
-
As a Trade Mark Attorney, here is my non-binding advice to Dennis:
If you do go down the commercial route and decide to license your invention to a company in return for money. Make SURE you demand a clause in the contract giving the company a time period to commercialise the IP, if they do not commercialise it in that time period then you can retake control of the IP.
That way you give the company say 6 - 12 months to commercialise the product, if they don't - it's yours again.
-
Wow, this thread has grown to be huge. It's getting to be very hard to track this thing. Who knows, it might even get it's own section in the website, eh non ?
-
yes, I'ld like to see a thread devoted to only dennis'es updates about the project. it's too tiresome to read all of it when you're looking for something. (is there a nick filter in the threads ?)
-
Good work Dennis. :-D
A 'miggy in a drive bay...mmm any chance of getting a Coldfire solution working? :lol:
For the record I would buy this product if and when it supports AGA and 68030@ (at least) 33MHz, works well amd runs OS3.9.
As a portable solution a touch screen would also be required before I parted with cash. I know a man who used to programme Amiga based touchscreen systems so if you need a driver let me know and I'll bend his arm so to speak. ;)
Well done again and keep it up. :pint:
-
Touchscreen, you say? I wrote the driver for the tabby graphics tablet, and still have the source... It would be trivial to convert it from reading the tabby format data into reading another form of data off a serial port.
Took a fair bit of computing power tho, as it wasn't the most efficient of things. It was also implemented as a commodity.
-
chiark wrote:
Touchscreen, you say? I wrote the driver for the tabby graphics tablet, and still have the source... It would be trivial to convert it from reading the tabby format data into reading another form of data off a serial port.
Would probably be better to do this with hardware in order to maintain compatibility with non OS-friendly software (i.e. games).
--
moto
-
Not sure there is a way to do it in hardware. Touchscreens at least when we used 5000 of them or so on our Amiga gaming machines ran off of a serial port. My driver emulated the mouse. Actually after I left VGT, lots of people used to use my driver to run there video toasters, with a touchscreen moniter, which works pretty cool. If dennis really wants/needs a touchscreen driver, he's welcome to mine, its even approved by the gambling commission :).
-Tig
-
Sounds like Tigger's the man to talk to then :-D
-
Bouncy! :-D
-
countzero wrote:
yes, I'ld like to see a thread devoted to only dennis'es updates about the project. it's too tiresome to read all of it when you're looking for something. (is there a nick filter in the threads ?)
Try this (http://www.amiga.org/search.php?action=showallbyuser&mid=6&uid=12811) - (get it by looking at the "show all posts" link for a given user's profile). Seeing as Dennis is posting pretty much only about his project, it's easy to see his progress.
-
Twin wrote:
If you do go down the commercial route and decide to license your invention to a company in return for money. Make SURE you demand a clause in the contract giving the company a time period to commercialise the IP, if they do not commercialise it in that time period then you can retake control of the IP.
Hmm, handy advise. I know quite a few people who didn't seek advice from lawyers until it was too late, and really got burnt in their deals.
Got to love this forum.
-
@Twin,
What of the original Amiga hardware patents?
I'd imagine after all these years they must be getting close to expiring?
?
-
Any more updates on Minimig? :D
-
My friend, you have GOT to get one of these to Ben Heckendorn (http://www.benheck.com) I get all goose-pimply at the thought of an hand-held Amiga with a touch screen and a wireless PDA keyboard, and he's the one to do it.
Keith
-
@Deadly Dad
Hey, thanks for that link. I like that NeoGeo cabinet. I`ve been working on a Mame cabinet myself.
-
Been following the thread for a while now. Things are getting exciting!
What Ben Heckendorn has done with the PSOne is what I'd like to see in an Amiga portable. Imagine that, but with a (slightly) expanded CD32 inside.
Still, I'll buy anything that develops from this project.
Great work, Dennis!
-
This is for me the most exciting Amiga project underway at the moment, The market for an open gaming handheld at the moment is unbelievably huge, OK so people are going to say check out the GP2X (I have one already and it is really really nice) but there are two points to note here
1 GPH the makers of the GP2X are totally disregarding the GPL and have caused an uproar which may not go away
2 Check the GP2X forums for the ammount of posts about emulating an Amiga on the GP2X, There are a hell of a lot and it has just about been decided that it is impossable at a playable speed (If you hear any strange sounds in the wind it's just the groans of the GP2X owners as we realised we ain't getting handheld mig)
Maybe we can help out a bit, I have plenty of web space for my business's so i can donate some space and i can design a little site with updates and so on, Maybe Dennis could send some info monthly as an update ? (Pictures would also be great) but still it can be updated from this thread anyway :)
This offer stands either way Dennis decides to go, Open or commercial, I couldn't care less which way he goes as long as he goes (I want whatever this turns into because i have an interest in FPGA, I don't work on them myself but someone i work with does)
Anyway good luck with your plans to try and get it ready for that show so soon :-o
PS congrats on the darts too, Two dutchmen in the final for the first time ever ;)
-
@ Flexinoodl
Supposedly GPH have told several reputable people in the GP2x community they will be releasing the firmware source for 1.3.0 next week. They have intimated it will be everything, not just the kernel.
Apperently one of the German devs or resellers of something will be the official download site for it.
I agree, two 200MHz ARM9 cores will probably never have a "solid" amiga emu on it. Maybe some games, I don't know.....
I think the only issue would be a good screen. Can Minimig drive an LCD directly? Would refresh rates with an LCD be an issue? (hopefully not)
I for one really hope this goes open source with a reservation on commercialisation in Dennis. That way hackers can play with it but can't sell it, but Dennis can still talk to manufacturers to produce it if he wants. That'd be great..... and hopefully make Dennis uber-wealthy and maybe he would then decide to buy up all the Amiga IP and resurrect it :) ......................maybe..............
-
Hi Dennis!
"Currently I am working on the scandoubler."
What about your scandoubler?
Perhapes I can help you.
I have done a smaller Project then yours:
http://www.symbos.de/trex.htm
But there I have a working Scandoubler in VHDL.
Currently I'm working on an 32/16bit RISC which can execute 68000 Code. The RISC needs 64KB RAM for the EMU Code and 1000 LC's.
Viele Grüße
TobiFlex
-
Another update:
I have finished the scandoubler. It was not as straight forward as I hoped it would be because I'm crossing clock domains. Minimig's bus (and the rest of the logic) is clocked at 7.09MHz. This is also the lores pixel clock. For hires screens I use a DDR scheme, in which a pixel is clocked out on both the falling and rising edge of the clock. For VGA output however, a pixel clock of 28.36MHz is needed (non-DDR). As it is now I use a blockram to buffer a single line of video data and to synchronize the VGA beamcounter to the Amiga's beamcounter. This way I have no connection between the 2 clock domains other than the blockram which nicely handles the clock problem.
Note that it is a scandoubler and not a flickerfixer. A flickerfixer requires that a whole video frame be buffered, for which the fpga has not enough memory.
I am now working on cleaning up the PIC code and adding the user-interface for selecting floppies, cores and rom images.
@motorollin
And do you have any plans to add AGA?
AGA would be a LOT of work. Currently I have no plans beyond OCS.
@tobiflex
Impressive project! Thank you for the offer on the scandoubler, but I just finished it. The FPGA board you are using looks very nice. How many LE's does the FPGA have?
Dennis
-
Flexinoodl wrote:
2 Check the GP2X forums for the ammount of posts about emulating an Amiga on the GP2X, There are a hell of a lot and it has just about been decided that it is impossable at a playable speed (If you hear any strange sounds in the wind it's just the groans of the GP2X owners as we realised we ain't getting handheld mig)
The GP2X Wiki (http://wiki.gp2x.org/wiki/Emulation#Commodore) suggests it could have Amiga emulation (from what I've read, OCS may be possible, AGA is almost certainly out of the question.)
It's probably going to be some time before the developers get their heads around the GP2X enough to get enough power out of it for even OCS emulation though.
-
@Dennis
How many LE's does the FPGA have?
There is a Altera EP1C6 with 5980 LC's on Board.
Viele Grüße
TobiFlex
-
Dennis wrote:
@motorollin
And do you have any plans to add AGA?
AGA would be a LOT of work. Currently I have no plans beyond OCS.
Dennis
Hi Dennis,
Thanks for the update and dealing with yet another AGA request (which resulted in my (NOT - as explained by Colin Camper) being flamed when I pointed out in an earlier thread would not be forthcoming due to lack of technical info, commercial worth and/or personal interest on your behalf).
Might I please enquire if there is enough technical info available and the estimated level of difficulty to expand/improve the OCS chipset to ECS compatibility?
Regards,
Gizmomelb
-
Dennis wrote:
Another update:
I have finished the scandoubler. It was not as straight forward as I hoped it would be because I'm crossing clock domains. Minimig's bus (and the rest of the logic) is clocked at 7.09MHz. This is also the lores pixel clock. For hires screens I use a DDR scheme, in which a pixel is clocked out on both the falling and rising edge of the clock. For VGA output however, a pixel clock of 28.36MHz is needed (non-DDR). As it is now I use a blockram to buffer a single line of video data and to synchronize the VGA beamcounter to the Amiga's beamcounter. This way I have no connection between the 2 clock domains other than the blockram which nicely handles the clock problem.
Note that it is a scandoubler and not a flickerfixer. A flickerfixer requires that a whole video frame be buffered, for which the fpga has not enough memory.
Dennis
Hi Dennis,
please forgive my asking, but from the VGA timings given on this page:
http://www.epanorama.net/documents/pc/vga_timing.html
wouldn't a 640x512@60Hz VGA timing clock be 26.853. ?
Also how do you handle the PAL 50Hz video output to NTSC/VGA 60Hz conversion? I'm in PAL land as well, but if you can convert PAL to (effective) NTSC for VGA output, then possibly having (if possible, and if it's cheap enough) a VGA to NTSC TV circuit included would double the miniMig's prospective market. Ok, its a kludge, but it could give the unit a higher commercial appeal (if that is what you are after).
Regards,
Gizmomelb
-
AGA request (which resulted in my being flamed
I didn't flame you. I just pointed out that you don't speak for anyone else, necessarily. Nor do I. I think you are being a little precious if you took that as a flame.
not be forthcoming due to lack of technical info, commercial worth and/or personal interest on your behalf)
There you go again! You don't speak for Dennis! Let Dennis tell us how much personal interest he has in AGA. He pointed out earlier that AGA would be 'A LOT' of work. That's all!
I tried to make this point earlier in the thread, with some humour, and maybe it didn't get across that way because of language or cultural reasons. I wasn't flaming you! :-)
-
Colin_Camper wrote:
AGA will not be forthcoming due to lack of technical info, commercial worth and/or personal interest on your behalf)
There you go again! You don't speak for Dennis! Let Dennis tell us how much personal interest he has in AGA. He pointed out earlier that AGA would be 'A LOT' of work. That's all!
Contributing to Colin's off-topic-ness and further polluting the signal to noise ratio of this thread:
No, I am not trying to speak on Dennis's behalf, and I apologise to Dennis if I have mis-represented his intentions. I am simply voicing my opinions based on Dennis's comments that AGA support would not be easy, and with the logic that researching and developing AGA support (for such a limited number of software titles, compared to OCS and ECS titles) would not be commercially viable.
Dennis's statement of intention is in the first message of this thread:
"My aim with Minimig is to built a complete OCS A500"
second post by Dennis:
"AGA: well, to be honest I haven't really thought about that yet. My first goal is to get an OCS system going 99.9% correctly."
third post:
"AGA would be a LOT of work. Currently I have no plans beyond OCS."
So from reading the above quotes from the author, I have logically deduced that Dennis has no personal interest in researching AGA support (which I do not consider a loss).
Commercially I feel it would be a 'classic' Amiga which would interest the general public (due to the sheer number of game titles which could be marketed with the unit), with the hackers and die-hards being especially interested if its an accurate 'emulation' like the C64 DTV and the Atari Flashback 2.
-
Nice,
but dont put someting so good in something so bad (archtiecture wise) -ie just like a pegasos.
actually urs is a little better. ehhehe
-
mdma wrote:
Maybe (Mr. 13 posts!) spend a little time on the forums before you start deciding what people do and do not want.
Maybe (Mr. 123 posts!) you should spend a lot of time posting on Amiga.org before you go around being rude to new members.
Signed,
Mr 4335 posts. ;-)
:lol:
Signed,
Mr Silly# posts.
-
@gizmomelb
28.36MHz is the clock required for PAL, NTSC would be 28.64MHz.
I dont think dennis is doing anything other than standard scandoubling to 31KHz@50Hz
-
Hi Alexh,
Thanks for the info.. hmm there could be a lot of dissappointed LCD owners - most LCD displays won't sync that low.
-
wouldn't a 640x512@60Hz VGA timing clock be 26.853. ?
I'm not doing 60Hz but 50Hz VGA like alexH said. As this is not a standard VGA mode, I was unsure what timings to use. I got the right values by measuring the VGA signal on my A3000. On the A3000, a single VGA line looks exactly like a single PAL line, but then twice as fast. This implies 2 times the PAL hires pixel clock, hence 28.36MHz.
Also how do you handle the PAL 50Hz video output to NTSC/VGA 60Hz conversion?
I don't.
There are a lot of monitors that do sync down to 50Hz however. I have 2 Philips monitors, a 15" 105MB CRT and a 19" 190B TFT and they both sync down to 50Hz without problems.
Dennis
-
Hi Dennis,
Thanks for the info - it's nice to see a retro system developer concentrating on PAL for a change :)
Just for the benefit of those in NTSC countries however - would changing your main 7.09MHz system timer to 7.16MHz cause any problems with internal timing, or video output?
Regards,
Gizmomelb
-
Mick Tinker made fpga AGA, point is, it might be easier if someone allready had done that.
I ques that he doesn't hope to get money with it anymore.
That just offtopic / or near ;)
-
Mick Tinker and his work are almost certainly gone forever :(
-
from the pics and info I found on this website:
http://access.amigaworld.de/index.php?menu=1
and especailly the info here:
http://access.amigaworld.de/index.php?menu=13
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
-
gizmomelb wrote:
from the pics and info I found on this website:
http://access.amigaworld.de/index.php?menu=1
and especailly the info here:
http://access.amigaworld.de/index.php?menu=13
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
Thats right... Access just used stock bought from the Commodore liquidation... no new chips were made ater C= went tits up.
-
gizmomelb wrote:
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
The Access and Boxer were two different things. The Access was actually built and sold for one thing. :-D But it did use existing AGA chips from Commodore. The Boxer prototypes did use FPGAs to clone the custom chipset. Alteras, if I recall...
I don't fully understand the concern about lack of detail about the AGA chips innards. I mean, the various AGA screen modes are known, and enough detail is known to write programs that use those modes... so what's the deal... who really cares exactly what's going on inside. In fact, it ought to be better, since no 'tricky stuff' was known that game writers used and now has to be emulated. :-?
Jim
-
alexh wrote:
Mick Tinker and his work are almost certainly gone forever :(
so whatever happened to Mick Tinker ? IIRC it was stress / personal ...
-
-
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
Yes, I think Petro had virtually promised them an indefinite supply of AGA chips for this machine.
However, when the BoXeR was in development - these chips were not available in any great quantity so Mick Tinker decided to implement AGA in FPGA. Understandable as an OCS BoXeR would have been seen as a retrograde step. The irony is that he might have succeeded (like Dennis) with OCS and gone on to sell quite a few! :-)
-
I'm only a Computer Science undergrad, but the whole project sounds very interesting. It would be fantastic if the whole thing was opened up to the community.
How large (physical dimensions - mm) is the project?
Samuar
-
@Dennis,
i don't know how var you are (didn't read the whole tread). But i have some service manuals of the A500. If you are interested mail or PM me.
-
Remember there was this inane argument about why Gateway/Amino abandoned AGA compatibility? Some pinheads said it was because AGA compatibility was dead-end while others said it was because a lot of technical data on thie schemtics had been lost in the various transfers of IP (intelectual property)... and that Jeff Schindler and others didn't want to admit to that.
Guess who was right? Someday an secretly made audio recording may be released that will reveal the truth once and for all.
Dennis, have you tried to contact Jeri or Mick? What about custom chip designers for the AGA Amigas? I suspect that making personal contacts would be helpful to you.
-
@ Boing
I heard stuff had been lost. I didn't know it was Gateway though...
Dennis, have you tried to contact Jeri or Mick? What about custom chip designers for the AGA Amigas? I suspect that making personal contacts would be helpful to you.
Dave Haynie has been posting over at Amigaworld.net - I mentioned Dennis and the minimig to him and he checked it out. Needless to say he was very impressed! :-D
>"Whoa! No, I didn't know abotu this. He's the Jeri >Ellsworth of the Amiga, eh? Sounds very cool, I will check >it out. Just did... this is great. No question the state of >FPGAs is up to this kind of stuff: rapid development way >beyond what was possible back in the day.
>
>Curiously, in 1992 or thereabouts, they did the 8520 in an >FPGA, in preparation for integrating into the CDTV gate >array. This was a very expensive (over $100) FPGA in a >"tower" board that plugged in place of the CIA itself. So >that's a measure of how far this has come: the whole >chipset in a $30 (single quantity) part."
Amigaworld : Dave Haynie (lead engineer of C= Amiga) opinion on Amiga Successors (http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=16943&forum=2&start=500&viewmode=flat&order=0)
-
Just as long as the bitmaps range of titles work on it, that is all that matters :)
-
gizmomelb wrote:
from the pics and info I found on this website:
http://access.amigaworld.de/index.php?menu=1
and especailly the info here:
http://access.amigaworld.de/index.php?menu=13
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
I'm allowing myself to go a little off-topic here, but:
Wasn't there some plans for making a new batch of Access's? What happened?
-
Wasn't there some plans for making a new batch of Access's? What happened?
Don't know, but I've asked a question now in the corresponding german thread in the forum of amiga-news.de (http://www.amiga-news.de/forum/thread.php?id=10252&BoardID=1).
But given all the time that has passed since then, I doubt there's still life in that project.
-
Dennis can you check your PMs, thx :-)
-
Hi,
another update:
I have been working on the PIC code and have been leaving the fpga core alone for a while. I have also tested a lot more games and demo's (the ones that work with 512K) and fixed a last bug in the blitter special line draw mode, which sometimes placed 2 pixels on a line, causing subsequent hardware fills to be messed up. Almost all of the software I try now works very well, except that some demo's which use the blitter a lot run twice as slow as in Winuae. When I speed up the blitter by giving it more cycles than in the original, these demo's run normal. The weird thing is that older versions of Winuae (0.8 and older) behave exactly the same when blitter-exact-cycle was switched on. (good example is the lemmings cracktro) I have to investigate this further.
Just for the benefit of those in NTSC countries however - would changing your main 7.09MHz system timer to 7.16MHz cause any problems with internal timing, or video output?
No, it wouldn't but it wouldn't make it NTSC either. For NTSC output the beamcounter logic has to be changed.
it doesn't actually appear (to me) that the AGA custom chips had been re-engineered at all.
How far did Mick actually got? Did he acutally complete an AGA implementation?
How large (physical dimensions - mm) is the project?
When it's finished, It will probably be the size of a floppy drive / eurocard / 10cm * 16cm.
Dennis, have you tried to contact Jeri or Mick? What about custom chip designers for the AGA Amigas? I suspect that making personal contacts would be helpful to you.
For now I'm still working pretty much on my own. But I am now realizing that I will never get it 100% OK just by myself, so I will need some help in the nearby future. That is why the open-source idea appeals to me.
Dave Haynie has been posting over at Amigaworld.net - I mentioned Dennis and the minimig to him and he checked it out. Needless to say he was very impressed!
Cool! :-) I have one of his machines right here on my Desk. A perfectly working A3000. (although I do suspect it's PAL's from fading, need to replace them soon)
Dennis
-
Any chance you could test Minix 1.5 for Amiga on your setup?
I would be interested to hear how well it works.
Also, how much has the project cost you so far?
Samuar
-
How far did Mick actually got? Did he acutally complete an AGA implementation?
It is my opinion that he got to a similar stage that you are at. Majority of programs work, but not all of them 100%, especially demo's. The extended design cycle trying to iron out the bugs is possibly what killed the project commercially. However reading through older magazines, the BoXer II (PCI with custom chips in FPGA's) was never demo'd working.
But I am now realizing that I will never get it 100% OK just by myself, so I will need some help in the nearby future. That is why the open-source idea appeals to me.
Cool. However for the would-be developers to come onboard immediately, someone would have to organise the devkits to some extent.
o The Spartan FPGA devkit
o The 68k and I/O PCBS
o All the components / connectors & cables.
Ready made kits is probably impractical, so DIY kits containing all the bits where you solder it together yourself will probably be the way to go.
Any idea on a cost, approx?
-
alexh wrote:
Ready made kits is probably impractical, so DIY kits containing all the bits where you solder it together yourself will probably be the way to go.
Any idea on a cost, approx?
I'd be up for a kit like that. I love hobby stuff like that. :-D
Hopefully, something like that is doable at a reasonable price.
-
It'd be cool to build one of these. Sounds like it is coming along very nicely still. Maybe this is the hobby for me in the future, I've always been interested in hardware.
Dennis, any chance of a picture of the hardware connected to a monitor running, say, Sanity Arte or Kefrens Desert Dream? I'm missing my hardware geek pictures!
Edit: Have you seen the Amiga floppy emulator link at the top of ANN? Russian site, but it looks like someone's build a fully floppy interface compatible memory card floppy device. Might be interesting to you?
-
JLF65 wrote:
alexh wrote:
Ready made kits is probably impractical, so DIY kits containing all the bits where you solder it together yourself will probably be the way to go.
Any idea on a cost, approx?
I'd be up for a kit like that. I love hobby stuff like that. :-D
Hopefully, something like that is doable at a reasonable price.
I don't know if it's suitable or cost effective for this kind of project but here is a nice little card for $199:
http://www.xess.com/prod035.php3
No 68k CPU but....it's a million gates to play with.
-
Very nice project, it's really very interesting!
What about emulating 68k? It could make the design easier and better for commercial (or not) manufacturing.
What about c-one like manner? Using the hardware for emulate other architectures (Atari ST, all 8bit ones, some 16bit ones...).
I hope you will be able to simulate an A1200/A4000 system with 68060 (or with coldfire atached without emulate?) and some nice rtg gfx stuff. This could be the dream of most amigans: a cheap and small amiga system with most that they want in a small (uATX?) board!
What about the floppy emulation project? Maybe a join and creating a team could be quite interesting, specially in researching in the amiga hardware and sharing your efforts (and why not publishing it for others?).
-
Sorry Dennis, that Mick Tinker talk is my fault :) but I can't help, I was really exicitng with BoXeR
I really dream that FPGA AGa would be possible, it should be with amiga 1200 accellator port, dream about it you could play ABII3D-tkd in 800 x 600 screen or Napalm :)
Again sorry I don't try make your work less valuable.
But I really would like to heard what would you say / speculate about AGA FPGA
-
TH0072 wrote:
...here is a nice little card for $199:
http://www.xess.com/prod035.php3
No 68k CPU but....it's a million gates to play with.
Neat! Nice that it has integrated VGA and keyboard/mouse... And that seems like a reasonable price too.
If only I had the knowledge to create something good with it!
- Ali
-
InTheSand wrote:
TH0072 wrote:
...here is a nice little card for $199:
http://www.xess.com/prod035.php3
No 68k CPU but....it's a million gates to play with.
Neat! Nice that it has integrated VGA and keyboard/mouse... And that seems like a reasonable price too.
If only I had the knowledge to create something good with it!
- Ali
That DOES seem like a really nice proto board. I might get one if I don't find something better.
-
JLF65 wrote:
That DOES seem like a really nice proto board. I might get one if I don't find something better.
I kinda like the XEM3010 from these folks http://www.opalkelly.com/
It's a bit pricey at $349 though...
The 'front panel' software interests me. It looks like you can build a logic analyizer right into your project.
One of these days, I'll be able to afford it. :-)
-
For $99 I still like the Spartan 3 starter kit. It comes with a lot for the price.
http://www.xilinx.com/xlnx/xebiz/designResources/ip_product_details.jsp?key=DO-SPAR3-DK
Something that really has not been mentioned is the fact that the newer Freescale ColdFire processors (V4e ColdFire core) MCF5475 etc .. are supposed to be very 68K compatible (except for the FPU) and are very fast (200 to 266Mhz). If the FPGA Verilog source code to Minimig gets released to public via GPL or other method, Dennis's FPGA design could be modified to interface to a Freescale ColdFire CPU (V4e ColdFire core) such as the MCF5475 which has an on chip DDR/SDRAM controller, a PCI host controller, and Ethernet. This project could evolve into a super fast 68K Amiga that did not use PowerPC 68K emulation !!
:-)
-
MegaBlast wrote:
Just as long as the bitmaps range of titles work on it, that is all that matters :)
Such good taste. :pint: On your first post too. Good to see new members and all that. :-D
-
I think all of us are aware of the lack of AGA info available but what about Akiko? Or even the so called AAA? Was AGA's replacement ever finished? If so, are those plans available? and how AGA compatable if available?
If the plans are available and work with AGA then could those plans be usefull?
-
I'm skeptical about this "lack of AGA info" that several people have mentioned. What exactly is lacking? It seems to be emulated well enough in UAE. For the purposes of adding AGA functionality to an existing OCS implementation in FPGA (when that point is reached) I can't imagine what other kind of information would be needed. AGA is mostly an incremental improvement over OCS, after all.
I have seen some general information online about AAA but I don't think it was register compatible with AGA.
-
I must admit it puzzles me too. :getmad: but I'm not a coder yet (give me time and that will change)so I have to presume others know better than me. :roll:
Having said that, if it can be done in software I just can't see why it can't be done in hardware.
I've stated before on this thread that AGA is a minimum for me, If the minimig gets AGA (or RTG) I'll buy it.
-
Wow, extremely worthwhile project.
I have come from the land of Atari to check this out (news travels :-) )
I've never owned an Amiga before, but a new "mini" amiga like this would tempt me an awful lot.
Obviously we'd love something like this for our beloved platform, but that's our problem ;-) Major congratulations to you Dennis, and by extension to the Amiga scene.
keep the spirit alive!
-
You know what excites me the most? This whole project gets us all one step closer to the allusive Amiga Laptop. Even in OCS/ECS, this would be an amazing feat.
-
by Dr_Righteous:
You know what excites me the most? This whole project gets us all one step closer to the allusive Amiga Laptop. Even in OCS/ECS, this would be an amazing feat.
I have that, more or less. A $100 Compaq e500 running WinUAE. :-P
It's actually quite good, esp when running AIAB. With Picasso gfx and very fast '020 emulation, its faster and prettier than my real A1200 ever was.
Worth considering...
-
Tripitaka wrote:
I must admit it puzzles me too. :getmad: but I'm not a coder yet (give me time and that will change)so I have to presume others know better than me. :roll:
Having said that, if it can be done in software I just can't see why it can't be done in hardware.
I've stated before on this thread that AGA is a minimum for me, If the minimig gets AGA (or RTG) I'll buy it.
I don`t think he cares if anyone buys it, I don`t think he`s even trying to sell it...
-
off topic, but it case you didn't see an earlier post
www.fpgaarcade.com
Atari ST FPGA development work going on.
sound/shifter/glue/mmu fpga design complete. I need to find time to build a board to test everything ...
-
Projects on www.fpgaarcade.com appear to be some pretty good stuff. The FPGA Colecovision conversion looks pretty interesting as well. You mentioned Amiga AGA on fpgaarcade site. Are you going to work on Amiga AGA or just Atari ST chipset ?
Keep up the good work !!
:-)
-
No, a Compaq laptop, or any other PC laptop running WinUAE is not worth considering to me. I have that capability already. I'm a hardware junkie. Software sucks.
-
Frags wrote:
[/quote]
I don`t think he cares if anyone buys it, I don`t think he`s even trying to sell it...[/quote]
Hehe, I think he needs an agent. :-D
-
Dr_Righteous wrote:
No, a Compaq laptop, or any other PC laptop running WinUAE is not worth considering to me. I have that capability already. I'm a hardware junkie. Software sucks.
Perhaps a little strong but I don't like WinUAE much either. I guess it's a matter of taste.
As for the "hardware/software is better" argument, the way I see it, it's all about balance. Too much reliance on hardware fills up space PDQ, too much software running on the same machine complicates maintanence. Finding your balance between extremes is a very individual thing. A gifted coder in single bedroom acommodation may well prefer a software approach, you prefer hardware. Choice and balance. Dennis's minimig is cool as it uses a spare drive bay. A space most of us have on at least one machine. :-D
-
Tripitaka wrote:
Dr_Righteous wrote:
No, a Compaq laptop, or any other PC laptop running WinUAE is not worth considering to me. I have that capability already. I'm a hardware junkie. Software sucks.
Perhaps a little strong but I don't like WinUAE much either. I guess it's a matter of taste.
As for the "hardware/software is better" argument, the way I see it, it's all about balance. Too much reliance on hardware fills up space PDQ, too much software running on the same machine complicates maintanence. Finding your balance between extremes is a very individual thing. A gifted coder in single bedroom acommodation may well prefer a software approach, you prefer hardware. Choice and balance. Dennis's minimig is cool as it uses a spare drive bay. A space most of us have on at least one machine. :-D
Very Zen. :-)
-
I did a detailed study of AGA and decided it was possible to do it in a reasonable amount of time. The Atari stuff is a lot easier so I am going to finish that off first. At that point I will look at it again. If the Minimig source becomes available then maybe we can do something with it.
/Mike
-
Mikej! If you remember I have mailed you already a some time ago. I keep crossing my fingers for if you can finish AGA in an fpga! Looks like i have to cross my finger for atari st at first than for Amiga =). The only thing is keeping amiga's still future is aga if any one could do it fpga than this is the way out to get really new hardware. Thanks a lot to you both guyz (Mikej and Dennis)!!! You givin' us some hope so far!
-
mikej wrote:
I did a detailed study of AGA and decided it was possible to do it in a reasonable amount of time. The Atari stuff is a lot easier so I am going to finish that off first. At that point I will look at it again. If the Minimig source becomes available then maybe we can do something with it.
Nice work on the Atari chipset, maybe it can run on my final Minimig board as an alternate core. (Would be cool, dual boot Amiga/Atari. Together at last.. :-) )
On your site you said that you have programmed demo's for the A1200, that will help when implementing AGA, as you probably know about all the weird, undocumented 'features'. I have never pogrammed the Amiga on that level. To be honest, implementing Minimig took much more time than I had anticipated. I think I have spent about 600Hrs on it in total, mainly on sundays and workday evenings.
Are you planning on a circuit level replica? Minimig is not. I figured that would be too much work, besides I don't have a logic analyzer. :-(
Anyway, looking forward to the ST core.
Dennis
-
Update,
I have fixed the blitter timing for 90%. It was a stupid bug in my blitter nasty handling. Also, I found some clues in the Winuae code which really helped. I am now (again) focussing my attention on the PIC code to clean up the floppy selection user interface.
Any chance you could test Minix 1.5 for Amiga on your setup?
I don't think it will run on 512K.
Also, how much has the project cost you so far?
Hmmm, well. I figured out very soon in the project that synthesizing verilog code can take a loooong time if you don't have a fast computer, so I actually bough a shiny new PC :-D . Besides that I have spent about a couple of hundred euro's.
Any idea on a cost, approx?
Not really, it all depends on large quantity pricing. So I can't give a real price.
What about emulating 68k? It could make the design easier and better for commercial (or not) manufacturing
Synthesizing an 68000 and ironing out all the bugs would be a project on it's own. I know I am not going to do it. At least not in the near future.
Dennis
-
Hi Dennis, good to talk to you.
As you say, the biggest problem with this sort of work is the amount of time it takes - if you were to charge on a standard hourly basis for this it would not be feasible. I thought the PacMan project would take a few weeks, in the end it took over six months - most of which was finding very subtle bugs in the cpu core.
Because of all the overscan tricks etc used in the Atari, I want it to be cycle exact to the 8MHz system clock. One trick I found with some of the other projects was to get one chip working at a time in a real system, preferably comparing the fpga core with the real chip using the analyser. I am backing off this slightly as it is seriously time consuming and you get into slightly wobbly behaviour with the whole thing becoming unstable.
The other problem is moving over to a single syncronous clock domain - I am running the ST chipset at 32MHz, with everything clock enabled at the correct slower clock speed. This may have to change if I can't get the 68K cpu core (assuming it is ever finished) to run this fast. So, I guess it is a circuit level replica, but tidied up a bit :)
I need to get myself some ram /flash on my 68K virtex1 board to run it for real, or bolt a low voltage 68K onto my virtex4. What cpu are you using ? 3v3 io on it ? Any major differences with the original 68K ?
It would be great to get both designs running together. Contact me directly at www.fpgaarcade.com if you wish.
Cheers,
Mike.
-
by Dr_Righteous
No, a Compaq laptop, or any other PC laptop running WinUAE is not worth considering to me. I have that capability already. I'm a hardware junkie. Software sucks.
I understand your pov, I used to be a hardware-only nut too. Ive had a room with an A1200 and an A500, a lowboy Arcade machine with 3 PCBs and an A500-Jamma hybrid, 3 consoles all hooked up to 3 monitors/TVs, a draw full of handhelds; Lynx, GBcolour, Gamegear and all the associated crap, games, carts, adapters, controllers...
At some point it gets too much.
Now, I could care less. If Ive got a PC/laptop sitting there in front of me, I may well as load it up with emulation software and save the hassle and clutter of real hardware.
Yeah, its nice to have the knowledge that there's a "real" 680x0 doing the work, but not nice enough to go back to that.
Having said that, I'd buy an A500-in-a-joystick if it ever came into existence, and I'd probably have it wired up to a 5" LCD within a week.
XX fingers crossed XX
-
The thing about emulation is, if you only ever had an A1200 with 030 or something like that, WinUAE is miles better.
But even though WinUAE on my 2GHz laptop is theoretically faster, it feels sluggish compared to my A4000 with CSPPC and Voodoo 3. I'm guessing its because Picasso96 performance is slower... it definately doesn't feel quite "right".
Mind you, watching AGA demos and playing games on the laptop doesn't feel right either; but i'm guessing if I used the TV out on the laptop and used frame syncing it'd be a lot nicer ;-)
-
If both Amiga and ST cores were on this one board as Dennis speculated, I'd wet my pants! and then I'd order one :-p
Life would seem just a little less cruel :)
@ MikeJ
i'm sorry to be presumptious, this is your project... but i was just wondering, do you think your ST core could be connected to a real 68000, as Dennis is doing on the minimig? The only reason i suggest this is that if the minimig has a 68k on board, maybe you could (just for this project) focus on the ST chipset and save time on emulating the 68k, and we'd have a dual Atari/Amiga board! Maybe that wouldn't take as long to do.
And of course, you could then work on a "super minimig" or "minimig 2", where your completed FPGA 68000 is included and the real 68000 is left off, making the board even smaller, and the ability to be faster. This would give you time to work on the AGA chipset too, which you said you had studied. Dennis may even help, who knows :)
Just some humble speculation. Good luck!
-
MikeJ wrote:
I need to get myself some ram /flash on my 68K virtex1 board to run it for real, or bolt a low voltage 68K onto my virtex4. What cpu are you using ? 3v3 io on it ? Any major differences with the original 68K ?
I am using the Freescale MC68SEC000. This cpu runs at 3.3V at has 3.3V I/O. Differences from the original are:
No 'E' clock, you have to simulate 'E' clock timing yourself.
MOVE sr, is privilidged and there is no real replacement instruction. I have not encountered a program yet that uses this instruction, so that is not a real problem.
Dennis
-
"do you think your ST core could be connected to a real 68000, as Dennis is doing on the minimig?"
Yes, that was I was thinking about in the previous post - to save time run with a real 68K. I could use the simple board I was playing with the 68K (you can see this on the fpgaarcade site), but I need to add some RAM and FLASH to get it working.
Dennis, thanks for the info (it's the one I thought you would be using). I will get hold of one and apply it to the spartan3 board.
Cheers,
Mike.
-
I think it needs an x86 accelerator on board, in addition to the 680x0. Nothing fancy, just a 1Ghz athlon or so. We could write a x86 Powerup/Warpup type system... x86 datatypes, drool...
Or not. Cool on it's own, but with the athlon it'd be revolutionary :-)
-
I think it needs an x86 accelerator on board, in addition to the 680x0. Nothing fancy, just a 1Ghz athlon or so. We could write a x86 Powerup/Warpup type system... x86 datatypes, drool...
[DREAM MODE]
I was thinking more along the lines of a VIA EPIA mini-itx board plus a PCI-card with Minimig on it. The PCI card gives you video out, joystick ports and OCS chipset. You could then run AROS on it. If the EPIA is defined as the defacto standard, you only need to write drivers for one platform, namely the EPIA mini-itx board. Besides, there are a lot of cool cases to be found for mini-itx.
[/DREAM MODE]
But this is just a dream, do not take me seriously on this
:-)
Dennis
-
Would be nice!
I still can't understand why the Amiga One wasn't based on a standard EPIA mini-ITX board... Would have made so much more sense than the proprietary, expensive and now non-existant PPC-based "solution" that was chosen instead...
- Ali
-
Dennis wrote:
[DREAM MODE]
I was thinking more along the lines of a VIA EPIA mini-itx board plus a PCI-card with Minimig on it. The PCI card gives you video out, joystick ports and OCS chipset. You could then run AROS on it. If the EPIA is defined as the defacto standard, you only need to write drivers for one platform, namely the EPIA mini-itx board. Besides, there are a lot of cool cases to be found for mini-itx.
[/DREAM MODE]
But this is just a dream, do not take me seriously on this
:-)
Dennis
My company already uses Via EPIAs for our embedded products I might be able to convince them to produce some Amiga clones for you! :-) If you could get the PCI cards made of course!
Martyn
-
InTheSand wrote:
Would be nice!
I still can't understand why the Amiga One wasn't based on a standard EPIA mini-ITX board... Would have made so much more sense than the proprietary, expensive and now non-existant PPC-based "solution" that was chosen instead...
- Ali
Primarily I suppose because then people could just go out and buy their boards wherever they wanted, rather that being forced to buy only from Licensees, and so Amiga Inc would lose the License fees they got on every board sold.
-
[DREAM MODE]
I was thinking more along the lines of a VIA EPIA mini-itx board plus a PCI-card with Minimig on it. The PCI card gives you video out, joystick ports and OCS chipset.
What would you think about the idea of having the card feed the videodata to the PCI-display controller of the EPIA-PC? I mean in the same way as TV-tuner-cards do. (PCI-tuner -> PCI-display controller.) You would get a flicker fixer "for free"?
May I ask if there will be Paula floppy disk drive controller implemented?
-
What is the cooling system for Minimig?
Is it heatsink, fan or none?
What is the power supply specification?
Is it ATX, hard disk power connector, +12V or other?
> VIA
VIA is famous for defective, low quality GSC capacitors.
The capacitors will leak because of heat problem.
You should use low speed fanless EPIA products, not the 1 GHz products.
-
Hi God, oh, I mean Dennis,
This is AWESOME.
I haven't read the whole thread, but was wondering (hope it hasen't been discussed yet), what is the chance of going past the 2 Megs Chip ram limit/curse of Amiga?
Also, can an FPGA be added, that could emulate a 68000, and a second one to do a 68020, that is much faster than any that have been made to date?
Can an FPGA be added that emulates a 68060, with the full math library in it?
I am definitely interested in this, and, it would be a knockout to get AGA (>2 MB limit) and/or PCI card it!!!!!!
-
Implementing an 68000 requires a lot of work - an '020+ much more. An '060+ is probably not possible with current FPGAs.
It'd be easier to use a PPC with JIT emulation or - the elegant way - a Crusoe with the proper code morphing software. Or a Coldfire V4e/V5...
However, this leads to exotic mainboards that would need development, and nobody will do that.
-
Zac67 wrote:
Implementing an 68000 requires a lot of work - an '020+ much more. An '060+ is probably not possible with current FPGAs.
It'd be easier to use a PPC with JIT emulation or - the elegant way - a Crusoe with the proper code morphing software. Or a Coldfire V4e/V5...
However, this leads to exotic mainboards that would need development, and nobody will do that.
Exactly a minimig FPGA on a PCI board would be the best solution. Then the user can decide which CPU to use.
-
Whoa...Just finished reading this entire thread and my it sounds like an awesome project, I wish you all the best with it Dennis :-)
-
Seems like dennis went with the best solution for his project. use a cpu thats already cheaply avaiable, do all the other stuff in fpga. Why impliment cpu in fpga from scratch when you could just buy one?
@dennis
Haha, you made hackaday.com. Your project blows the water out of 95% of what pops up on there in terms of technical geekness. Your hardwork is an inspriation to us all.
-
I wonder...
Amiga OS4 seems to be able to use the custom chips if present, one AmigaONE with Minimig could be (made) as if they were again available, running games, demos as if in the real thing... :roll:
-
@bloodline
Exactly a minimig FPGA on a PCI board would be the best solution. Then the user can decide which CPU to use.
The force is strong with you... :-D
-
Exactly a minimig FPGA on a PCI board would be the best solution. Then the user can decide which CPU to use.
Actually, I think this is a bad idea, since you NEED standardisation. Standardisation is what made the original Amiga so wonderful to work with, and the lack of it is what made the PC such a pain to work with.
-
@Dennis
Thanks for your regular updates - they are appreciated by many, many people.
I had a couple of thoughts - first off;
You and Commodore share something in common! You're both from The Netherlands! :-) If I were you I might email them with your progress and photos - they have shown themselves to be interested in things Amiga and even have a link to a Dutch Amiga site from their website!
Secondly....
HDL Synthesis... if you wrote a book explaining your project and M.O. from start to finish you would sell A LOT of books. The books out there are really crap for $70-$100.
-
Dessimat0r wrote:
Exactly a minimig FPGA on a PCI board would be the best solution. Then the user can decide which CPU to use.
Actually, I think this is a bad idea, since you NEED standardisation. Standardisation is what made the original Amiga so wonderful to work with, and the lack of it is what made the PC such a pain to work with.
You could first start by using those platforms already covered by AROS, then the world who knows...
:-)
-
Zac67 wrote:
Implementing an 68000 requires a lot of work - an '020+ much more. An '060+ is probably not possible with current FPGAs.
It'd be easier to use a PPC with JIT emulation or - the elegant way - a Crusoe with the proper code morphing software. Or a Coldfire V4e/V5...
However, this leads to exotic mainboards that would need development, and nobody will do that.
Hi Zac67,
Okay...
/Dream mode on
2 FPGAs.
1 does OCS/ECS (SW switchable), standard speed. (With, higher than 2 Megs chip ram, SW switchable, to .5, 1, 2, 256 Megs chip ram.)
2nd FPGA, does as high speed as is possible for OCS/ECS, and 2 Megs chip ram, SW switchable to .5, 1, 2, 256 Megs chip ram.
3 CPUs
1 standard 7.12 MHz 68000
1 68030 50 MHz or 68040 50 MHz or 68060 66 MHz
1 Dragon ball (was the max 66 MHz?) fully 68000 compatible CPU
+ a flicker fixer circuit
On a PCI card and/or in a PDA!!!!!!
/Dream mode off
Whew!
-
Lando wrote:
InTheSand wrote:
Would be nice!
I still can't understand why the Amiga One wasn't based on a standard EPIA mini-ITX board... Would have made so much more sense than the proprietary, expensive and now non-existant PPC-based "solution" that was chosen instead...
- Ali
Primarily I suppose because then people could just go out and buy their boards wherever they wanted, rather that being forced to buy only from Licensees, and so Amiga Inc would lose the License fees they got on every board sold.
Perhaps Amiga Inc could have made a security dongle of some sort and restricted OS4 sales to be bundled with the mini-ITX hardware only?
- Ali
-
Atheist wrote:
Okay...
/Dream mode on
...snip...
/Dream mode off
Whew!
(http://www.itcall.com/exf/images/em/anne.png)
-
@Karlos
:lol: :lol: :lol:
-
Hi Dennis,
Great work, the Amiga was one of the best computers ever - great job.
I'm really interested in building one of these myself :)
Would you recommend the...
http://www.xilinx.com/xlnx/xebiz/designResources/ip_product_details.jsp?key=HW-SPAR3-CPLD-DK
or should I wait for the...
"The Spartan-3E Starter Kit target availability is February, 2006."
Also, will you be making the source code for the FPGA, PIC and the various circuit board layouts available?
Thanks,
Anthony
-
@Atheist:
FPGA 'code' is loaded on powerup, so there's no real need for a 2nd FPGA chip, as long as you can fit everything into one - you just reload the one you have. :-)
-
For those that are interested in this kind of thing, Xilinx are GIVING AWAY FREE evaluation boards for the Spartan 3E :
http://www.xilinx.com/products/boards/s3esamplepack/index.htm
This is the updated version of the Spartan 3 FPGA used in this project (but the evaluation version is the smallest version).
-Anthony
-
I've just picked up one of these - currently fitting pacman into it for a quick giggle. I have ordered one of the larger 3E starter kits, but they still seem to have supply issues.
/Mike
www.fpgaarcade.com
-
Just look at this beauty!
(http://commodore-gg.hobby.nl/minimig.jpg)
-
pixie wrote:
Just look at this beauty!
(http://commodore-gg.hobby.nl/minimig.jpg)
:-)
-
I signed up just because of this thread.
Now this was one LONG and INTERESTING read.
Whats they say in "Field of Dreams" ? If you build it, they will come !
Now for the questions......
Having built it to read games from ADF, how difficult would it be to incorporate support for reading IPF images, now that there are a LARGE number of Images out now.
I am sure the CAPS/SPS team could help.
I only see the problems being in SOME of the timings, as the IPF's are LITERALLY images of the ORIGINAL disks. That means that the copyprotection scheme's are present, something which is not present in ADF's.
Is there a plan for increasing RAM for the unit from 512k to 1024k, I understand that this is going to mean the core usage on the FPGA will grow, but for games like Pinball Dreams/Fantasies etc, 1MB of ram *DOES* have its benefits.
Probably not 1MB CHIP, personally i Always liked the 512Chip/512fast personally.
But any way to add more memory would be good :-)
To sum it up, I am overwhelmingly impressed, and hey... I WANT ONE !!!!!!!!!!!!!!!!!!!! :)
I have BASIC soldering skills, so even a DIY thingy for us amateur solderers.
Respect.
Brian.
[edit] oh, i read about the scandoubler......., I primarily use LCD, which run at Horizontal 37.92, and Vertical of 60.62 (at 800x600, cant get 640x480 from what i can see).
The manufacturer's website says Horizontal range 31-63, Vertical range 56-75.
What do ya reckon ?
Also, If using the card thingy to load ADF's to play the games, how difficult would it be to make the card act like a Harddrive ? And could it be upgraded to use Kickstart 2.04+ for the Autoboot facility that 1.3 did not have ?
-
There is a cancelled plan to create Pensa 2 FPGA based CPU from Starbridge:
"StarBridge has developed what it calls Hyper-Specificity cores that can perform a mix of scalar, vector and DSP applications, depending on how Viva synthesizes the algorithms. When the operations are instantiated on Xilinx 4062-8 FPGAs, the resulting processor chips are called Pensa processors. StarBridge has implemented 10 Xilinx FPGAs on the central board in its HAL-15 system: eight for general-purpose Pensa processing, and two dedicated to I/O operations."
Is it possible to use Minimig with the Pensa/Starbridge computer?
Perhaps the system can be used as a remote game server for multiple PC in a game center?
Is it possible to buy license for X86 PC IP core, combine it with Minimig+68K core and create a switchable multi architecture system?
http://www.starbridgesystems.com/
http://www.nallatech.com/
http://www.annapmicro.com/
-
I think most of Amigans would flip knowing that it would run with something resembling a x86 :roll:
-
Having built it to read games from ADF, how difficult would it be to incorporate support for reading IPF images, now that there are a LARGE number of Images out now.
Haven't really looked into that, getting it to run ADF's was quite a challenge already.
Is there a plan for increasing RAM for the unit from 512k to 1024k, I understand that this is going to mean the core usage on the FPGA will grow, but for games like Pinball Dreams/Fantasies etc, 1MB of ram *DOES* have its benefits.
The RAM is external to the FPGA. The only reason I have only 512K at the moment is that my development board hasn't got any more. When I move the design to my own board, I will put more RAM in it.
Also, If using the card thingy to load ADF's to play the games, how difficult would it be to make the card act like a Harddrive ? And could it be upgraded to use Kickstart 2.04+ for the Autoboot facility that 1.3 did not have ?
Harddrive emulation is high on my wishlist when I have my own board running. The current plan is to have a second (internal) MMC slot which acts like a harddrive. Minimig runs kickstart 2.04, 3.0 and 3.1 so autobooting should work.
I haven't read the whole thread, but was wondering (hope it hasen't been discussed yet), what is the chance of going past the 2 Megs Chip ram limit/curse of Amiga?
Hardware-wise, it is no problem whatsoever. The problem is software. AFAIK, Amiga-OS only recognizes the lower 2MB as chipram.
Would you recommend the...
I don't know. I have the spartan-3 starterkit (400000 gate version) and it is very nice, the only drawback I can think of is lack of extra ground pins on the IO headers.
Dennis
-
I am still overwhelmed at this way this project is coming together, and to think of the negativity that was being displayed earlier in the thread.
If everyone listened to the negativity when Amiga was originally on the drawing board, we wouldnt have had the amiga.
People have to remember this is only still in prototype stage, the possabilities are HUGE.
Sure we could go all the way to a new ATX motherboard (or PCI card), with the FPGA as its core, but where does it end ?, Dennis probably has his own ideas where he wants the project to go, and its great we still have people with the knowledge to bring new life.
Now to Open source..., sure its good to have multiple heads on the project, with differing talents to make the project go further, but again, where does it end ?, do we really need an open source project like LINUX ?, so many differing versions, and some versions are not completely compatible with eachother.
What would be good, is the project available to many people for further development, but One group or person(s)doing the final release, that way you do not have lots of partly finished projects or lots of partly compatible side projects.
I cant attend the amiga conference that dennis will be bringing the project to (Different continent, but i am keen to see pics/videos of the project in Action.
Actually, Dennis..., have you tried running a game other games, like Shadow of the beast and see how the parallax scrolling copes ?
I watch this thread with eagerness :)
-
@Dennis:
Hardware-wise, it is no problem whatsoever. The problem is software. AFAIK, Amiga-OS only recognizes the lower 2MB as chipram.
I can trow 8Mb in WinUAE without a problem and it seems to recognize...
-
Dennis wrote:
Harddrive emulation is high on my wishlist when I have my own board running. The current plan is to have a second (internal) MMC slot which acts like a harddrive. Minimig runs kickstart 2.04, 3.0 and 3.1 so autobooting should work.
I was hoping to see a feature like that creep in. It's strange to think that my first hard drive was 128mb and that now it might become possible to stick a small 1GB card in this for a fraction of the cost. :)
I hope the people who came out with the C64 stick notice this project of yours and perhaps fund you in some way for your hard work to date. This could be a great little "toy" for Christmas this year if bundled with some licenced games and could easily sell as many units as that DTV-C64 (or whatever it's called - and I have one of those).
Does the idea of seeing the Minimig mass produced and sold on QVC appeal to you? ;)
-
@whiteb
The reason there was so much negativity is because many Amiga users have been pulled by the heartstrings for so long, they tend to take any type of news of any "new" developments with a pillar of salt.
I still remember back in 1998 hearing about a new system, then all the talk about software. Back in the days of the Amiga being in development there were no "false promises", "lies", "shallow dreams". It was more like a "yeah you must be dreaming" instead of at "get your fake ****** out of here". That's just what happens when so many things go wrong.
I myself know how that is , because me and some other people were working on a Sega Saturn emulator back when Sega decided to just pull the plug on the system while the heart was still beating. That mad a lot of people :pissed: . What made it even worse were the false promises of software to keep coming out for the Saturn. Developers for the US market dumped the system like a bucktooth, piggytail girl on a date.
Enter our team. We decided to figure out ways to try to emulator two processors and get them to work with the other systems ( 2 sh-2, a sh-1, 68000 cd controller). We working on a design for the flow of emulation, and one of the guys ended up getting evolved with this girl two weeks in the process (that just wanted him for his money) that seemed to convince him to get drunk and high everyday.
Not even a week later he ups and leaves us. The other guy was his friend, he stopped coming by the coffee shop that we met at, and I was forced to cancel the project. People accused me of not having a project at all.
I think I still have the source code to the stuff on a dead Western Digital hard drive somewhere. I would pull it off it if the thing would power up. But now we have other Sega Saturn emulators now and it's expensive to use data recovery places.
Remember the moral of the story kiddies..
Once bitten, twice shy.
-
My Spartan 3E sample pack arrived today (thanks to the nice PR people at Xilinx UK). Apparently is normally only available to the US???
Anyhow, I'll be ordering a Spartan 3E Starter Kit ASAP.
Dennis: Will you be releasing source code to your minimig?
I was thinking about using your kit to do the GFX/Sound and use an ARM to emulate the 68K. A fairly fast ARM should be able to emulate an 8Mhz 68K in software without any problems. You could have an 8Mhz emulation and a "as fast as you can" emulation.
That way we could have a "native" ARM mode to run super fast, an unused 68K instruction could be used to switch off emulation. A small ARM OS could hold the emulator and be used to return to 68K "mode".
An ARM cpu should be quite cheap and they usually have quite a lot of functions built-in. Some have IDE and/or Flash memory controllers and could also take the place of the PIC.
What do you think?
-Anthony
-
@pixie
I can trow 8Mb in WinUAE without a problem and it seems to recognize...
That is only because WinUAE hacks it. ROM only detects upto 2MB.
-
Couldn't the same patch could be implemented on the fpga version? Because UAE is opensource the patch should be easy to find and add to the fpga versions rom after its loaded.
-
DoubleDutc wrote:
Couldn't the same patch could be implemented on the fpga version? Because UAE is opensource the patch should be easy to find and add to the fpga versions rom after its loaded.
Try setting 8megs in UAE, and then booting with 1.3. It will not see the 8megs. You need to to patch the software, that's easy for software, but probably not so for hardware.
-
pixie wrote:
I can trow 8Mb in WinUAE without a problem and it seems to recognize...
Where does UAE map the extra 6Mb?
I just had a quick look at the A500's memory map, the only place I could see it fitting would be in the 8MB space meant for Autoconfig /Fast Mem..which just so happens to be right after the 2MB chip mem area, but then that means you're limiting yourself to 2MB of space to add any fast mem unless you change to a CPU with a 32bit address bus :-(
Edit:
Just had a look at UAE, the extra 6mb of chip was in at $200000 right after the normal 2mb chip.
Strange thing was it was in 2 different nodes, with only MEMF_PUBLIC | MEMF_CHIP set, instead of MEMF_PUBLIC|MEMF_CHIP|MEMF_LOCAL|MEMF_24BITDMA|MEMF_KICK that the normal chip mem has..but I guess this is just UAE adding to the memory list instead of patching the function in exec that searches for the upper limit of chip mem.
Anyway, a BlizKick patch should be pretty easy to sort that out if Dennis wants to go for more chip mem. :idea:
-
As long as you can get the 'chip set' to work with >2 MB, you can flag the mem as chip RAM any way you like - a small tool changing the mem flags or a resident module doing it or even a patch to the ROM, which is loaded from flash anyway.
As long as the memory can be addressed for graphics etc. there's no point in flagging it as fast mem - it doesn't make it faster.
Many, many years ago I wrote a small system patch 'MakeChip2Fast' that additionally flagged chip RAM as fast RAM - for a friend of mine with an A1200 who couldn't run a certain program that required fast. Worked as a charm and Amy was displaying 2 MB chip AND 2 MB fast -> 4 MB total. :lol:
-
For anything that's emulating an OCS machine I'd be happy with a system that included 2MB chip memory and 2MB fast memory. Nothing more is required to be honest - the only game that required 4MB was Sim City 2000, and that was AGA anyway.
If you were designing a cut down version for a 'tv game' type system then you'd have 1MB in total (512/512) and say a 16MB flash chip containing enough ROM to boot a floppy image, and some 18 floppy images, allowing 9 two-disk games. You'd want the 1MB for the games that actually made use of it.
Now what games would you want on such a device? Hmm ... Speedball 2. Robocod. Zool. Xenon 2...
-
I was thinking about using your kit to do the GFX/Sound and use an ARM to emulate the 68K. A fairly fast ARM should be able to emulate an 8Mhz 68K in software without any problems. You could have an 8Mhz emulation and a "as fast as you can" emulation.
If you put the arm into the FPGA its Ok. But as external Part it is not reasonable. Then you connect better the 68000. What about the Timing of an ARM Emulation?
I will put an 68000 Emulation into the FPGA. But I have not cycleexact emulation.
Currently needs my own RISC 1200LC's of a Cyclone FPGA and 128KB RAM for the EMUCode. The most 68K Opcode work now but not all. I hope finish my work in the next 4 weeks.
Viele Grüße
TobiFlex
-
If you put the arm into the FPGA its Ok. But as external Part it is not reasonable. Then you connect better the 68000. What about the Timing of an ARM Emulation?
The ARM would be external, it would emulate the 68k thru software. As long as its near as possible to 8Mhz I don't see there being much of a problem. Don't forget that there were different CPU speed Amigas out there. As long as the other hardware is as similar as possible it should work for 90% of games, etc.
The other benefit of having the ARM external is that it doesn't use valuable FPGA space, this could be better used to enhance the original chipset. As I said before, there could be a native ARM mode to speed up some tasks.
-Anthony
-
Hi All,
Thank you for talking about the 2 MB chip ram barrier.
Well, as I recall, and being no expert, in AOS1.3, when you went from Agnus (512 K Chip), to Fat Agnus (1 MB Chip), and then to MegaChip 2000 (2 MB chip), no SW was involved at all. It just "worked".
Now, as the 68000 was a 16/32 bit CPU, is it that 8 Megs of fast ram, and then a max of 2 MB chip ram, and then the rest of the 16 Megs of the 68000's physical capability, which is 16-10= 6 Megs, was reserved to address adresses for all of the Zorro bus' I/O?
Is that why only 2 Megs of chip ram is possible?
Is the 68000 only able to address a 16 Megabyte address space?
Could the Chip ram go to at least 4 Megs, and leave 4 Megs for the Zorro bus??
I do go on about this as it is significant to demo coders, and what they can achieve without this barrier, or at least some more breathing space!!!
Could the chip and fast ram be flipped? 8 chip, and 2 fast?
It's real easy (sarcasm), just figure out how the Agnus changed from 512 K to 1 M to 2 Megs. Duplicate.
Amiga Rooollz!
-
Well, as I recall, and being no expert, in AOS1.3, when you went from Agnus (512 K Chip), to Fat Agnus (1 MB Chip), and then to MegaChip 2000 (2 MB chip), no SW was involved at all. It just "worked".
No software, but the Motherboard had to be hacked (traces cut etc).
I have a 4MB trapdoor expansion for my A500 (4x1MB PC Simms), I have a Fat Agnus (1MB CHIP), and i had to cut a trace to help with the 1MB chip..., 2MB chip had like 5 traces to be cut from memory, and i didnt want to go there.
Now, as the 68000 was a 16/32 bit CPU, is it that 8 Megs of fast ram, and then a max of 2 MB chip ram, and then the rest of the 16 Megs of the 68000's physical capability, which is 16-10= 6 Megs, was reserved to address adresses for all of the Zorro bus' I/O?
Well, My A500, coupled with my A530, and the 4MB trapdoor expansion, i have 12.5MB of ram available (1MB chip, 11.5MB fast). A530 has the full 8MB.
-
> ARM - 68K emulator etc
Is it possible to combine Minimig with Omega Northbridge?
(switchable system)
Project Omega:
FPGA Northbridge for StrongArm and emulate 24 bit Acorn RiscPC:
"The base model Omega uses a 300 Mhz Intel Strong ARM processor and the normal Southbridge 'Super Combo' chip.
The Acorn MEMC and VIDC are replaced by the MicroDigital Lightning© chipset and the standard Northbridge chip is replaced by a special MicroDigital designed version on an FPGA. Similarly the System chip is a CPLG device. As you can see, the main feature of this design is that instead of the usual interfacing logic most of the work is done by Complex programmable Logic Devices (CPLG) or Field Programmable Gate Arrays (FPGA) CPLDs and FPGAs can be considered as a sort of 'raw' circuit block."
-
Atheist wrote:
Is that why only 2 Megs of chip ram is possible?
Nope, it's cos the existing chipsets don't have a wide enough address bus to be able to read/write to any address bigger than 2Mb.
Is the 68000 only able to address a 16 Megabyte address space?
Yup, and the same is true for some other CPUs too, such as the 68EC020 inside the A1200.
Does anyone here have any experience with Freescales MC68306 (http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MC68306&nodeId=0162468rH3YTLC61650795)?
Although it only has the usual 24bit address bus, can it's internal DRAM controller be used with memory mapped into Z3 space ?
It looks like it could be a good replacement for the normal 68k in MiniMig, adding a few features like fast ram and another/faster serial port that would normally take a lot more work to add on.
-
Actually a 68000 could address more than 16 MB, when paired with an MMU. This being pretty painful, the 68010 was made to overcome the problem. Also by using memory mapping techniques you could extend the barrier.
Guys, this is getting very OT, you might consider opening another thread...
-
Actually the Freescale MCF547x/MCF548x series would be a good starting point for new hardware. These are Coldfire V4e cores with MMU and FPU. Supports DDR I memory, USB2.0, 10/100 ethernet, PCI.
MCF5474 266MHz Retail pricing: $38.74 at qty 1 (digikey.com)
The only draw back is 388 pin BGA package. It would likely need 4 layer PCB with plate through holes for mounting.
BTW: Dragon 1200 accelerator - MCF5475 (5474 with encryption)
http://dragon.amigaworld.de/index.php?lang=en&page=2
-
About CPUs:
http://www.logicpd.com make FireEngine modules, that have MCF547x (or MCF548x with CAN.) The prices of these boards are, of course, not very cheap. The budgetary pricing on Freescale's site is several hundred $ in 1000+ quantities.
The MCF5407 is in PQFP 208 package, that could be hand soldered, but it requires 1.8V core and 3.3V IO voltages, which suggest at least 4-layer PCB. 2-Layer boards are possible to make or much cheaper to have made. Or perhaps it would be possible to place a tin foil + an insulator under the PQFP package for the core voltage.
68SEC000 or perhaps 68HC000 seem manageable for a hobbyist. I was thinking of buying used Compaq LTE 5xxx series laptops, and removing the PC-parts, and then making a board with 68SEC000 and some Amiga compatible chipset. If Dennis' design will be available, then I'd use a Spartan3 or 3E, or Actel ProAsic3(if it is suitable.)
-
If Dennis' design will be available
Dennis: You haven't posted in a while, will you be making your design available (with pic and vhdl or verilog source code)?
-
I hope Dennis has a back up of his code in the hands of a reliable third party, it would be a double tragedy if something happened to Dennis and the minimig was lost forever!!!
-
Update:
I have been working a lot on blitter timing issues lately; I am now very close but it is still not perfect. Although everything now works without graphic errors, some demo's run much too slow. With games everything seems OK now. I have also discovered a bug with (I think) the sprite engine; Fire & Ice works like a charm except that the hero (the blue dog) isn't visible! OTOH, if I try this game on winfellow....
I will put an 68000 Emulation into the FPGA. But I have not cycleexact emulation.
Wow!, you run an emulator on your own CPU? How fast will it go?
Dennis: You haven't posted in a while, will you be making your design available (with pic and vhdl or verilog source code)?
See my earlier post in this thread, I still don't know as I am afraid someone grabs the code and makes it commercial . Since my first post I have got 7(!) offers to cooperate on commercialising Minimig. :-o
I hope Dennis has a back up of his code in the hands of a reliable third party, it would be a double tragedy if something happened to Dennis and the minimig was lost forever!!!
You are scaring me. Really! :nervous:
Dennis
-
Dennis: You haven't posted in a while, will you be making your design available (with pic and vhdl or verilog source code)?
See my earlier post in this thread, I still don't know as I am afraid someone grabs the code and makes it commercial . Since my first post I have got 7(!) offers to cooperate on commercialising Minimig. :-o
How about releasing some sample source code that will give people who are new to FPGAs some insight into getting Amiga quality video and sound.
For example you could release the sound code and video code to do a standard Amiga gfx mode on a tv - but no blitter, copper or sprites.
This way its useless to anyone who wants to rip you off, but gives beginners some useful info on how to generate good sound and get output to the rgb input of a tv.
?
-Anthony
-
@DoubleDutc
What? If you want to learn about VHDL or Verilog there are loads of websites, even several full examples of Arcade Machines. Dennis will release his work, when and if he see's fit to.
Sample source code would be useless to you, unless you understood RTL. It's parallel structures are unlike anything in C (perhaps with the exception of threads).
If you are truely interested in doing something like an Amiga in an FPGA then the first place to start is reading the Amiga Hardware Reference Manual. Then perhaps examin the clocking schemes and I/O of each chips so you have a framework with which to work, a top-level. Work through the registers within each chip, the memory space, buses. Then work on the state machines that link them together and finally concentrating on the timing of the statemachines and the I/O to bring everything together.
If you are asking dennis for details on the techniques he used to drive VGA pins out of an FPGA I am sure that he'd show you ohms law.
http://www.fpga4fun.com/PongGame.html
The sigma delta modulator, that however is a very clever beast indeed. I was very impressed when Dennis said he'd developed one instead of using a regular parallel DAC. Just shows his maturity considering this isnt his full time occupation. Point you in the direction of things like this he might.
http://www.fpga4fun.com/PWM_DAC.html
http://www.beis.de/Elektronik/DeltaSigma/DeltaSigma.html
http://direct.xilinx.com/bvdocs/appnotes/xapp154.pdf
Just my opinion.
-
Hi Dennis,
once again fantastic work you've done - I'm really looking forward to seeing the photos (and hopefully video) from the Amiga party on the 18th! I'd love to see this commercial as much as I'd love to see it go open source. It's your baby so you should run with it how you see fit.
Don't listen to all the pressure from the users on here about it should work with 68xxx, PPC etc.etc. you have it working how you want first. For a base Amiga 500 'emulator' then OCS, KS 1.3 and 0.5 CHIP/0.5 FAST memory is all that is needed for compatibility for a huge percentage of the software available for the Amiga.
Should you want to release a small amount of VHDL code to the Amiga community, then could I suggest/request it is the code for the CIAs and PAULA chips - so that an open source CATWEAZEL style of card could please be developed (read/write Amiga format disks and have Amiga joystick/mouse ports)? The CATWEAZEL card is not manufactured any more so there is no other alternative product that you'd be undercutting.
-
As far as I've understood Dennis doesn't have a real implementation of Paula's FDD I/O, rather he mimicks floppy access on a flash drive...
I was thinking about something like that, too. An easier project to start with FPGAs - I've tinkered with that for years really, but don't have the time anyway. I'd like to build
- a HD capable FDD controller for Amy, either transparent (buffering and half-speeding the track) or dedicated and full speed (someone else would have to write a driver)
- an Amiga FDD format interface for a PeeCee, probably through USB
- a native USB controller for Z2/3 (probably far out)
- an AGA expansion for my A3k (way out...)
-
DoubleDutc wrote:
Dennis: You haven't posted in a while, will you be making your design available (with pic and vhdl or verilog source code)?
See my earlier post in this thread, I still don't know as I am afraid someone grabs the code and makes it commercial . Since my first post I have got 7(!) offers to cooperate on commercialising Minimig. :-o
How about releasing some sample source code that will give people who are new to FPGAs some insight into getting Amiga quality video and sound.
For example you could release the sound code and video code to do a standard Amiga gfx mode on a tv - but no blitter, copper or sprites.
This way its useless to anyone who wants to rip you off, but gives beginners some useful info on how to generate good sound and get output to the rgb input of a tv.
?
-Anthony
Best to look at UAE source, much easier to understand than a pile of verilog code.
If generating some video signals or audio is what you are after, there are much much easier ways to do this than using an fpga.
I'd like to learn some fpga stuff myself, but i think using his project as a source for a newbie like me would be hardcore overkill, and i probaly learn more from using simpiler projects as an experimental test bed.
I think that the main value of dennis's project is that it shows that a single guy can still pull off an amazing feat all on his own, in his spare time. When i look at all the layers and layers of hardware and software in a modern system, i often think to my self, "I'm supprised that this all actually works"
10 to 20 years from now, i bet some guy will do the same thing he's doing now, only emulating what currently call "modern"
Sorry for going off on a tangent. If you want to learn how the amiga works, check out UAE source, or look at code on aminet. If you want to learn how fpgas work check out the fpga tutorials all over the net. This guys project is probably going to go over the heads of an unexperienced hobbiest.
-
Quote:
I will put an 68000 Emulation into the FPGA. But I have not cycleexact emulation.
Wow!, you run an emulator on your own CPU? How fast will it go?
I hope the RISC can execute the Software so fast as a 10-14MHz 68000.
This is a special RISC. He is only drafted to emulate the 68000.
Here is a Code example(AHDL) how the RISC Jump into the nativ RISC-Code for the 68000 Instructions.
------------------------------------------------------------------------------------
------------------------------------------------------------------------------------
if i_A32 AND stat[1..0]==B"00" THEN --Jump to next RISC EmuCode
(flag_ztmp, flag_c, flag_z, n_flagshigh).ena=VCC;
stat[0] = B"1"; --cancel stored RISC Command
IF OP1out[15..12]==H"6" THEN
code68K[]=(OP1out[15..3],B"111"); --choose A7 (Stack)
ELSE
code68K[]=OP1out[];
END IF;
code68K[].ena=VCC; --OP Code merken
CASE OP1out[15..12] IS
WHEN H"0" => --i befehle
IF NOT OP1out[8] THEN
addrmode[]=B"011"; --0x10000-0x13fff
END IF;
WHEN H"1",H"2",H"3" => --MOVE
IF (OP1out[11..10],OP1out[8..6])==B"00111" THEN -- ->st nn.w or st nnnn.l
riscaddr[2]=B"1";
IF OP1out[9] THEN --(nnnn).l
addrmode[]=B"001";
END IF;
END IF;
WHEN H"4" => --i befele
IF NOT OP1out[8] THEN
IF OP1out[11..9]==B"111" AND OP1out[7..4]==B"0111" THEN
addrmode[]=B"101"; --0x1c000 RESET,NOP,STOP...
flag_c=68KFlags[4] AND (OP1out[11..9]==B"000" OR OP1out[11..6]==B"100000"); --NEGX, NBCD
ELSE
addrmode[]=B"011"; --0x14000-0x17fff
END IF;
END IF;
WHEN H"5" =>
n_regwr=VCC;
IF OP1out[7..6]==B"11" THEN --DBcc, Scc
registerin[7..0]=condition;
flag_c=condition;
ELSE --ADDQ/SUBQ
registerin[2..0]=OP1out[11..9];
registerin[3]=OP1out[11..9]==B"000";
registerin[7..4]=GND;
END IF;
WHEN H"6" => --bra, bsr, bcc
addrmode[]=B"010";
riscaddr[15..5]=(B"010111",OP1out[7..0]!=H"00",B"11",OP1out[11..9]!=B"000",OP1out[8]);
riscaddr[2]=VCC;
flag_c=condition;
WHEN H"7" => --MOVEQ
addrmode[]=B"010";
WHEN H"8",H"C" =>
flag_c=68KFlags[4] AND OP1out[8..4]==B"10000"; --SBCD, ABCD
WHEN H"9",H"D" =>
flag_c=68KFlags[4] AND OP1out[7..6]!=B"11" AND (OP1out[8],OP1out[5..4])==B"100"; --SUBX, ADDX
WHEN H"e" => --shift befehle
IF NOT OP1out[11] AND OP1out[7..6]==B"11" THEN
addrmode[]=B"100"; --0x18000-0x1bfff
flag_c=68KFlags[4] AND OP1out[11..9]==B"010"; --ROXL,ROXR
ELSE
flag_c=68KFlags[4] AND OP1out[4..3]==B"10"; --ROXL,ROXR
END IF;
WHEN OTHERS =>
addrmode[]=B"000";
END CASE;
Viele Grüße
TobiFlex
PS. Sorry, the Preview show me no TABs in the Source :-(
-
Since my first post I have got 7(!) offers to cooperate on commercialising Minimig.
That sounds good. I suppose it is still to early to say, but what type of commercialising do they want to do, Amiga in a joystick or a complete minimig board?
-
Personally if it went commercial I wouold love to see an "Amiga In A Stick" but with an SD port to allow people to add thier own ADFs to run :-D
-
I've just been informed on this project and tried to read the entire thread, which is next to impossible.
This project is indeed impressive, and as I understand it has endured a lot of flame and skeptisism. (Did that Jethro Tull bloke ever reappear, or did he burn himself on his own pyre?)
Anyway. As I see it, is this project another sign that the tide keeps swinging between the rule of large corperate developments and enthousiasts taking matters into their own hands.
The industry gave is FPGA technology to play with as much as all the electronic components of the olden days!
Once this amiga on a stick has materialised, just leak it to the community and see what happens. Dennis is just laying down the ground works and already people start working with it!
Really, Damn the lincences! Once the community take charge of this, the industry will eventually give way and in the end be glad that all this development work is done for free!
Let's face it, nobody in any community ever won a court battle over intellectual property generated from hobby projects.
The amiga community isn't worth saving, the amiga community is in effect a figment of our collective imaginations.
But the so called amiga community is very usefull in generating new technology.
@ Dennis,
Please. Don't try to hold on to it like some paranoid mad scientist. Nobody will get rich from this anyway. As you can see there are people here who can add to this, and work with it. Let them do it. We don't even have to work together as a team to let it happen.
There is no right, or rightous way to do it.
It's just technology, it will tell you of it's done right by working properly!
Oh dear. I've been on a rant again! :insane:
-
I've just been informed on this project and tried to read the entire thread, which is next to impossible.
This project is indeed impressive, and as I understand it has endured a lot of flame and skeptisism. (Did that Jethro Tull bloke ever reappear, or did he burn himself on his own pyre?)
Anyway. As I see it, is this project another sign that the tide keeps swinging between the rule of large corperate developments and enthousiasts taking matters into their own hands.
The industry gave is FPGA technology to play with as much as all the electronic components of the olden days!
Once this amiga on a stick has materialised, just leak it to the community and see what happens. Dennis is just laying down the ground works and already people start working with it!
Really, Damn the lincences! Once the community take charge of this, the industry will eventually give way and in the end be glad that all this development work is done for free!
Let's face it, nobody in any community ever won a court battle over intellectual property generated from hobby projects.
The amiga community isn't worth saving, the amiga community is in effect a figment of our collective imaginations.
But the so called amiga community is very usefull in generating new technology.
@ Dennis,
Please. Don't try to hold on to it like some paranoid mad scientist. Nobody will get rich from this anyway. As you can see there are people here who can add to this, and work with it. Let them do it. We don't even have to work together as a team to let it happen.
There is no right, or rightous way to do it.
It's just technology, it will tell you of it's done right by working properly!
Oh dear. I've been on a rant again! :insane:
-
I do not know if this has been talked about on this forum.
Back in 1996-97 a company called Dracos Systems made a computer that used a CPLD or FPGA to create enough compatability to use Amiga OS. (I hope I find that issue of CU Amiga) I guess I do not remember the specs. Would creating enough compatability to get RTG and AHI going be tough?
Thanks for listening.
-
To run the exec kernel, you need at least :
- the INTREQ / INTENA registers
- the SERPER / SERDAT registers (for debugging routines)
- the 8520s (for the time-slicing done in the timer.device)
-
@trekiej
Maybe you mean the Draco (http://amiga-hardware.com/showhardware.cgi?HARDID=43) from MacroSystem?
-
@ Senex
Yes you are correct. I looked at some mags after I posted. (oops)
I noticed in the past that some manufacturers sold video cards, accelerators, and sound card for the Amiga. I was wondering if all they needed was a main board with a FPGA to glue it all together.
It would be nice if someone could build an ATX Mother Board with a group of FPGAs that could be used for whatever purpose they wanted. This probably not a new idea(C-one).
Thanks.
-
@Dennis
Keep up the amazing work, the project seems to be progresssing extremely well.
I noticed that there isn't a dedicated website for this project, if you are interested in setting up a website I am willing to donate some hosting if required.
-
Dennis, how about another progress report? I'm interested to see how far you've progressed.
-
Dessimat0r wrote:
Dennis, how about another progress report? I'm interested to see how far you've progressed.
Its getting demoed at the show tomorrow, so I figure by Sunday at the latest we'll be hearing all about how the demo went.
-Tig
-
Ohhh, I see -- rock on :D
-
I ám sorry for my terrible English, but I have some pictures for you of the Minimig… and yes It’s real ;-)
What is the Minimig? it’s a ocs amiga in an FPGA acts and feels like any other amiga.
I askt Dennis a few questions about his plans for the minimig. It will not be a AGA amiga, maybe a ECS. Its not clear what it will be in the future, I think it will be a small amiga in a box to play games on.
Its not going to be a new amiga just a small OCS/ECS maybe with some improvements like better sound.
The minimig works exactly like Dennis says so stop having doubts about him ;-)
I personally hope that in the near future it will be a addon card for let’s say the amiga one and works with os4 to be backwards compatible with the old software. That you can buy a amiga with of without classic compatibility. But I fear it will be a game device. So in the end just a joystick in a amiga I hope not.
The designer of the cat weasel device was talking to Dennis about I believe just that, one chip one small amiga. And sure there is money to be made in that, but I believe and hope it will become something better that that.
Nice speaking with you Dennis, and good luck.
For all the people how still believe this can’t be done please go on, it will make his day.
Regards Sebastian
(http://home.versatel.nl/sebasvdm/1.jpg)
(http://home.versatel.nl/sebasvdm/2.jpg)
(http://home.versatel.nl/sebasvdm/3.jpg)
(http://home.versatel.nl/sebasvdm/5.jpg)
Second on the left is Dennis ;-) Not as old as we expected :-)
-
Very nice! Are there any other reports on it from the event? I want to know what games were played, and how they ran, etc.
-
Sell me one!!!! :-D
-
Games played on the minimig? I have seen lemmings and turican running just fine a few demos to, but remember it has just 0,5 mb chip and 0,5 fast so most of the games don’t run for now, but lemmings and turican run smooth. The rest of the event well, Mmzz I have seen lots of realy old stuff like the commodore pet. Some amiga’s noting special. I was there only for the minimig. I am not really interested in really old people showing me there latest creation of guess the number game on the c64 ;-)
The coffee was 1 freaking euro, but came with a FREE cookie, the sandwiches 2 euros, we stopped at a local gas station for food on the way back. To make it short, If you like sitting all the day behind a c16 c64 or amiga go there but I have seen better ;-)
-
voyager wrote:
The designer of the cat weasel device was talking to Dennis about I believe just that, one chip one small amiga. And sure there is money to be made in that, but I believe and hope it will become something better that that.
I think Dennis has everyright to make some money for his troubles, and amiga in a joystick is not something to be shamed of. Though of course I would like to see this project as a standalone computer as well, and I believe it is possible to release it as different platforms, just dennis has to be careful about the patents and ownerships of stuff.
-
Just came back from the show (Voyager was faster. BTW, i'm the guy in the green jacket ;) ). More pictures for you all at:
http://amigaroxx.zuurkool.com (http://amigaroxx.zuurkool.com)
The Minimig runs like a charm. Watched several demo's (e.g. Mental Hangover) and played different games like Lemmings and Parasol Stars. Only PS showed a spritebug on top where the points are displayed.
I truly see a bright future for Dennis, if he makes the right decision. I think Jens Schöenfeld can help a lot.
It surely was a nice show with some familair faces even after such a long time. Keep it up Dennis!
-
ROCK ON!!!
-
Thanks mendark, great pictures - I wish I could've seen it.
(BTW: The C64's serial bus is no I²C, it's IEEE-488 'serial CBM' version, sometimes called 'IEC'.)
-
Thanks mendark, great pictures - I wish I could've seen it.
(BTW: The C64's serial bus is no I²C, it's IEEE-488 'serial CBM' version, sometimes called 'IEC'.)
D'oh! Thank you Zac67! Fixed it!
Sorry you could'nt not see it in real life. It's absolutely amazing and Dennis is so calm about it.
It's just amazing!!!
(will start tomorrow with drawing some possible cases for the minimig :idea: )
-
mendark wrote:
(will start tomorrow with drawing some possible cases for the minimig :idea: )
***COMPO TIME***
Oh come on, we should have an Amy'05 style competition for the design and the winner could win......er....er....
-
Even if it only becomes an Amiga in a joystick type thing, make it hackable and the hardcore Amiga fans as well as mainstream nostalgic A500 owners will like it!
*Starts to save up some money for a couple of minimigs in a Tac-2 shell (transfers some money from Boxxer savings account
-
Good to hear something may come out of this :)
I hope we don't end up with just an "Amiga in a joystick" though. What I'd like to see is a little box with keyboard and mouse ports (USB or PS/2) and an Amiga-style joystick port so I can use my old joystick (or USB with some sort of adapter so I can use my old stick if I want.)
Not that the self-contained joystick idea isn't good, it would be nice to be able to choose which joystick to use.
-
I would also like to see this offered as a small computer.
Especially if Dennis has designed in some advanced features over the original OCS chipset (better sound, more colours on screen and such).
I think something with a compactflash reader in a small enclosure with 2 Amiga joytstick/mouse ports and a ps/2 keyboard port has the potential to really take off.
Price it at $199 (or even less) and watch it sell like theres no tomorrow.
Make it so the CPU can either be upgraded, or switched to a higher end model, and I would buy several of them!
-
It is great to some photos of Minimig in action.
I think that if Dennis does decide to work with Jens good things can happen for himself & Minimig where Minimig can grow in ways that benefit Dennis, Jens, and the rest of the Amiga community. Jeri Ellsworth, logic designer for one set of C64 core logic for C_1, & Jens Schoenfeld did work together to make the C_1 FPGA board with a C64 logic core come into existence. The paths that the C_1 C64 logic design took would also be pretty good for Amiga/Minimig:
A) A C_1 like FPGA Developer board with SDRAM, a CPU Card slot, VGA video connector, PS2 mouse/keyboard ports, and a couple of PCI slots for a C_1 like Minimig for users such as myself wishing to enhance this design to add things such as an IDE interface, Coldfire V4e CPU, DDR SDRAM, Compact FLASH interface, etc. The Verilog HDL or a Xilinx NGO/Altera encrypted net list for the core logic would have to come along with a C_1_Minimig board purchase to make it interesting to enhance/expand.
B) An Amiga Direct-to-TV joystick that is hackable to turn it back into a some kind of Amiga where one can add back on a joystick port, PS2 keyboard, PS2 mouse, VGA output, and a SD_CARD slot so that everyone can play old games & run old Amiga software that once worked on the Amiga500/1000.
Also, in the short term, if Dennis does upgrade his Xilinx Spartan3 development board to the newer Spartan3E development board from Xilinx($149), Minimig could get 64Mbytes of memory which is more than a stock 68000 could even address.
http://www.xilinx.com/xlnx/xebiz/designResources/ip_product_details.jsp?key=HW-SPAR3E-SK-US&sGlobalNavPick=PRODUCTS&sSecondaryNavPick=BOARDS
:-)
-
broken wrote:
Price it at $199 (or even less) and watch it sell like theres no tomorrow.
Make it so the CPU can either be upgraded, or switched to a higher end model, and I would buy several of them!
I'd buy one in an instant. Shall i get some Amiga Joysticks out of the cupboard already ?
-
Maybe some dontation fund could be set up to buy him the newer Spartan 3E board?
-
As an Industrial designer who has worked on the product development of around 50-plus retail products that sell throughout the US, Europe and Australia, I see Dennis facing a few issues with this were it to go commercial.
The biggest problem Dennis will have with minimig as an A500-in-a-stick is licencing the IP. ie working out how much of the minimig concept he actually owns, and what he owes to the current Amiga IP owners. IP law when it comes to reverse engineering (which minimig is) is vague and frought with potential disaster.
The other issue is distribution and marketing, many of the people who developed the C64DTV have been short-changed by their corporate "partners". The situation is difficult for a hardware minded (as apposed to business minded) person to action against without very good legal advice and the required funds.
Compared to the C64DTV I wonder how much mainstream appeal an A500-DTV would have? Possibly much less, due to the Amiga being less well known. Research?
At the end of the day, I hope we see an A500 in a stick at some point in the future, and I hope Dennis gets suitable financial reward for his efforts.
Good luck Dennis, it seems it's not to good to be true.
-
coldfish wrote:
As an Industrial designer who has worked on the product development of around 50-plus retail products that sell throughout the US, Europe and Australia, I see Dennis facing a few issues with this were it to go commercial.
The biggest problem Dennis will have with minimig as an A500-in-a-stick is licencing the IP. ie working out how much of the minimig concept he actually owns, and what he owes to the current Amiga IP owners. IP law when it comes to reverse engineering (which minimig is) is vague and frought with potential disaster.
The other issue is distribution and marketing, many of the people who developed the C64DTV have been short-changed by their corporate "partners". The situation is difficult for a hardware minded (as apposed to business minded) person to action against without very good legal advice and the required funds.
Compared to the C64DTV I wonder how much mainstream appeal an A500-DTV would have? Possibly much less, due to the Amiga being less well known. Research?
At the end of the day, I hope we see an A500 in a stick at some point in the future, and I hope Dennis gets suitable financial reward for his efforts.
Good luck Dennis, it seems it's not to good to be true.
As a hobbiest who's started an awful lot of projects ( and never completed hardly any of them) I'd say the toughest part is actually finishing and polishing the whole thing.
-
I really hope to see a production model of the Minimig on sale as I'd love to have one attached to my TV with my favourite games held on a single SD card (Something small enough that the wife won't notice it).
I hope that the main device comes it it's own box where you can plug in a power supply, video cable, PS2 keyboard, joystick and mouse. I wouldn't be too keen on an Amiga-in-a-joystick design as I've broken way to many joysticks on real C64s and Amigas over the years. :-)
-
The biggest problem Dennis will have with minimig as an A500-in-a-stick is licencing the IP. ie working out how much of the minimig concept he actually owns, and what he owes to the current Amiga IP owners. IP law when it comes to reverse engineering (which minimig is) is vague and frought with potential disaster.
My understanding is that Minimig doesn't even use reverse-engineering for the Kickstart ROM, but the original as-is. As such, it'd need the license for the KS ROM from Amiga Inc (or whoever owns the rights really). Good luck with that.
-
coldfish wrote:
As an Industrial designer who has worked on the product development of around 50-plus retail products that sell throughout the US, Europe and Australia, I see Dennis facing a few issues with this were it to go commercial.
The biggest problem Dennis will have with minimig as an A500-in-a-stick is licencing the IP. ie working out how much of the minimig concept he actually owns, and what he owes to the current Amiga IP owners. IP law when it comes to reverse engineering (which minimig is) is vague and frought with potential disaster.
The minimig isn't reverse engineered, no one disassembled an A500 and then wrote a specification for him to use... he built a computer compatible with publicly published documentation, as published by Commodore themselves!
It's different if he wants to use the kickstart... but then he might not need a complete kickstart.
The other issue is distribution and marketing, many of the people who developed the C64DTV have been short-changed by their corporate "partners". The situation is difficult for a hardware minded (as apposed to business minded) person to action against without very good legal advice and the required funds.
Compared to the C64DTV I wonder how much mainstream appeal an A500-DTV would have? Possibly much less, due to the Amiga being less well known. Research?
The biggest problem I see with the market is that one can already buy a "Sega joystick game thingy" which runs lots of games that the Amiga ran... For the games which are more unique to the Amiga, it will be a struggle to find the IP holders and then obtain a licence...
At the end of the day, I hope we see an A500 in a stick at some point in the future, and I hope Dennis gets suitable financial reward for his efforts.
Good luck Dennis, it seems it's not to good to be true.
We all wish Dennis good luck!
-
Like others, I just hope that the IP side of things doesn't completely kill this from either a commercial or hobbyist perspective.
Personally, I think there's a market for an amiga in a joystick style thing, but also a market for an amiga on a board too. The joystick, I guess, will be far more volume-driven and will be battered on price-point: c64tv's are now dirt cheap, Sega embedded systems are likewise.
I really do hope that this comes to something and doesn't go the way of other potentially great amiga projects.
If you're after commercialising it, "trust no-one" seems to be the story (and that's no disrespect intended to individual computers!). Start business like, and keep business like.
If you're going to GPL it, then wow...!
-
I wrote games for the Amiga and know several others who also did. We sold the rights (to the Amiga version) to various software publishers - such as Psygnosis, Ocean, US Gold; as did most others who did not work in-house.
There are two main questions when including the "bundled" games. Who owns the rights to the same games published on this new machine (is it an Amiga)? Then (if you find out who owns the rights), whre are these people now?
Then there is the question of royalties...
I personally don't mind if people have a free copy of a game I wrote years ago. BUT, I do mind if its bundled with a machine being sold for a profit. I think the same can be said for many others. I think there will be copyright owners seeking royalties on such a device.
Even if a game is not bundled, but the user allowed to insert a SD card (or similar) with disc images that the user has copied on there. Dennis is opening himself up to new laws designed to stop this kind of thing happening. I'm sure its now illegal to build and sell a device that will bypass a scheme designed to prevent copying; for example the disc protection in the games that deigned to prevent copying, but will now be bypassed because of the nature of the medium being changed. I personally wouldn't mind this happening, but I know others who would.
If there was no memory card slot and rights were obtained from the correct owners then I suppose all these problems would go away. :)
-
bloodline wrote:
coldfish wrote:
-snip
The minimig isn't reverse engineered, no one disassembled an A500 and then wrote a specification for him to use... he built a computer compatible with publicly published documentation, as published by Commodore themselves!
It's different if he wants to use the kickstart... but then he might not need a complete kickstart.
-snip
The biggest problem I see with the market is that one can already buy a "Sega joystick game thingy" which runs lots of games that the Amiga ran... For the games which are more unique to the Amiga, it will be a struggle to find the IP holders and then obtain a licence...
On the reverse engineering issue, I hope what you say is true. I suppose it isnt hugely different to UAE, where routines implimented in original hardware are emulated by different routines in software, except those routines are now emulated in hardware. I have a feeling though, given the Amiga's history, that someone will crawl out of the woodwork to argue for their piece of the royalty pie.
Those "Sega joystick game thingy" are at least a good example showing a more complex 16bit device can be produced at the same pricepoint as simpler 8bit "in-a-stick" devices, ie dirt cheap! Radica games seems to be pretty good at these things.
http://www.radicagames.com/manufacturing.php
-
First off great work Dennis!!
And now I'll play some devil's advocate here though:
I have an Amiga in a Joystick it's called a modded X-box running a UAE port on it. I could be wrong but most people with the proficiency to hack out an "Amiga in a Joystick" to add a keyboard and a mouse would also be able to at least softmod an Xbox and install UAE on it and even with a stock HDD at least 5GB of free hdd space for adfs. I, personally, have not tested a slew of games/demos running UAE on the Xbox but, using a joystick only, frustration kicks in REAL fast even with the option to input keys from the menu. Let's see that's "Start -> 2 x Down-> A -> 9 x Down -> A" .. and I get a B for my Bard's name. So for games that require a keyboard there very simply is no easy way around not having one.
-
Wow i don't know what countries you guys live in but we still have one or two liberties left in the UK, No laws exist here that would stop you building something that could be used to circumnavigate copy protection, Otherwise you would not be able to buy a PC or a game console or a tape deck or a VCR.
Please think about what you are typing before trying to scare monger in this thread "I think such n such law exists" is such a silly comment.
-
Actually Flexinoodl your wrong. I live in the UK too, its now illegal to manufacture or sell a machine that will bypass macrovision, sell psx, xbox, or ps2 "chips", or basically do anything that will bypass protection. Its also the same in the US. I think someone won as case against Sony in Australia, but on appeal I think they lost.
If you want to see what laws are broken do a search for DMCA for the USA and EUCD for Europe and specifically the Amended Copyright, Designs and Patents Act in the UK.
I don't agree with the above laws, I actually signed petitions to have them stopped alongside other patent and copyright laws; but those with the money pushed them through as always. :(
It looks like you have one less liberty than you thought. ;)
-
But would this device fall under these laws? The Minimig doesn't override any copy protection even with a card reader attached. The fact that it can access ADF files it not illegal as there is nothing stopping anyone distributing software as an ADF file. ADFs are not illegal.
If this was illegal then all home VCRs sould be banned simply because it is "possible" for them to play a tape that may have been pirated elsewhere. Heck, that would also make all DVD players illegal if they are capable of playing DVD+/-R disks.
In the case of things like PS2 and XBox mods, the lawyers were taking on "chip" makers who were modifying actual hardware for the sole purpose of accessing prirate software. All the Minimig does is allow access to a file type stored on a memory card. Now, if the Minimig had some built in routines to get around passowrd protection in games then you might have a case.
That's my opinion of the situation. :-D
-
I hereby apologise profusely for ever doubting the authenticity of this project.
No hard feelings Dennis. :-D
-
I hope we don't end up with just an "Amiga in a joystick" though.
Well, the C64 joystick was made fully hackable thanks to Jerri's efforts, I don't see the trouble in gettng that made if it leads this project into fruition (commercializing of the thing)
Personally, I rather have a big toy company distributing this at Toys R Us than having a {bleep} "amiga-friendly" "company" trying to do it and failing miserably. Last thing I'd like to see is this project going bad,, and it seems Jerri got it quite good and you can turn any C64DTV into a full-on computer with great new capabilities!
-
Any new developments on Minimig?
-
Dessimat0r wrote:
Any new developments on Minimig?
Lol, I knew this would happen.
The masses are impatient for more! :-)
@Gadgetmaster
Good on you for apologising, it takes a big man.
@everyone
Anyone know if any video footage was taken at the recent demo. I'd love to see it.
-
It has been a while since my last update but I have not been sitting still. First of all, thanks to all who have visited the HCC meeting, it was nice to meet people who are interested in my project.
As for Minimig; I have fixed some bugs and have started work on moving Minimig to it's own pcb board. For projects of this size to get finished, it is always important to keep the goals clear and my goal from the onset was to make an A500 clone, nothing more (well perhaps a little more :-D ) and nothing less. To keep this goal "for real" I have trimmed the specs down a little. The specs of the board I am now designing is as follows:
board: 2-layer (yes, that'll be a challenge) 150mm x 150mm
cpu: freescale MC68SEC000 @ 10MHz
memory: 2Mbyte of 70ns SRAM (0.5Mbyte for kickstart)
fpga: Xilinx Spartan-3 XC3S400-4PQ208C
microcontroller: PIC18LF252-I/SO
2 9-pin controller ports
2 PS/2 ports (mouse and keyboard)
1 15-pin VGA port
1 3.5mm stereo audio jack
1 9-pin serial port
1 power input jack 8 - 12VDC
1 MMC card slot
2 leds (power and drive)
1 push button (OSD menu)
1 switch (31/15khz video out)
Although this board might seem pretty basic, it is still a hell of a task to get every little detail and every litte "design challenge" right and this should keep me busy for weeks to come. (especially because I want to make it so small) Once this board is built I can finish Minimig and finally add keyboard support, OSD menu and hopefully play all 1Mbyte games!
Gadgetmaster wrote:
I hereby apologise profusely for ever doubting the authenticity of this project.
No hard feelings here either :-)
By the way, has anybody seen Jethro_Tull ?
Dennis
-
Dennis, thanks for the update !!
You have made some really good progress.
Even as an A500 clone, it still sounds as an interesting development board.
With only a 2layer PWB, that is going to be tricky with the Spartan3 FPGA in a QFP since the QFP 208 package leads will have a lot of lead inductance. You will have to be very careful with close decoupling cap placement, protecting against power/ground bounce (wide power & wide ground trace widths), and termination on the bus lines. Internal series terminations are available on the Spartan3 I/O for the bus lines from FPGA to 68K CPU & FPGA to SRAMs (due to fast edge rates on these modern ICs in the sub micron process (.15um) Vs the old slow edge rates found on the old 74LSxxx logic found in some old Commodore Amiga boards). I see that you are a hardware Engineer as I am; thus, you know what I am talking about with edge rates & Signal Integrity (SI)issues.
Also, one more thought for you is that 2Mbytes of SRAM is probably going to be expensive as compared to a single x16 wide SDRAM. Have you considered using a single 126Mbit (8Mx16 => 16Megabytes SDRAM) SDRAM in a 54pin TSSOP package instead ?
http://www.micron.com/products/dram/sdram/part.aspx?part=MT48LC8M16A2P-75
Xilinx (LogiCORE) & Altera (MegaWizard) both have SDRAM controllers in their IP libraries. Here is the link to the Xilinx SDRAM controller appnote & Verilog Source code for it:
http://www.xilinx.com/products/design_resources/mem_corner/resource/xaw_dram_sdram.htm
Are you going to sell some development boards to others (blank or populated) to also do development as well or is your new PWB to help further progress your internal development of Minimig ?
:-)
-
To anyone:
I wonder if Amiga INC could use this to end the hardware shortage? I do not know if anyone has talked to them or even who would be the best to ask the question.
-
Sounds NICE
-
Personally, I'd like to see a DVI out connection option also, so that I can have my LCD monitor in it in digital mode, and possibly a TV out (could the VGA connector handle this?)
Out of interest, why does it seem to be easier to generate an analogue signal for video out, rather than a digital one in many cases? You could have digital out, and then some extra circuit board that jacks onto the digital out connection, converting it to analogue out.
-
Have you any idea how much a DVI-D encoder/phy chip costs?
TV output via RGB SCART is possible (i.e. no extra chips, just a few resistors) it depends if dennis has any more free pins on his FPGA to route out the signals before the scandoubler. However a SCART socket is big and dennis appears to be working towards an ultra compact form factor.
Composite output for our American friends would require an extra video encoder chip, not expensive, but the "A520" control logic is extra design work.
Out of interest, why does it seem to be easier to generate an analogue signal for video out, rather than a digital one in many cases?
Digital signals, especially buses are not designed to travel long distances. DVI-D encodes the data into a very high-speed serial stream for transmission. This encoding is quite intensive and the high speed serial required for uncompressed 24-bit SXGA and above resolutions makes the phy complicated and expensive.
A cheap Analog "hack" is to use little more than a few resistors to act as a Digital to Analog converter (DAC).
-
It's a good plan to restrict yourself to a specific configuration of hardware. Get this A500 clone hardware system fully working first, and leave the other stuff for future projects.
Why did you cut the speed of the 68000? The extra speed comes in handy for some of the 3D games. Maybe a 14MHz 68000, at exactly 2x the original clock speed would suffice?
As long as it plays Lionheart, Populous II and Alien Breed, I'll be happy if you can come up with something that is remotely affordable, even on a DIY project basis.
-
If the CPU speed would be the maximum from the start there is nothing left to hack ;)
-
Best thing ever happened since the A1200.
-
Doppie1200 wrote:
If the CPU speed would be the maximum from the start there is nothing left to hack ;)
My thoughts exactly.... Overclocking is always good fun, and could perhaps even be done with software in this case.
-Paul
-
Great news regarding the specs,
one observation (sorry if it has been discussed already - didn't have time to go through thr thread):
could the memory card reader be SD/MMC (instead of plain MMC), so as to ensure it's future proof (with mini-SD and so many secure-digital format spin-offs) - just an idea.
Well done and looking forward to seeing the final product :)
-
someone trying to stir up trouble about this hobby project: http://www.amiga.com/forums/viewtopic.php?p=22166#22166 (http://www.amiga.com/forums/viewtopic.php?p=22166#22166)
-
@Dennis
Since you're doing a re-design of the pcb, why don't you go nano-itx? ;) (just jokin', man! :D )
about the amiga.com forum thread: world is full of morons, never mind!
-
With only a 2layer PWB, that is going to be tricky with the Spartan3 FPGA in a QFP since the QFP 208 package leads will have a lot of lead inductance. You will have to be very careful with close decoupling cap placement, protecting against power/ground bounce (wide power & wide ground trace widths), and termination on the bus lines. Internal series terminations are available on the Spartan3 I/O for the bus lines from FPGA to 68K CPU & FPGA to SRAMs (due to fast edge rates on these modern ICs in the sub micron process (.15um) Vs the old slow edge rates found on the old 74LSxxx logic found in some old Commodore Amiga boards). I see that you are a hardware Engineer as I am; thus, you know what I am talking about with edge rates & Signal Integrity (SI)issues.
I know, that is my main concern at the moment. The PCB will have the bottom layer dedicated to ground plane and the top layer to routing. (routing will be simple, very few crossing wires). Beneath and around the FPGA I plan to create some local power planes/islands with plenty of decoupling caps to tie them to the global ground plane. This way I should be able to minimize ground bounce and power supply noise even with a 2-layer PCB. But...If it doesn't work out I will certainly move to 4-layer. At the moment I'm looking into the current ratings required for the VCCINT and VCCAUX LDO regulators.
Also, one more thought for you is that 2Mbytes of SRAM is probably going to be expensive as compared to a single x16 wide SDRAM. Have you considered using a single 126Mbit (8Mx16 => 16Megabytes SDRAM) SDRAM in a 54pin TSSOP package instead ?
SDRAM is very difficult to work with in the Minimig environment. The Amiga chipset accesses the memory in a very random manner. With SDRAM this is not ideal as you have a lot of overhead when fetching random data. However, when browsing to the Digikey catalog I found out that you can now buy PSRAM (pseudo SRAM) with the ease of use of SRAM and the price of SDRAM. That would be ideal for Minimig.
Are you going to sell some development boards to others (blank or populated) to also do development as well or is your new PWB to help further progress your internal development of Minimig ?
At first it will be mainly for my own development purposes.
Why did you cut the speed of the 68000? The extra speed comes in handy for some of the 3D games. Maybe a 14MHz 68000, at exactly 2x the original clock speed would suffice?
Price
Dennis
-
But...If it doesn't work out I will certainly move to 4-layer.
Let's keep our fingers crossed for the 2-layer - much cheaper!
At first it will be mainly for my own development purposes.
Maybe, when evrything has come together and minimig is to all intents and purposes a perfect A500 Amiga and this has been demonstrated to the whole Amiga Community.....
You could do a KLF and take all the prototypes, plans, verilog files to a remote Scottish cottage and burn it all. :-D
Edit: Whoops! Forgot to thank Dennis profusely for his update!
-
Dennis,
I want to thank you for what you are doing. I also want yo u to post more of your updates. No rush on the release, just like hearing the interesting things that you are doing with minimig.
Thanks again,
Clyborg
-
I concur.
-
First, I would like to thank Dennis for his great work and talent, bringing something appealing and fresh (although a A500 is not new :) ) to the Amiga community.
I would like to write some words about minimig and how the people are reacting / expecting.
I read here many people asking for AGA support, better sound, better graphics and so on.
As far I understood, the minimig will be an A500 "emulator". I presume this is Dennis goal. No more, no less, but a *good* A500 machine like.
Some people are asking for more features, since what they really are expecting is a new Amiga... Not the minimig.
I don't think it should be a computer. Why ? Well, the A500 was never a good option for using a hard disk - no HD support from the kickstart. Yes, there were addons to connect a HD, but to implement it on the minimig it will have a cost of money and developing time.
You could also use the A600 or A1200 kickstart instead, but the issue remains : adding hardware support for HD.
I think the minimig must be cheap in order to sell. If Dennis starts to add HD support, paralel and serial ports, RGB connector, and a lot of extras needed to use it as a computer, the price will not be low enough to sell many units. Don't forget that most of us would buy an minimig because our Amiga past! The current generation don't have that kind of attachment to the Amiga spirit... If the price rises close to the nowadays consoles ... Well, you know what will happen.
So, I think the Minimig should be a plain A500 with some "fast" memory and a joystick plug to optionally plug a Amiga joystick, or even a PC one (don't know how to simulate a Amiga joystick from a PC one, but Dennis should surelly know). A mouse port and a keyboard plug is also mandatory, since many games need keyboard input.
Minimig should be an entertaining platform and not a complete computer. Otherwise, it will be costy and will sell poorly.
IMHO
-
ptek wrote:
First, I would like to thank Dennis for his great work and talent, bringing something appealing and fresh (although a A500 is not new :) ) to the Amiga community.
I don't think it should be a computer. Why ? Well, the A500 was never a good option for using a hard disk - no HD support from the kickstart. Yes, there were addons to connect a HD, but to implement it on the minimig it will have a cost of money and developing time.
I also extend kudos to Dennis on a great project. Even if it never becomes a commercial product or opencore, it's still a needed bit of inspiration. :-)
But you're wrong on the HD addon. The 500 uses the same kickstart as the 2000, and it's had autobooting HD support since 1.3. Heck, I even had a HD on my 1000. IDE interfaces are trivial, and it wouldn't surprise me to see a premade core already out there for scsi. The key question will be are there enough pins left on the FPGA?
-
I agree with the low cost target for an A500 clone; however, other additional interfaces such as IDE and Compact FLASH are not complicated and not expensive to add. If Dennis does release the Verilog IP in some form with the sale of a develeopment board from him, I am willing to work on some Verilog to add a PIO mode IDE interface and a Compact FLASH contoller to his design (Also would need to add some 74LCX16245 transceivers to deal with the 5volt IDE & Compact FLASH devices). Maybe, he will add some 68K bus (with a clock) expansion connectors on his new single board layout so some of this can be added. However, it is up to Dennis to figure out how he wants to roll out Minimig.
:-)
-
edited.
-
Dennis, this is amazing! I've done hardware and embedded design myself, and I can definitely appreciate the amount of work (and fun!) that goes into it.
For my undergrad CS project 2004-early 2005, I designed and laid out a working ARM-based robot controller board (a prototype successor to the popular Handyboard). It was very cool being able to design the hardware AND do the software for the board at the same time.. its an experience that Dave Haynie mentioned is increasingly rare, and I cherish it.
Currently, I'm having a blast combining my love of music and electronics by resurrecting the C64's SID chip using an FPGA and a dash of analog circuitry. PhoenixSID has been a wild ride with lots of learning and work (and more to come). It's also lots of fun (and once again, I'm doing the hardware and the drivers and demos)! But this is only a sound chip compared to an entire working computer that you've designed! Major kudos!
And yeah, I totally loved my Amigas (started with an A2000, learned 68k asm on it, and later got an A1200, which I upgraded all the way to an '060 tower). Didn't have to get a PC until 1997 :)
Can you tell us about how and why you started the project? What were the initial insights and inspiration? How did you go about building the whole system?
PhoenixSID 65X81 project page:
http://myhdl.jandecaluwe.com/doku.php/projects:phoenixsid_65x81
George
-
@Dennis
Any hackable extras you can add would be a God-send to the community :) (apart from the original God-send :) of Minimig itself)
I know 2 layer doesn't encourage it, but any thoughts on a form of expansion slot? Some form of header or slot.
-
Some form of header or slot.
I agree - this would be fantastic. I disagree with the guy who just wants a joystick - a full featured A500 with expansion would kickstart a whole new community with Dennis as it's God, sorry, mascot! :-D
Bty - Dennis! Have you made any progress or have any thoughts since your last update?
-
Dennis, any more updates on Minimig? It's been a while since you last gave us an update.
-
But you're wrong on the HD addon. The 500 uses the same kickstart as the 2000, and it's had autobooting HD support since 1.3.
Actually it was to be in Kickstart 1.2, but due to unfortunate bug (http://www.iki.fi/sintonen/why_ks_1.3.txt) in the Autoconfig(TM) autobooting code, it would not work.
The irony is that KS 1.3 was later marketed with the notion that it would add autobooting that was allegedly missing completely in KS 1.2. Basically 1.3 was just bugfix, but it was sold as a feature upgrade...
-
Actually it was to be in Kickstart 1.2
Surely the 2000 had HD Autoboot right from the start?
-
@Colin_Camper
Surely the 2000 had HD Autoboot right from the start?
Not with A2091 and KS 1.2.
That's why the A2091 had a jumper to disable autoboot for KS 1.2.
See Commodore A2091 SCSI Interface (http://www.l8r.net/technical/t-a590.shtml)
-
I've read the thread and have some thoughts:
* Replaceing the PIC18LF252 with an CPLD could solve the MFM encode/decode performance issues. And aswell allow for other feature goodies (ether?). Because the CPLD maintains it's bitstream during powerdown it can be used to bootstrap the main fpga and load KS into ram.
* Moveing MFM encode/decode to the FPGA could help floppy performance issues.
* Ethernetbooting would eliminate the need for swapping physical media. Try all software on aminet without ever leaving your amiga keyboard ;)
(althought I think getting it to work properly is the priority #1)
* The CPU should have an 7.14 MHz setting to enable a compatible speed. And avoid possible incompability.
* If there's not enough I/O on the FPGA, one could pack data into serial bus in a southbridge/pci-e alike setup. And link it with another fpga/cpld that has the actual external I/O (ZorroIII bus.. etc).
* Copybypass laws could be circumvented by haveing a real floppy interface. And leaving some "testing pads".. Also the PCB could be sold without bitstreams to minimise capability liability.
* Useing rewritten (AMOS?) KS + WB may help to rid of known bugs and potential license issues that the original have.
* GPL license would enable others to write additions like AGA, ZorroIII etc..
I posted to 'comp.sys.amiga.hardware' in 2005-October regarding Amiga gatecount. And useing the pictures on the original Amiga (Daphne?) I came up with roughly 350k gates. So when I saw your project I knew there was some substance to it. Especially since it was a 400k chip..
Maybe it's possible to get some information on chips from the original amiga designers/developers ..?
I wonder what current generation will come back and revive.. :-)
Keep up the good work anyway!
-
* Useing rewritten (AMOS?) KS + WB may help to rid of known bugs and potential license issues that the original have.
Rewrite KS + WB in AMOS? I think you meant AROS :-D
-
Wow, this project is amazing, it would be interesting to have an Amiga DTV, i tried to read entire post, and i read most of them, some people want to have all kind of connectors, but it will be expensive, and if we want to see this on a shop, it must be cheap. A solution to this problem is make it like C64 DTV, that have soldering points to add things like keyboard, disk driver, joysticks, video output connections, and left to the user add this hacks.
I really amazed also with fpga's, i wondering to buy one, and made my own 8 bit computer, i studied some of digital electronics at universty (but i din't passed the exams, i will hope that one day will pass the exams). I saw a spartan3 starter kit with 200000 gates by 100 eur on a german shop, it will be enough to implement a simple 8 bits computer?
-
* Copybypass laws could be circumvented by haveing a real floppy interface. And leaving some "testing pads".. Also the PCB could be sold without bitstreams to minimise capability liability.
I can't see it being an issue, it's up to the user to supply the software.
* Useing rewritten (AMOS?) KS + WB may help to rid of known bugs and potential license issues that the original have.
Yup. Needs to be chipset compatible though.
* GPL license would enable others to write additions like AGA, ZorroIII etc..
They don't a license to enable that and GPL may actually disable some projects if they wanted to remain closed source. Also, GPL could be problematic if someone ever wanted to integrate it into another existing system which contained existing commercial IP.
-
* GPL license would enable others to write additions like AGA, ZorroIII etc..
They don't a license to enable that and GPL may actually disable some projects if they wanted to remain closed source. Also, GPL could be problematic if someone ever wanted to integrate it into another existing system which contained existing commercial IP.
I like BSD license but then some greedy corp(tm) might just rip the author. I think this would be a pity.
Btw.. when will next status report pop up? =)
-
@dennis
>>>I use a Xilinx Spartan-3 400K gate FPGA. All in all the custom chips (even OCS) are quite complex and at the moment (without the sound) occupy about 60% of the FPGA. <<<
Great project you have Dennis.
I saw a pci card design that uses FPGA chips. Maybe you can take that design and run with it. I'll look it up, and try and comeback and post it, if time permits, or contact me at admin@troikang.com. :-)
Update: Dennis take a look at the open-graphics card specs.
http://lists.duskglow.com/mailman/listinfo/open-graphics
:angry:
-
Hi all,
My name is Vinay, people call me Nay. I am currently a student at Kings College London and as part of my third year engineering degree program i have decided to do a project involving the Amiga.
My actual project involves me documenting the eUAE emulator, in order to do this i have to understand how the Amiga works on a Hardware level and then look into how the software mimics this.
That is just a breif introduction.
i was just wondering how long Dennis has been working on his project and would my documentation have helped befroe you began this task.
Thanks
Nay
-
It could help making an AGA version. I think it would take all the documentation one could get :).
-
I must admit, I've refrained from posting on this one, because I thought it was probably a hoax.
Now - having read in Total Amiga how it has been previewed in February at the HCC GG - I've got to say congratulations, and keep going :-)
Certain other people could learn a lot from your approach of showing some hardware in development (and I don't mean the opportunistic "Trust us, we're honest, can we see the specs?" people (who if they are honest SHOULD KNOW BETTER).
The other people I refer to will go nameless ((harrumph!! Hellbox, Troyyka, AAAaaakk)) - sorry, just clearing my throat there :-)
Since you're a talented individual WITH hardware to show (only a couple of months after a low key announcement), and they're - allegedly - professional companies who made strong announcements (nearly two years ago in some cases) this really shows them up. I know that is not your intention or aim, it is their own lack of actions which show them up as unprofessional.
Anyway, congratulations - I hope it is fun for you. Keep going!
-
Any more status updates Dennis? :-?
-
@Troika
Now an Amy/PV/A1/Peg with builtin minimig capability - that would be very cool!
Bty - I would love to see an update from Dennis - and I bet he is reading this thread and itching to give us one. However I suspect he has taken pity on his wallet and is in secret talks with a commercial partner! IMHO of course!
-
Hi all,
Progress was slow but steady the last few weeks. I expect to finalize the schematic of the Minimig board this week. I only have to add the PS/2 and audio connectors to complete it. If all goes well, I will start routing next week. Routing will be pretty easy as I have the designed the board in such a way, that all routing from the FPGA to the various other chips and connectors will be straight wires without crossings. I have bought enough parts to build a couple of boards for my own use. The schematic took such a long time because I have been busy at work and, to be honest, Minimig is not as much fun anymore as it used to be. You know how it goes, as soon as it works, a new project comes up... (diy hifidelio-like CD-player-with-harddisk anyone? :-D :-D ) But, I said I would finish this and so I will!
Dennis
-
Dennis wrote:
You know how it goes, as soon as it works, a new project comes up... (diy hifidelio-like CD-player-with-harddisk anyone? :-D :-D ) But, I said I would finish this and so I will!
Yup, been there and almost made the Tshirt ;-)
It's also {bleep}ing annoying how the things you think are simple are the bits that take the most time and vice versa.
Keep at it though, this looks like a great project with a lot of potential.
-
Thanks for the latest update Dennis. The Minimig is the only ongoing project where we get any sort of worthwhile progress report and it's facinating to follow the progress you've made.
-
Great work Dennis! Much appreciated!
-
Dennis keep up the great work !!
Since you are closer to having a few PWBs for Minimig, do you know what you are going to do with Minimig in the future ?
Are you going to GPL the source Verilog for Minimig for others to add features on to it or are you going to take Minimig commercial in some direction ?
:-)
-
I imagine the Future of the Project like this :
1. I Download a .zip File, extract Schematics, Firmware,
PCB Layout and some Docs and try to build my own minimig :-D
or
2. I order a complete, tested and working minimig PCB,
pay 99 Euro or whatever and have Fun with it. :-)
which is more likely ?
BTW, thank you for being honest about your Lack of interest in the Project, i can only hope you put it to good use, somehow.
-
Considering you started on a special board. Maybe it's good to make pcb space for a real D/A for video output such that one can get highquality video output ..?
Another option is to provide RGB video data in 4 bits + clock and then let an external pcb process it with either options:
* Cheap resistors (analogue).
* 3x 4-bit D/A (analogue).
* DVI/HDMI transmitter chip (digital).
I think the main point is to ease output of video in another way than the default way.
DVI / HDMI output can be accomplished with a chip like this:
http://www.analog.com/UploadedFiles/Data_Sheets/641001752AD9389_0.pdf
Pinout of the connectors used for digital video (LCD):
http://www.networktechinc.com/technote.html
Conversion of audio from one clock domain to another can be accomplished by writeing output sample to a latch, and then use the other part of source clock to read it out to an outputlatch, that will provide the same sample until the next one is transfered. Ie if Amiga use samplerate of 4-16 kHz and output domain will use a constant samplerate of 44.1 kHz, this would be required. Another issue is how to handle volume controls in an all digital domain.
As for package. A joystick like 64DTV with an flashslot on the side and solderpads for more expansion on the inside is one suggestion.
Another is simple a simple _small_ box with the original connectors.
Which also brings up the issue of Zorro bus. Is the construction in such way that adding zorro will be realistic. Or is the internal bus in such way it's not at all compatible with any zorro stuff..?
Would the cost of useing a fpga with slightly more gates be significant ..?, was thinking of putting the m68000 into the fpga. Or adding other "hardware" in fpga in the feature.
Any news on licenseing issues .. ?
-
was thinking of putting the m68000 into the fpga.
I think Dennis already thought about this - no suitable free 68xxx cores exist - I'm not sure about licenced one though! This would be a nice solution as it would free up a lot of I/O pins for lots of great things! :-D
-
You know how it goes, as soon as it works, a new project comes up..
Like I said before....
You could do a KLF and take all the prototypes, plans, verilog files to a remote Scottish cottage and burn it all. :-D
-
@Dennis
I have bought enough parts to build a couple of boards for my own use.
Did you manage to build those boards?
Did you ever speak with Troika?
Keep up the good work! :-)
-
Hi Dennis,
Saw this link on the C1 page.
www.symbos.de/trex
Do you reckon this development board would be powerful enough to be used for your Amiga ECS cicuits?
Thank you for continuing the work on the minimiga please keep up the good work.
Regards
John
-
Hi,
Are you going to GPL the source Verilog for Minimig for others to add features on to it or are you going to take Minimig commercial in some direction ?
It most probably won't go commercial. I'm more and more thinking of releasing Minimig in some form of GPL license.
1. I Download a .zip File, extract Schematics, Firmware,
That is most likely.
Considering you started on a special board. Maybe it's good to make pcb space for a real D/A for video output such that one can get highquality video output ..?
The VGA connector with resistor DAC that I'm now using (and which will also be included on the new board) gives surprisingly good video output. You must not compare this to, for example, the DTV-64's output. Minimig outputs RGB instead of composite which makes a huge difference. Quality is as good as an A500.
I think the main point is to ease output of video in another way than the default way.
My main concern at the moment is to get Minimig finished in some form, and this usually means keeping things as simple as possible.
Conversion of audio from one clock domain to another can be accomplished by writeing output sample to a latch, and then use the other part of source clock to read it out to an outputlatch, that will provide the same sample until the next one is transfered. Ie if Amiga use samplerate of 4-16 kHz and output domain will use a constant samplerate of 44.1 kHz, this would be required. Another issue is how to handle volume controls in an all digital domain.
Using a simple latch to convert the sample rate would not improve sound quality a lot, as that is not a good way to convert audio sample rates. You would have to apply proper filtering. But, the sound output of Minimig is currently not that bad. The one-bit converter does a pretty good job and Minimig sounds about as good as an original A500.
You could do a KLF and take all the prototypes, plans, verilog files to a remote Scottish cottage and burn it all.
Well, I could also put it in a box and put it on the attic, where most of my other projects are.. :-P
Did you manage to build those boards?
Currently busy with routing...
Did you ever speak with Troika?
No.
Do you reckon this development board would be powerful enough to be used for your Amiga ECS cicuits?
The FPGA is probably big enough, but you would need to convert the design to use SDRAM instead of SRAM and you would need an 68000 in some form.
Dennis
-
The 'Vidiot' circuit used in all OCS/ECS Amigas (A1000, A500, A2000, A3000, CDTV?) is also just a resistor network - nothing wrong with that.
The high density audio DAC you've built is also a very good choice, might even be better than the original (it provides for some oversampling, somewhat similar to SACD).
I'm looking forward to future updates from you - and very much to building my own Minimig sometime in the future!
-
Dennis,
Will you sell some blank PWBs or just release gerber files so that we can build up a Minimig as well ?
Keep up the good work !!
:-)
-
@Dennis
I'm more and more thinking of releasing Minimig in some form of GPL license.
Wow! You would be, like, the Linus Torvolds/Richard Stallman of the Amiga.
1. I Download a .zip File, extract Schematics, Firmware,
That is most likely.
I actually almost started dribbling thinking about this!
But, the sound output of Minimig is currently not that bad. The one-bit converter does a pretty good job and Minimig sounds about as good as an original A500.
I can't wait!
Well, I could also put it in a box and put it on the attic, where most of my other projects are..
That will be the rough at the edges but working "Time machine", "cold fusion reactor" and "teleportation" projects, eh? :lol:
Did you ever speak with Troika?
No.
Who cares! If you go GPL, any of them can add minimig functionality to OS4, AROS or Morphos! What a gift to the Amiga community. :-)
Thanks for the update, Dennis! Good luck!
Wow! This is a great body of work! :-)
Edit: Corrected perverted subject line
-
@Dennis
You arriving on Amiga.org with this project is the best thing to happen to the Amiga in well over a decade, and I mean that sincerly!
-
@redrumloa
Yeah! I agree. Well said!
-
Here is a cool little project much like the Minimig that could also carry over to "newer" Amigas as well.
http://linux.embebido.org/ecb_at91/
-
It looks quite nice but I fail to see any relevance with Amigas, new, classic cloned or otherwise! :lol:
-
That's the problem with most people today. They look but do not see. They listen but do not hear.....
-
Another problem with people today is trying to shoot people down instead of just explaining yourself...
-
Well, if people would go to the web site( http://linux.embebido.org/ecb_at91/ ) and actually read the material instead of looking at the pictures, they'd see that this little project has a lot of relevance for all Amigas, past, present and future.
The project uses UBOOT, can anyone say A1?? It could be modified to use a PowerPC processor instead of ARM. It runs Linux at present and could run UAE...instant embedded Amiga.
So for someone to say that they fail to see the relevance of the project just shows at best a lack of imagination and at worst a lack of desire to read the material and a desire to respond without being informed.
It isn't my job to explain the projects of others or the inherent relevance, but you insisted.....
-
@ferrellsl
I think you will find the :rtfm: attitude doesn't fly here on AO like it does on Linux sites. So it may possibly run UAE? A bit of a stretch for an obvious amiga connection.
Note I'm not saying it's wrong to mention it, it's interesting.
-
I went to the site, i read it and i seen that it was an ARM based single board computer that runs linux that bears as much relevance to Amiga or Minimig as any ARM based single board computer (very little imo).
It may not be your job to explain others projects but it is your job to explain why you think these projects are relevant to a particular topic. We can't go around imagining what your opinion is.
-
Sheesh....enough with the "holier than thou" attitudes already. I was asked to explain my point. I explained my point. End of story. I'm not asking for a commentary or a flame war. If you like the project, fine, if you don't like the project, fine. I really don't need anyone's feedback. I was pointing out a project similar to MINIMIG that could be modified/adapted and looked interesting.
Most of the people following this thread were skeptics and unable to make the "stretch of an obvious connection" that MINIMIG was anything more than vaporware too. But it obviously runs classic AmigaOS. It's a miracle that Dennis continued to move forward with his project with all the "attitudes" following this thread.
-
@ferrellsl
Attitudes? Your kidding right? The only attitude I see here is your own flaming in the last couple comments. :roll:
-
I give up. The skeptics with attitude win. I'm outta here.... :crazy:
-
Aaaaw...
-
@ferrellsl
Why on earth would anyone modify an ARM based SBC to use a PPC instead?
Switching to a PPC would mean such a fundamental redesign of the board that you probably would be better off designing a new SBC from scratch.
There are also several PPC based SBC's supported by U-boot one could use instead.
Neither have your proposal any FPGA installed which is a key component on the Minimig.
A more suitable board is the ML300 from Xilinx.
http://www.xilinx.com/products/boards/ml300/
Supported by U-boot this board is based on a high end Virtex-II Pro FPGA with integrated PPC 405 core.
w
-
I didn't say anyone should modify THIS project to run a PPC processor. I was pointing out how it related to Amigas past, present and future. I think an Amiga SBC using ANY processor would be a great thing. It's the concept of an Amiga SBC using modern processors that I find intriguing. Size and cost are kept to a minimum, an area where past Amigas haven't excelled.
Past Amigas such as Amiga Anywhere (not what I really consider "true" Amigas) use ARM processors. So does this project. So it isn't such a huge stretch to see this thing running classic Amiga programs just like Amiga Anywhere or the MINIMIG.
In regard to current Amigas, there are many similarities between this project and MINIMIG and the PPC systems out there. The concepts of this Linux SBC project could be applied to an Amiga SBC in the future. I also consider MINIMIG to be a current Amiga even if it is built primarily to relive the Amiga's past glory.
-
You obviously suggested it in your earlier post.
ferrellsl wrote:
The project uses UBOOT, can anyone say A1?? It could be modified to use a PowerPC processor instead of ARM. It runs Linux at present and could run UAE...instant embedded Amiga.
IMO this SBC project is no more related to an Amiga than modifying a vacuum cleaner into running Amiga OS4. :-D
w
-
At the risk of igniting the flames further: both the vacuum cleaner and its OS would then both suck! :-o
Actually, like 99% of the remaining Amiga users, I've never used OS4, just couldn't resist the opportunity for a crap joke!
- Ali
-
@Ali
I'm with you man. People on here are way too vicious....
-
:lol:
Not exactly what I had in mind, but nice one. :-D
w
-
The Amiga SBC Nazis have spoken. No Amiga SBC for a you!
Bad Seinfeld joke..... :-D
-
@ferrells
That's the problem with most people today. They look but do not see. They listen but do not hear.....
Ah Yes, Glasshopper! :-D
So for someone to say that they fail to see the relevance of the project just shows at best a lack of imagination and at worst a lack of desire to read the material and a desire to respond without being informed.
Ah Yes. Velly solly, Glasshopper! :roll:
It's a miracle that Dennis continued to move forward with his project with all the "attitudes" following this thread.
We are not worthy! :boohoo:
-
t's a miracle that Dennis continued to move forward with his project with all the "attitudes" following this thread.
No offence, but most of the "attitude" started when you posted. To be fair, you have to admit, Dennis' project is using the 68k procssor (or at least a compatable) and an FPGA to emulate the custom chipset.
What you put up, while I actually agree is interesting, is a large stretch of the imagination. Saying that it can run UAE is valid but my 300MHz Palm can only just about emulate an A500. And converting it to a PPC processor would mosst likely need a complete board redesign as the ARM and PPC have completely different pinouts (from what I understand).
Anyways :-) Dennis have you had the proto boards done yet? :-D Anymore updates :-D :-D
-
I fail to see how the post (inserted below) had any attitude attached to it. I only followed up after some people with attitudes made some harsh comments. Seems like I'm running into the same ciritics who told Dennis his project was vaporware or just plain impossible to create a SBC Amiga.
Posted on: 2006/4/24 23:43
Re: Reconfigurable T-REX C1 development board by TerasIC
Here is a cool little project much like the Minimig that could also carry over to "newer" Amigas as well.
http://linux.embebido.org/ecb_at91/
-
Any progress, Dennis?
I'm really interested in buying one.
:-D
-
@ferrellsl
Please don't get me wrong, I didn't mean your post had any attitude :-) I meant the attitude started after your post.
Ref: posts about when dennis started, I think people were sceptical, how many times has a company said "we have a new Amiga mobo" only to have it delay after delay then not to turn up?
I just don't think that the ARM has enough processing power to be of any use as an Classic Amiga SBC (not at 180MHz). Although if someone were to port AROS to it, now THAT would be cool :-D
-edit- updated wording
-
No offense taken. I agree with you whole heartedly. Hopefully we'll have some new Amiga hardware in the near future which will decrease the level of negativity and skepticism in the Amiga community. But I really can't blame any Amiga enthusiast for being mean and nasty after all the broken promises.
It's just frustrating to see a great OS die for lack of decent hardware.
-
It's just frustrating to see a great OS die for lack of decent hardware.
Lack of hardware isnt the problem, hardware is everywhere.
If Hyperion had bit the bullet and ported to x86, even just a few select configs, we'd all be talking about how great our new OS4 boxes are and they'd be looking at sales beyond a cuppla' thousand units...
(flame me, meh)
-
Well said coldfish.
-
@ferrellsl
Nothings wrong with your attitude in the first post, but when people start questioning your suggestion, you choose to jump down in your trench shooting at everyone instead of providing a reasonable explanation why you think the forementioned project is related to Amiga.
Later on you characterize us as SBC Nazis and sceptics that thrive on tearing down new initiative.
If an idea can't survive a couple of critical questions, then usually it's not such a good idea as one might think.
w
-
@mr_white
Dude, you need to seriously look into buying a sense of humor.
-
Okay, So now I'm a SBC Nazi, a sceptic that thrive on tearing down new initiative and a dude with no sense of humor.
Anything more to add?
Again you choose to respond by characterizing other members.
w
-
Hey, I'm sorry if I've offended you. I noticed that in your posts you appear to be very knowledgable when it comes to hardware. I also noticed your Kick ROM post....anyway, in your opinion, how difficult would it be to design an Amiga SBC using a PPC processor? Could it be designed in such a way that the buyer/kit builder could flash his own ROMs to this system and avoid the whole OS licensing problem? Would it be affordable or at least competitive with a PowerVixxen? I suppose one could even provide sockets for classic Amiga Kick ROMs????
I know, a lot of dependencies, but I thought I'd ask.
-
ferrellsl wrote:
Hey, I'm sorry if I've offended you. I noticed that in your posts you appear to be very knowledgable when it comes to hardware. I also noticed your Kick ROM post....anyway, in your opinion, how difficult would it be to design an Amiga SBC using a PPC processor? Could it be designed in such a way that the buyer/kit builder could flash his own ROMs to this system and avoid the whole OS licensing problem? Would it be affordable or at least competitive with a PowerVixxen? I suppose one could even provide sockets for classic Amiga Kick ROMs????
I know, a lot of dependencies, but I thought I'd ask.
Why, Why why why why why why why... and thrice WHY would anyone consider using a PPC cpu in 2006... I have to know, becuase it seems to be the most pointless excercise in witless hardware design.
Once does not choose a dead CPU for a new design! If this thinking had been around at the time of the original amiga... The A1000 would have used a 6502...
-
Because they are CHEAP, CHEAP, CHEAP, CHEAP.......And PPC chips aren't dead. Apple still makes plenty of computers with PPC chips......
And because it wouldn't require that the OS be rewritten to support a different processor. Look how long it's taken AROS to get up and running on x86 processors. Valiant effort, but lengthy.
And face it, AmigaOS and Amiga hardware have been relegated to the land of the hobbyist. All those who have visions of a great Amiga return using an x86 or other processor are deluded indeed!
As for witless hardware design, why would anyone design an Amiga SBC computer with an FPGA and a 68000 CPU to run classic AmigaOS 1.3? Because it's a fun hobby. Just ask Dennis, the guy who started this thread and the MINIMIG. If you want to delve into AmigaOS for x86, go look at AROS.
Do you want to wait another 10 years for AmigaOS 4 to run on an x86? I don't.
-
Making OS4 to run on x86 wouldn't require to rewrite it from scratch, you really think that it's 100% pure PPC assembly? Beacause that's the only case in which it would be neccessary to rewrite it.
Those who want OS4 for x86 don't want it beacause they want to compete with Windows or something. That would be crazy. OS4 for x86 means access to a cheaper and instantly available hardware. Some say that this would require a lot of time to provide support for all these different kinds of x86 motherboards but who said that it would have to support all of them at startup? Linux didn't. But even one supported x86 MB would be a better HW situation than what we have now. I also cannot read the posts that say: "noone would use OS4 on x86 beacause they can have Windows with all the software they can imagine". None of them have ever said what would make those people use OS4 when it is for proprietary PPC hardware that's expensive and not even available. However, some of the current Windows users would be able to try OS4 at a low cost without need to buy a new computer.
Wow, huge off-topic, sorry for that.
-
No, a 100% rewrite would not be necessary, but it would still entail rewriting a significant part of the OS. Honestly, I'd love to see AmigaOS 4 running on the latest x86 hardware, but I doubt it'll ever happen as the current owners of AmigaOS are using the old Apple sales model of licensing each copy of the OS to a closed hardware platform. You'd think that they would have learned by now that if it didn't make Apple a huge success, then it won't work for them either.
Maybe they'll take a hint from Apple's latest foray into the x86 world and develop an x86 version of AmigaOS 4.
I know, I'm dreaming or on drugs or something....... :-D
-
ferrellsl wrote:
Because they are CHEAP, CHEAP, CHEAP, CHEAP.......And PPC chips aren't dead. Apple still makes plenty of computers with PPC chips......
No they don't. As of NAB (last week) the 3rd of 4 Apple platforms has moved to Intel, they only have to move the big desktops and the G5 desktops are basically done being manufactured, so it will happen soon, I would not be surprised to see the Desktop announcement at the next big show. Sometime in June (July at the latest), the last PPC mac will be get shipped.
-Tig
-
I would not be surprised to see the Desktop announcement at the next big show. Sometime in June (July at the latest), the last PPC mac will be get shipped.
Apple have long term contracts, if they specify PPC machines Apple have to ship them. To that end they have a contract with freescale for CPUs to be supplied until 2008.
This wont be done publicly but Apple will be shipping PPC machines for a long time yet.
-
minator wrote:
Apple have long term contracts, if they specify PPC machines Apple have to ship them. To that end they have a contract with freescale for CPUs to be supplied until 2008.
People will not be buying PPC parts for Apple computers past this year. Anyone telling you otherwise is wrong. Apple doesnt buy parts from freescale since they dont manufacturer there computers, we buy the parts for the computers, as do Solectron, Celestica and Flextronics (some of which may not make Apple anymore, but all of which did at one time.
This wont be done publicly but Apple will be shipping PPC machines for a long time yet.
No, they really won't. The last of the PPC macs are being manufactured or may have already been manufactured. Across the EMS companies who make Apple, all the lines are being converted to the intel boards, its not in any way cost effective to keep a PPC line or two going at this time. The parts are more expensive and the volumes are even lower. Noone wants them except at the cut rate prices they are no offering to clear PPC inventory.
-Tig
-
What about everyone who paid for Applecare? if their CPU fries and its covered by Applecare, Apple is not going to tell them to get an X86 box.
No, Apple is going to need to buy PPC CPUs and assorted PPC Mac components for a couple years still. Not in as large a volume as right now, but enough to be able to fulfill their extended warranty contracts.
Apple is not Eyetech. CPU modules are not mythical things normally found in dragon hoards.
-
How much of available gates on the SP3-400 fpga have you used up?, thinking if there's space left to implement an 68k as verilog code.
The FPGA is probably big enough, but you would need to convert the design to use SDRAM instead of SRAM and you would need an 68000 in some form.
Regarding memory.. what about putting some sdram socket such that one can make use of lot's memory for graphics programs or the alike..? the socket don't have to be supported. Just be physicaly wired such that anyone that feels inclined to code could "upgrade" without needing to make another pcb.
Any estimates what fpga size would be needed to implement A4000T or A3000 .. ?
-
Everytime I see this thread posted, I look at here with hopes of seeing dennis post some update about the MiniMig (remember the original thread anyone ?). But all I see is ppl flaming eachother about nonsense and beating dead horses. If you have some respect for other forum members or the thread starter, please do your off topic talk in relevelant threads.
-
I agree, for historical purposes it would be nice if this thread was cleaned from all not mini mig realated chat.
We do not want to keep such nonsense anyway and makes the lot alot better readable.
-
CodeSmith wrote:
What about everyone who paid for Applecare? if their CPU fries and its covered by Applecare, Apple is not going to tell them to get an X86 box.
First of all Apple doesnt even do apple care, another company does. As part of Apple care a number of replacement parts have been set aside, however I will point out that even before the switch if you fried an older computer, you had a really good chance of getting a faster computer returned to you if they did not have the same speed parts available, and there is strong encouragement on a DOA board for you to upgrade to a new computer at a discount, in which Applecare will be pushing the intel macs.
No, Apple is going to need to buy PPC CPUs and assorted PPC Mac components for a couple years still.
Again, I will point out Apple doesnt buy parts for computers at all, Applecare is handled by another company who has gambled they have enough replacement parts for there needs. If they start running out of PPCs, they may order more, but that not apple, that would be low quantity, and it has nothing to do with Apple still making PPC computers. Apple isnt ordering anymore PPC computers from its manufacturers, there time is over at Apple.
-Tig
-
CodeSmith wrote:
What about everyone who paid for Applecare? if their CPU fries and its covered by Applecare, Apple is not going to tell them to get an X86 box.
First of all Apple doesnt even do apple care, another company does. As part of Apple care a number of replacement parts have been set aside, however I will point out that even before the switch if you fried an older computer, you had a really good chance of getting a faster computer returned to you if they did not have the same speed parts available, and there is strong encouragement on a DOA board for you to upgrade to a new computer at a discount, in which Applecare will be pushing the intel macs.
No, Apple is going to need to buy PPC CPUs and assorted PPC Mac components for a couple years still.
Again, I will point out Apple doesnt buy parts for computers at all, Applecare is handled by another company who has gambled they have enough replacement parts for there needs. If they start running out of PPCs, they may order more, but that not apple, that would be low quantity, and it has nothing to do with Apple still making PPC computers. Apple isnt ordering anymore PPC computers from its manufacturers, there time is over at Apple.
-Tig
-
[size=9]First I was going to post an off-topic comment about my humble opinion, that the current developing AmigaOS should not be ported from PPC to x86, but instead of that you should keep it PPC code, and to companion it you should make a PPC computer emulator for the x86 IBM-PC architechture which would give us millions of PPC platforms to run PPC AmigaOS on.[/size]
But on-topic questions, that Dennis can answer:
What kind of computer is used to compile the MiniMig sources? (How much RAM? Which version of Xilinx software?) How long does it for the computer to process the design from sources into a downloadable bit-file?
-
@Dennis
Good Luck!
I hope you are having success with the minimig board. :-)
I understand if you have better things to do (Even Holland can have nice weather at this time of year!)
Please report back here when you have an update for us.
Thank you for all your efforts so far! :-)
-
Wow, I have been away for a while and 2 pages of new posts! Not all Minimig related however :-? but this is Amiga.org :lol:
I have to admit progress has been slow lately. I have been working on the layout and expect to send the design to the PCB manufacturer in a couple of weeks.
What kind of computer is used to compile the MiniMig sources? (How much RAM? Which version of Xilinx software?) How long does it for the computer to process the design from sources into a downloadable bit-file?
I am using a Shuttle barebone with a prescott pentium4 @ 3GHz and 1Gbyte ram. I am running Xilinx webpack version 6.3.03i. From source to bin file (Minimig uses bin files which are bit files without a header) takes about 2 minutes. I actually bought this system especially for this project. Before this I only owned a compaq laptop (duron 900MHz / 256Mbyte ram) and offcourse my 3000. Compiling on the laptop was not fun, especially when I had some PDF's open on the background because then it would start swapping. The shuttle however is perfectly adequate for this project.
Dennis
-
@ferrellsl
No offence taken, but in general I think we all get better allong if we discuss issues instead of persons. :-)
Answering your question about the hardware it shouldn't be any problem designing an Amiga PPC-based SBC.
It could also be done avoiding the licensing issue. However, this is hardly an issue for potential Amiga System providers. Providing a system which which undermines the license programme would be financially downright stupid.
Reason for this is:
- You risk being sued down your a*r*s*e by Ainc for not obtaining a license.
- You extend development time/costs due to the fact that you have to hack the OS to find out how it works and you will have no support from Hyperion.
- A lot? of potential customers will not buy your product either because they are discusted by your immoral behaviour against Hyperion/Ainc, don't take the risk of buying your product since support is limited and uncertain.
At last would you buy a product which you have no guarantee will run the OS you expect it to work with?
Doesn't seem very tempting does it?
@Dennis
Getting back to the topic: Can you tell us more about which development tools, hardware and software, you've used for this project?
w
-
Wow, I haven't read the forums here in a LONG time and seeing all the Minimig talk and pictures is very cool! Nice to see someone has finally done this.
Dennis, it's only a matter of time till we see one of these in a joystick :roll: so why don't you make a deal where you can sell it now and still release it GPL in the future?
Amiga Inc would probably love to make any dollar they could, just don't expect any quick email replies. When I requested a dev kit from them it took something like 6 months before they replied and offered me a free one. I'd switched email accounts by then and didn't even see it till just the other day.
Ever consider adding an LCD controller and a small display. It would make a kick @$$ handheld video game.
-
Not sure if Amiga Inc actually own the rights to the classic Amiga chipset anymore. Do they?
It would make a kick @$$ handheld video game.
Yeah right... Not. Although I love the Amiga, it's too old, too crusty to compete with anything except the Sega Nomad. In FPGA form (After you include the LCD driver chips, the screen etc.) probably too expensive?
-
alexh wrote:
Not sure if Amiga Inc actually own the rights to the classic Amiga chipset anymore. Do they?
Gateway kept patents but I think some of them expired anyway. It depends on when Congress extended them.
It's the OS copyrights I'd worry most about and Amiga still owns that.
It would make a kick @$$ handheld video game.
Yeah right... Not. Although I love the Amiga, it's too old, too crusty to compete with anything except the Sega Nomad. In FPGA form (After you include the LCD driver chips, the screen etc.) probably too expensive?
First of all, most of the LCD driver logic could be integrated into the FPGA. LCD is all digital logic and needs no external D/A conversion like a monitor. It shouldn't need really complex logic. It may even allow the simplification of parts of the miggy chipset.
The 68K could be rebuilt and go into the same gate array. Many of the existing CPU cores try to duplicate timing exactly, aren't very efficient and are way bigger than they need to be. For example, the z80 core on OpenCores (T80?) is something like 40% larger than a new one floating around that's just as compatible. It *might* require a larger FPGA to include the CPU but if a company is producing that many it would be much cheaper to produce a custom chip than use even the smaller FPGA.
It would be no challange to the 3D PSP or other high end handhelds. More like a gameboy killer.
I guess I could just buy a GP32 and emulate it.
:-D
-
I'd like to see any non MiniMig stuff deleted or moved from this thread. I'd support moving previous ff-topic stuff too.
And I'd like AGA support instead of just OCS. :-) But I'll give you time... :-)
-
Hi, all
I made a page on Wikipedia.org for the Minimig. It can be accessed here:
http://en.wikipedia.org/wiki/Minimig
I'd like those who know stuff about this project to fill in info on it on that page. That is better than having to read through pages and pages of posts here on Amiga.org. There are no restrictions as you don't have to make an account. Just press Edit to add some new information.
About the Minimig itself, this is a very important piece of hardware. I don't think many people realize how important it is in non-Amiga related terms: It could be used for embedded hardware that sits currently without graphics display capability. There is a huge market for that. The question is how cheaply could it be mass produced? If it can be done cheaply enough, there ought to be a huge market in the very-low-cost area of graphics display hardware.
I'm not thinking about games alone, but simple devices that need to display graphics with a certain amount of graphics power on a small QVGA sized display.
I seriously hope this can be commercialized and made into mass production, how cold and greedy that may sound. :-)
It would show that there is plenty of life left in the technology and who is to say that it can't be developed into a full AGA system or even beyond that? The technology is very mature and there are many people who know how to program it. It's a clear advantage.
Dennis, I hope you make it. Good luck, man.
-
the Minimig itself is a very important piece of hardware.
Yeah? Great work by the author, I think it took a lot of dedication for him to get this far. With home projects like this, as soon as they start to show signs of life they get boring very quickly and you get itchy to move to something new.
Minimig in my opinion is creative, nostalgic, very innovative and technical design (usage of the Sigma Delta Modulator) but important...not too sure I would call it important.
I don't think many people realize how important it is in non-Amiga related terms: It could be used for embedded hardware that sits currently without graphics display capability. There is a huge market for that.
Erm, no there isnt (talking as an ASIC engineer IN the embedded hardware market) If I wanted a gfx display for my embedded system (assuming I dont have PCI or some other interface where I could connect an off the shelf gfx solution) I'd just use the free, VGA controller from OpenCores.org (http://www.opencores.org/projects.cgi/web/vga_lcd/overview) As it (naturally) supports the standard VGA register set it's much easier to program for (and less powerful, but who needs power?) VGA drivers already written for Linux work out of the box etc. If I need more power I a bitblt.
If I am in a REAL embedded market, and I am going to ASIC, and I have a true memory overhead, I would use a line based sprite renderer "borrowed" from one of the Arcade FPGA projects. Only needing 2 lines of memory instead of 1 (or even 2) frames of memory has significant impact on RAM usage and bandwidth.
-
This is fantastic! I remember talking about this a while ago and a few people being skeptical. I always thought it could be done.
Any chance of making an AGA version?
As for copyright issues, if it's ok to emulate the hardware in software, why is it not ok to emulate it in different hardware? I mean it's not like you're copying the actual circuits.
-
Tricky wrote:
Any chance of making an AGA version?
As for copyright issues, if it's ok to emulate the hardware in software, why is it not ok to emulate it in different hardware? I mean it's not like you're copying the actual circuits.
As said before the current focus is to get plain A500 with OCS working properly. Other stuff is nice, but have to wait..
There might be difference in copyright between sw and hw emulation. Would be nice if someone filled in on this.
henrikmk, I updated your minimig page.
Dennis! & others:
There's exist an MC68000 verilog source:
http://www.opencores.org/projects.cgi/web/k68/overview
Source code example:
http://www.opencores.org/cvsweb.cgi/~checkout~/k68/rtl/verilog/k68_decode.v?rev=1.1.1.1;content-type=text%2Fplain
This means it may actually be feasable in short term to rid of the physical MC68000 incarnation.. (depending on bugs :)
-
@freqmax
Dennis! & others:
There's exist an MC68000 verilog source:
http://www.opencores.org/projects.cgi/web/k68/overview
This have been mentioned earlier in this tread.
Other 68k IP cores, more feasible to implement technically, exist though.
C68000 (http://www.evatronix.pl/products/docs.html?id=2&product=QzY4MDAw)
D68000 (http://www.dcd.com.pl/acore.php?idcore=20)
w
-
The free 68K IP cores I've seen were still incomplete last I checked and you'd need to check the licensing terms for the commercial ones before considering them.
-
Dennis, remember to keep us updated! :D
-
This have been mentioned earlier in this tread.
Other 68k IP cores, more feasible to implement technically, exist though.
C68000 (http://www.evatronix.pl/products/docs.html?id=2&product=QzY4MDAw)
D68000 (http://www.dcd.com.pl/acore.php?idcore=20)
Which ones could you get the source for tonight ..?
Which cores would add constraints on redistribution .. ?
Ofcourse the one found in opencores.org maybe isn't the state of art, but it can be debugged and made to work. Sure beats starting from scratch or license headache.
-
-
@freqmax
Which ones could you get the source for tonight ..?
Don't necessarily need the source code. A protected IP block would suffice. I don't see why you need it tonight, but if you are in a hurry why don't you contact any of them and ask?
Which cores would add constraints on redistribution .. ?
At this point there are no other usable cores to redistribute. If you are in a hurry, why will a half-finished core full of bugs do you any good?
Ofcourse the one found in opencores.org maybe isn't the state of art, but it can be debugged and made to work. Sure beats starting from scratch or license headache.
We don't know that, Mabe the core needs to be fully redesigned in order to work. Licensing doesn't necessarily have to be a headace. It's all dependent on the terms of license and those providing it.
w
-
@Fransexy_
How fast could be the emulated 68k on actual FPGAs? 20 mhz? 40? 200? more?
The 68k would not be emulated it would be synthesized, wich basically means it would work just as an actual 68k.
Latest Virtex-5 65nm FPGA's can run on clock speeds at 550MHz.
w
-
They can be a single FlipFlop at 550MHz
Anything complex and the speed is reduced greatly. 10's of MHz is almost certainly possible. But these Virtex chips cost a *lot* of money especially compared to MiniMig's spartan II and a real 68k!
-
MiniMig is a great proof of concept. It could be invaluable if enough people with vision and some investment money helped out. I'd give my left nut for a handheld or laptop Amiga (1536x620 display please). I also have some investment money it it were a safe venture.
[[With home projects like this, as soon as they start to show signs of life they get boring very quickly and you get itchy to move to something new.]]
Which is why I'd like to see it evolve to encompass a very fast 680X0 (either as actual processor or in FPGA), but more importantly, evolution to AGA status and beyond. Since 680X0s exist as readily-available chips, I see AGA compatibility as a more needed direction. I'd like to see the 68020 as the minimum processor (than the 68000). Maybe the MiniMig could ultimately have a processor interface that allows the user to select different processor boards (maybe exploiting those trap-door processor accelerators out there). And then farther down the road... PA Semi has superfast and cool PPC chips available by year's end IIRC.
>not too sure I would call it important.
Go away asshat. If you're not interested, just go away. Sounds like bitter dregs and negativity to me. The MiniMig is the closest the we have to a real "new" Amiga.
>There might be difference in copyright between sw and hw emulation
Let's not even go there. The "people" at Amiga Inc should be kissing Dennis's butt right about now. Anyway, he's using verilog code IIRC. So it is software. And he's put together the hardware blocks to run it on. You install KS/WB yourself. If UAE is legal (sans OS), so is this.
Speaking of the dynamic duo at Amiga Inc., when is somebody going to go over there with some nooses, torches and pitchforks? What are the DOING these days anyway? Besides hogging the Amiga OS that is?
-
Ok, I'm responding to several posts here.
The freely distributable 68K cores vary in how complete they are a great deal. The one I first looked at was completely missing the supervisor mode. One on open cores looks almost closer to being a coldfire than a 68K but I haven't looked it over in detail. If you had lots of free time you could probably get it working but you may spend a lot of time rewriting the existing code.
The fastest I've seen people mention testing 8 bit cpu cores at was around 40MHz max. And that seems to depend on a very good design since not all of them run that fast. If built into a custom ASIC at least one simulates to well over 200MHz. With the complexity of the 68000 instruction set you'd need a pretty good design to get that kind of speed. You might be able to run a higher MHz with a pipelined design but when you start doing that you need cache, a cache controller, instruction preloading and branch prediction to keep the pipline from stalling. Then the FPGA size goes up, the cost goes up, the development time... you get the picture.
I don't think the combination of a licensed core + added size of the FPGA would be very economical but it would be doable. I don't think you could sell many all in one joysticks at the pricepoint that would give you. If you could move it to a custom chip for manufacturing and produced it in sufficient quantity it wouldn't be quite as bad. I'm guessing it might still be too high for a Joystick game system. A laptop would be doable but it wouldn't even compete with Intel laptops that are several years old.
-
Which is why I'd like to see it evolve to encompass a very fast 680X0 (either as actual processor or in FPGA), but more importantly, evolution to AGA status and beyond. Since 680X0s exist as readily-available chips, I see AGA compatibility as a more needed direction. I'd like to see the 68020 as the minimum processor (than the 68000). Maybe the MiniMig could ultimately have a processor interface that allows the user to select different processor boards (maybe exploiting those trap-door processor accelerators out there). And then farther down the road... PA Semi has superfast and cool PPC chips available by year's end IIRC.
I think at least a 1200 clone would be what to shoot for in a Joystick game.
If I were to do a CPU in verilog I'd go for an EC060. If I remember right (and I might not) the 060 has the smallest instruction set to duplicate and is still supported by the OS. I also think it lacked the math functions but I can't be sure.
-
I think you're right that the EC wasn't supposed to have FPU or MMU (though there was some fun with some chips labled with EC having both of these). However, for a joystick Amiga, it's just for games, so wouldn't the incompatibility issues with an '060 be a bad bet?
-
Oliver wrote:
However, for a joystick Amiga, it's just for games, so wouldn't the incompatibility issues with an '060 be a bad bet?
Agreed... Wouldn't an '030 be better?
- Ali
-
Also without the MMU you cannot use MMULib and MuRedox and would make it harder to get all the whdload installs going. An EC030 would make more sense.
I took a look on the freescale website but I found the available 680x0 cpu's rather expensive. The 68000's are affordable but an 68060 is crazy!
-
Ouch! Even the '030s are expensive!
- Ali
-
Registered simply to post this and waffle on a bit :)
To Dennis - Fantastic thread. Reading thru all 30 pages (but not everyones posts) i see the fans, the sceptics, the proof, the photos, the progress along the way and a lot of techie stuff that i don't understand much of, and it makes great reading. The project is very nice and a finished product would be highly sought after. I see in the last couple of pages that your drive for this seems to be withering now which is something i see in so many coding projects. A great shame after so much work. Its always those final tidyup steps that get boring but you must be determined to finalise it. The pleasure of having a fully finished working project will be great.
Loads of specs have been thrown about and a lot of it makes no sense to me. DAC this, timing that, FPGA registers, PIC bumpf etc etc. Heres my simple questions requiring simple answers.
Will it plug into a TV ? And will it plug into 14", 15", 17" or 19" (and above) standard svga monitors either crt or tft ?
You said that when you put it onto "your own board" that you'll add more memory. Does this mean more memory usable by the "amiga" meaning more chip or fast ram ?
Could 1gb (or bigger) memory cards be used to store larger amounts of adf files ? Is the OS stored on the memory card too ? IF it was commercial could the OS be hardcoded with a removable memory card being for games only. Would there be some way to connect it to the pc for dumping different ADF files onto the card or upgrading the OS etc. Maybe having the OS in a rom and allowing upgrades (much like a bios upgrade).
You say that you will likely go open source with pdf schematics etc. Sigh. Sounds like someone else will snap this up and make a {bleep}load of money from it then. Foolish i think, but its your decision. I'd seek legal advice or even take up (or talk to) some of those offers you've had. Man, you've got 30 pages of interest here in a niche corner of the web. Get some marketing behind your project and you've got thousands of sales. Get together with the guy doing the similar ST thing, make yours support both and sales could be massive !! You've done all the hard work now, now you should be concentrating on reaping some rewards.
I'm sure theres plenty of people here, and elsewhere with some good marketing, that would love to reward you for this work.
Thanks for a project which has made very intersting reading and i really do hope you finalise it and get some sort of marketing. Even if you sell the project for a lump sum and then wash your hands of it, at least a company could then produce a product we could all use.
My hope would be for a type of mini amiga with swappable memory cards and firmware upgrades and NOT a joystick type thing.
Good luck.
Hitm4n.
-
at least a company could then produce a product we could all use.
I wouldn't bank on it! :-(
If he sold it to Amiga Inc, they would do nothing with it, frighten off any potential partners, and their cheque to Dennis would probably bounce. :-?
Then to rub salt in the wounds, they would bring out another zero value added version of Intent and call it MiniMigDE. :boohoo:
At least if it were open sourced or GPL'd someone competent would pick it up and produce product.
However, I (Humbly)agree that Dennis should finish this and make a nice packet from it. :lol:
Maybe he could ask Dammy to make minimig an AROS bounty and agree to release minimig under the AROS license when a predetermined sum has been raised?
-
Then he shouldn't sell it them !! Making it open source means someone else will probably profit from it when he's done all the hard work, not good in my opinion. Theres plenty of companies that can gain required licenses and make a product that Dennis feels is what the customer wants. Be more involved in its development to market.
-
Hi, Dennis
There is any status update?
:roll: :popcorn:
-
Oliver wrote:
I think you're right that the EC wasn't supposed to have FPU or MMU (though there was some fun with some chips labled with EC having both of these). However, for a joystick Amiga, it's just for games, so wouldn't the incompatibility issues with an '060 be a bad bet?
Screwed up the post totally so here is a 2nd attempt.
For a Joystick A1200 clone and EC020 at 14MHz is all you need.
Actually, you could have a 68000 with a 32bit buss at that speed, just run the system RAM as fast.
That should be doable.
It's when you start getting 40MHz FPGAs or 200+MHz custom ASICs that you have to start optimizing things.
We are crossing 2 different goals in the discussion.
-
Doppie1200 wrote:
Also without the MMU you cannot use MMULib and MuRedox and would make it harder to get all the whdload installs going. An EC030 would make more sense.
No argument there but for a joystick it's not needed.
For a desktop/laptop you are better off with and external CPU which could be off the shelf or custom.
Designing a full blown 030/040/060 would be a big jump from using an existing 68000 core. *If* you could get something running at 200MHz it won't matter if it's an 060. Cost would definately matter to produce a part that fast.
I took a look on the freescale website but I found the available 680x0 cpu's rather expensive. The 68000's are affordable but an 68060 is crazy!
If you think that is bad check the number of gates in an 060 and see what an equivalent FPGA will cost.
-
jdiffend wrote:
For a Joystick A1200 clone and EC020 at 14MHz is all you need.
Actually, you could have a 68000 with a 32bit buss at that speed, just run the system RAM as fast.
That should be doable.
Actually, I see some of the commercial cores are capable of 65MHz but that is in a freakin expensive FPGA.
On low cost FPGAs it would be around 30 which means 28 if you are using a multiple of the standard Amiga clock. That's more than fast enough for the joystick and it might suck batteries. If you could program the clock multiplier on the fly you might be able to save batteries on games that need less speed and still keep up with games that need more.
But the more complex the design the more expensive to produce.
-
I agree an 060 in an FPGA is far fetched. I was not suggesting that. But Dennis choose an 68000 wisely. The price per CPU is ~$2,- But for an 060 the price skyrockets!
I think ~$300,- per (obsolete) CPU is just too much. I think a minimig would benefit from having a standard A1200 compatible slot. Existing accelerators could be fitted.
-
@Doppie1200
If people want a better CPU than 68000, using a Coldfire chip would be much cheaper than a real 680x0. They are almost compatible and there's code available that emulates the missing operations. Dennis' decision to go with the 68000 also made the design easier, with fewer things to go wrong. It's usually best to start simple, and make things more complicated once you know that things work. Now that he's demonstrated it working, people can think about more advanced CPUs, AGA etc.
Hans
-
Hans_ wrote:
@Doppie1200
If people want a better CPU than 68000, using a Coldfire chip would be much cheaper than a real 680x0. They are almost compatible and there's code available that emulates the missing operations. Dennis' decision to go with the 68000 also made the design easier, with fewer things to go wrong. It's usually best to start simple, and make things more complicated once you know that things work. Now that he's demonstrated it working, people can think about more advanced CPUs, AGA etc.
Hans
And nobody has provided a working Coldfire solution yet.
Sure it's possible but odds are it won't be playing any of your old games.
-
jdiffend wrote:
And nobody has provided a working Coldfire solution yet.
Sure it's possible but odds are it won't be playing any of your old games.
According to this (http://www.microapl.co.uk/Porting/pr981015.html), all main 68k processors can be emulated by CF68klib in both user and supervisor modes so it should be possible to run even the old hardware banging games that take over the entire system. I thought that the emulator might be user-mode only but it's not.
Of course, as you said, no-one has released a working coldfire accelerator yet. There have been a few abandonned projects and other "spare time ones" that will take a long time to happen. There's definitely more work to be done using a coldfire chip than using a real 68k chip. For the price difference (and the higher processing power), it might be worth it though.
Hans
-
@Hans
Agreed, it is ofcourse possible to migrate to higher lever CPU's regardless of the fact them being compatible or not. What I'm trying to point out is that a next version of minimig would benefit from an A1200 compatible CPU slot because:
1) CPU extentions (020/030/040/060/PPC) are available 2nd hand (=cheaper then another custom board)
2) Smooth migration path because existing hardware can lift the minimig to other levels
3) The interface offers enough to access the complete chipset so other CPU boards can be designed by those creative enough
I understand that this has some drawbacks;
1) Rather strange connector (nothing like used on modern hardware)
2) Limited 24bits address bus available
3) Limited bandwidth (due to exotic connector)
This is me dreaming this time. Just a thought.
-
Hans_ wrote:
According to this (http://www.microapl.co.uk/Porting/pr981015.html), all main 68k processors can be emulated by CF68klib in both user and supervisor modes so it should be possible to run even the old hardware banging games that take over the entire system. I thought that the emulator might be user-mode only but it's not.
Yes but at a much reduced speed. Emulating a 68000 at a few MHz is different than emulating a game that requires a 14MHz A1200.
Of course, as you said, no-one has released a working coldfire accelerator yet. There have been a few abandonned projects and other "spare time ones" that will take a long time to happen. There's definitely more work to be done using a coldfire chip than using a real 68k chip. For the price difference (and the higher processing power), it might be worth it though.
Hans
See the lengthy discussion of this topic in other threads.
-
Yes but at a much reduced speed. Emulating a 68000 at a few MHz is different than emulating a game that requires a 14MHz A1200.
Surely a 166 MHz or 200 MHz 4e coldfire can do better than that.
Hans
-
Hans_ wrote:
Yes but at a much reduced speed. Emulating a 68000 at a few MHz is different than emulating a game that requires a 14MHz A1200.
Surely a 166 MHz or 200 MHz 4e coldfire can do better than that.
Not for full emulation and hardware traps aren't 100% compatible. Go read one of the Coldfire related topics.
-
Not for full emulation and hardware traps aren't 100% compatible. Go read one of the Coldfire related topics.
For the few games that don't work, they could be patched directly. Or the 8Mhz 68k fully emulated in software by the 200Mhz coldfire for complete compatibility.
-
I ran into a film editor at the cafe the other night and I asked him about video editing software for Wintel. I don't recall how it came up but he asked what I used. I said i had at least one of everything, Wintel, Mac, Amiga and others.
He said he always hears great things about the Amiga. He then asked if there was ever an Amiga laptop.
I said there were projects like PAWS, but that the MiniMig appears to be the the best possible answer to any devlopment there. It's small, and the pictures i've seen imply it's modular enough that it could take some interesting physical configurations. Laptops, maybe even bulky PDA.
I think having the modularity to allow later MiniMigs to exploit existing A1200 processor cards would be a great idea. Batteries are much better than they were a couple of years ago, so the major remaining puzzle pieces would be LCD support (I really think the AGA hardware maximum of 1536 pixels should be aimed for). Either we have to get LCDs that support PAL/NTSC resolutions and timings (Do any 1600x1200 displays support NTSC timing?) or there would have to be a way for the later MiniMigs to support SVGA timings (so we could use common LCDs) while still allowing real Amiga resolutions.
On Edit: The ECS had the same super-hi-res limit as AGA, but with only 4 colors.
-
DoubleDutc wrote:
For the few games that don't work, they could be patched directly. Or the 8Mhz 68k fully emulated in software by the 200Mhz coldfire for complete compatibility.
Few? Do I really need to point out that most games are copyprotected and patching directly is going to take some considerable effort?
-
i've been following this since january, and i'm still grinning from ear to ear. thank you dennis, and i wish you the best of luck in completing your project!
however, although i dream of someone taking the amiga hardware to where it should have gone, but for crappy circumstance, this project is definately ambitious enough as it is. also, it's rad enough as it is. sure, i could buy an A500 off of ebay to play games and demos on, but it is such a pain in the bum getting the ADFs on it. a clone with a flashcard drive would be so much better.
i wondered if maybe the original idea was to build a flashcard replacement for DF0: and it took off from there?
-
boing wrote:
I ran into a film editor at the cafe the other night and I
..
He said he always hears great things about the Amiga. He then asked if there was ever an Amiga laptop.
..
there. It's small, and the pictures i've seen imply it's modular enough that it could take some interesting physical ..
years ago, so the major remaining puzzle pieces would be LCD support (I really think the AGA hardware maximum of 1536 pixels should be aimed for). Either we have to get LCDs that support PAL/NTSC resolutions and timings (Do any 1600x1200 displays support NTSC timing?) or there would have to be a
..
As always whenever there was graphics, and esp tv related amiga had real power. FPGA are powerfull enough for a complete system.
Lcd support is proberbly not very hard. It doesn't got so tight restrictions on timeing asfaik. As long as your updates are faster than 10 fps. So it should be able to take pal/ntsc timeings.
One thing that will cause problems thoe is rescaleing!, if lcd native resolution doesn't match video output there is trouble. So either one have to use parts of lcd or dvelve into cosinus transform land and high speed dsp.
As minimig outputs digital video interfaceing should be a piece of cake. Provided one can live with only useing say 1536 pixels out of 1600.
Another approach is to add lcd native resolution as a native video mode into amiga. Should not be hard in the hardware sense. But to get amiga sw to handle it and esp os might be another ballgame :)
-
Hello Dennis,
we eagerly await being able to purchase one of your minimigs!
i would like to add my own hypothetical tidbit to this debate, maybe to supplement the ocs chipset that dennis has reproduced , we could have a picasso96 or cyberGFX compliant graphics chipset to supplement the gamingplaying abilities of the minimig with more serious workbench orientated stuff.
Another idea is to have a competition to create a new amiga minimig case with the spirit of amiga in the design.
i go now
-
nadoom wrote:
Another idea is to have a competition to create a new amiga minimig case with the spirit of amiga in the design.
Cool idea. Though, mine will be going in an A1000 case :-)
--
moto
-
(http://www.bb-shopping.com/prod_images/image_cat39/396_image1_keyboard.gif)
This would be a cool case, if it didn't have those compartments, and it was just one huge compartment.
-
Dessimat0r wrote:
This would be a cool case, if it didn't have those compartments, and it was just one huge compartment.
A good sharp knife would make short work of those, they're only plastic I take it.
-
do you reckon we could squeeze minimig into an old external amiga 1000 external disk drive? now that would rule!!
-
motorollin wrote:
Trust me, price will not be an issue. Come on, I've just spent over £500 building an A1200 system!!! I'm sure the cost of manufacturing and distributing Dennis' Minimig wouldn't even come close to that, allowing a healthy profit margin and still sell at a price Amiga enthusiasts would welcome!
moto
Wow, impressive amount of Dosh there you spent, and i thought my buying an A4000/CS060/Scandoubler for Au$700 was bad enough (£250), and an A1200/33Mhz 030/FPU/20GIG HDD.
[edit] oh, just realised i just repied to a post from last year :(
-
whiteb wrote:
Wow, impressive amount of Dosh there you spent
I have spent a lot more since then... :-) Mediator, Blizzard 060 et al do not come cheap!
--
moto
-
@Dennis: :bow: Awesome work !! :-o :bow:
:pint: :pint: :pint:
But it seems a long time since Dennis last update.
Has the thread been moved to any other place ? :idea:
-
I have found this link about Amiga AGA chipset. It is currently in Spanish.
Please note that I cannot asure if this data is right or wrong. However, it seems incomplete, but may be of help.
The link:
http://www.upv.es/amiga/120.htm
(I am on exams at the university so I won't have time to translate them in a while... :-( )
[EDIT:]
Some help to read it quickly:
Registro : Register
Nombre registro : Register name
Direccion : Address
Tipo de acceso : Access type (Read/write)
Lectura : Read
Escritura : Write
Hope it helps!
-
It's just a short version of what you can find in aga_guide.lha on aminet. Sadly enough, nothing special or new. Thanks for the link though.
-
@Dessimat0r:
Hi,
Wich brand and model is this "keyboard-with-case" ?
Or where did you find it ?
Thanks!
-
nadoom wrote:
do you reckon we could squeeze minimig into an old external amiga 1000 external disk drive? now that would rule!!
I think that would depend on how many connectors you want on the back more than the size of the rest of the board.
The C64 DTV unit fits in a Joystick. Actually, it doesn't even fill it. The gate array for an Amiga would need to be larger, it would have more data and address lines but that wouldn't take up more space than the old A1010(?) drive.
-
http://www.ebuyer.com/UK/product/87216
Here you go ;)
-
I dont think the C64dtv is an FPGA. It's an ultra low cost ASIC. They even have reduced the price further by not packaging the chip (putting it in plastic with legs) they have just glued the silicon to the board, bonded it out and put resin over the top. It can save perhaps $0.25 per chip and is a common practice in ultra low cost consumer apps.
-
Hi,
Finally a new update. I have been very busy with some non-amiga related stuff so it has taken some time to finish the PCB for Minimig. But nevertheless, the layout is finally finished. You can see a screenshot here:
Minimig rev 1.0 PCB screenshot (http://www.amiga.org/gallery/index.php?n=1470=13)
The PCB is a 2-layer design. This should make it easy and cheap to manufacture. I have also managed to keep the board relatively small, it only measures 120mm * 120mm!
The next step will be to actually make this board and to check if the FPGA will run reliably on only 2 layers. I have paid a lot of attention to grounding and power supply decoupling but one never knows for sure. I hope my next update will contain a picture of a populated and running board.
Dennis
-
Hey, really nice looking layout. Thanks for the post.
120mm, hey? An amiga 500 the size of a CD. Cool. What a fantastic job you've done!!!
Cheers :pint:
-
Dennis, it's only a matter of time till we see one of these in a joystick so why don't you make a deal where you can sell it now and still release it GPL in the future?
I'm not interested in the commercial route. I want to keep full control of my work and do what I please, without being binded by some sort of agreement or contract.
Any chance of making an AGA version?
No
Dennis, remember to keep us updated! :D
I will! :-)
Will it plug into a TV ? And will it plug into 14", 15", 17" or 19" (and above) standard svga monitors either crt or tft ?
As long as the TV has a scart and the monitors support 50Hz, it will run. I have tested Minimig on a 19" TFT and it worked...
You said that when you put it onto "your own board" that you'll add more memory. Does this mean more memory usable by the "amiga" meaning more chip or fast ram ?
All the RAM is on a common bus running at 7.09MHz. The RAM is being used as chip, fast and kickstart area at the same time by dividing up the ram in 3 area's using odd cycles for the amiga chips and even cycles for the 68000.
Could 1gb (or bigger) memory cards be used to store larger amounts of adf files ? Is the OS stored on the memory card too ? IF it was commercial could the OS be hardcoded with a removable memory card being for games only. Would there be some way to connect it to the pc for dumping different ADF files onto the card or upgrading the OS etc. Maybe having the OS in a rom and allowing upgrades (much like a bios upgrade).
Everything is loaded from the MMC card, including FPGA core, kickstart ROM and ADF files.
You say that you will likely go open source with pdf schematics etc. Sigh. Sounds like someone else will snap this up and make a {bleep}load of money from it then. Foolish i think, but its your decision. I'd seek legal advice or even take up (or talk to) some of those offers you've had. Man, you've got 30 pages of interest here in a niche corner of the web. Get some marketing behind your project and you've got thousands of sales. Get together with the guy doing the similar ST thing, make yours support both and sales could be massive !! You've done all the hard work now, now you should be concentrating on reaping some rewards.
The reward is being recognized from my work. To quote Mark Knopfler: "He's got a daytime job, he is doing allright."
i wondered if maybe the original idea was to build a flashcard replacement for DF0: and it took off from there?
No, but some people have contacted me to develop this part of the Mninimig as a standalone product. I wasn't interested as the full Minimig project is just much more fun. :-D
Dennis
-
@Dennis
3 words... gimme gimme gimme ;-)
Any plans to add a simple PIO mode IDE port like the Gayle chip does?
Failing that, a 'moon on a stick' would be nice :-D
-
Great to seen an update from Dennis' side. Some other developers could learn from this :-).
120x120 is nice, it could even fit a MacMini case ;-) [or my lunchbox]. Layout looks impressive to me, but said that... I'm not a hardware developer.
I can see Doobrey wanting IDE but the memory-card should give plenty of space to play with. I mean, these cards don't cost the world and come in rather big capacities. My old A3000 had a 850 MB harddisk and that was enough to keep me playing for years. The only thing you'd probably be missing is a possibility to add a CD Drive or something like that. But personally I'd not miss it.
Hmmm... the way it looks like now I can start day-dreaming of a Minimig as in-car entertainment for the back-seat gang while driving those long holiday trips.
Keep up the good work Dennis!
-
amiga_3k wrote:
...Minimig as in-car entertainment for the back-seat gang while driving those long holiday trips...
So, will that make them quieter, or noisier?
-
Hi Dennis,
I look foward to reviewing the MiniMig schematic, layout, and Verilog source when you feel ready to release the design via your chosen method (GPL etc ...).
Keep up the great work !!
:-) :-) :-)
-
Hi Dennis,
The picture of the two layer PCB of Minimig looks good.
After your debug of MiniMig, Will there be any blank PCBs available for purchase so that others can start to try to build up a Minimig as well to experiment with ?
:-)
-
props to you sir for going the non-commercial route! i'm gonna have as much fun constructing this as i'll have playing with it. :-D
also, thankyou very much for answering all our questions!
one more: is the core finalised, and if so, how is the current compatibility?
-
LoL "never give up never surrender !"
Galaxy Quest 1999 :-) :lol:
-
VGA to SCART cables are remarkably easy to get hold of, contrary to what I expected.
Check this out, for instance:
http://cgi.ebay.co.uk/Scart-to-15-pin-HD-Lead-VGA-Style_W0QQitemZ9737789091QQcategoryZ295QQrdZ1QQcmdZViewItem
..and there are loads of auctions like this on eBay.
However, it doesn't look like this has any audio connectors, which could be a problem, as you would need that in games.
-
By the way Dennis, have you already manufactured this PCB? What manufacturing process will be used to create it? Will you send it off to a company to be made, or will you make it yourself?
-
@Oliver:
Probably noisier, but don't tell my wife; I need a good excuse for getting me a MiniMig :-)
-
What about those of us who don't have the skills to build a Minimig ourselves? Will you be offering a build service? I would happily pay for this.
--
moto
-
Dessimat0r wrote:
However, it doesn't look like this has any audio connectors, which could be a problem, as you would need that in games.
I believe there is a mini stereo socket on the top edge of the layout.
As far as building goes, I guess a pre-written chip would be handy for most people too. I think Denis stipulated such, at some earlier point. I suppose it would be a bit of a drag supplying these for the next couple of years though.
Moto, if you get the parts, you could also pay a local technician to assemble it for you. I've done this kind of work for people before, for a minimal fee. I'm not a technician though, just an engineering student with much more hobby experience than real world work.
-
Will you send it off to a company to be made, or will you make it yourself?
Adam from AckControls can create a prototype very very fast...
-
Dreams come true.... :)
...but i want an A4000 keyboard port! ;P
-
Well Dennis, you've gone from being an unknown to most of us, to possibly looking like a white knight in only a few months.
As you know, the PPC options are looking a bit bleak right now, and that Coldfire card is still MIA. The Amiga community right now seems to mostly desire portability and enhanced Amigas.
Your MiniMig is a great proof of concept and you appear to be succeeding where Mick Tinker (for whatever reasons) did not. I hope you do not rest on your laurels, and instead are willing to either make this grow in as many directions as possible, or will at least permit others to license and work with you to develop families of devices that will grow the MiniMig in the directions that the users want. I see MiniMig as a starting point, and hope you do too. This might be a bit of a hobby for you, but for the Amiga faithful, this is the first sign of a second coming.
I hope you understand the emotional attachment people have to the philosophy of the Amiga architecture (to say nothing of their fond memories and experiences and years learning about it and money invested in). Clearly these people want to see a real Amiga come back in the 21 century even if it is a bit of a niche. I for one hope that you have an eye towards the future, or if you don't, are willing to work with people who do. For the sake of the Amiga architecture and the people that love it. Even if you don't have the time or interest it would be a shame to see your work and designs not progress to fill the needs of actual Amiga fans.
I'm hoping to see ECS support soon. This shouldn't be too difficult should it, since CBM essentially rebundled how fetches were exploited (IE., 4 planes worth of fetches became 2 planes but allowed for twice the number of pixels per line). I must admit to being perplexed why you don't appear interested in AGA support. Is this beyond your skils? Maybe we could find another who would be willing to tackle just the AGA graphics aspect of this new evolution of the Amiga.
Additionally, I hope that the 68000 is not the end of the line. Somebody mentioned making use of expansion boards. Maybe by use of a connector that allows the expansion board to sit under or over the MiniMig. That seems like a reasonable goal to me since it will allow versions of a MiniMig to be cusomized or cost-reduced.
On the previous page somebody mentioned 16 or 24-bit graphics support. If AGA support isn't something you can personally tackle (or wouldn't enjoy), then I would certainly encourage consideration of enhancing the graphics abilities via some 16 or 26-bit gfx chips. Ideally with overlay ability similar to the Firecracker board. Again, if for whatever reason these evolutionary paths aren't interesting to you, or you lack the time or skills (doubtful) then I'm sure I speak for the community when I ask that you be open to the idea of working with others who have an interest in growing your new Amiga incarnation in these directions.
Frankly, the MiniMig is too important to let it coast, or not see it for the only opportunity the Amiga really has (not to mention that the form factor alone is an important development). Despite the usual negativity from some posters here, you have shown your hardware to be real and accomplished much in a short amount of time. From what you say, I don't think you even appreciate what you've done. It's nothing short of astounding and I hope you do whatever it takes to allow it to grow to fill customer's desires.
On another topic, can your design support the genlock capability? There is still a market for video bulletin boards and the like, where a genlock compatibility would be useful.
-
Hi Dennis,
if you want, I can to realize the first prototype of this card
-
Dennis wrote:
All the RAM is on a common bus running at 7.09MHz. The RAM is being used as chip, fast and kickstart area at the same time by dividing up the ram in 3 area's using odd cycles for the amiga chips and even cycles for the 68000.
[nitpick]
Wouldn't that turn the all the FAST-MEM into SLOW/Ranger-MEM ?
[/nitpick]
[whine]
WANTONEMEWANTONEMEEEEEWANTONENOW
[/whine]
-
Oliver wrote:
I believe there is a mini stereo socket on the top edge of the layout.
Sorry, I meant on the cheapo VGA to SCART lead ;)
-
>>>>No, but some people have contacted me to develop this part of the Mninimig as a standalone product. I wasn't interested as the full Minimig project is just much more fun.
Dennis<<<<
Nice job on your project Dennis! When you are done, feel free to contact us as we have several FPGA type projects you might wish to get into for our motheboards. We could use a person of your skills.
JBuckly
Troika NG
-
@Boing:
From what I understood is that the OCS chipset data is loaded into the FPGA at boottime. From that I imagine that it should be possible to load ECS or AGA (not too sure if AGA would work with a 68000 processor) instead (if the FPGA could handle it).
Going a bit further into this day-dreaming I would not be surprised to see a Minimig-community, dedicated to do the ECS and / or AGA re-development. It could even have a go at incorporating different processors, enhanced gfx, and so on...
This little Amiga-wonder could well become one of the first OpenSource/OpenHardware computers creating perhaps a little revolution in computing.
-
@Troika
We could use a person of your skills.
True. Your so called "progress" with fried Amy's and board layouts done in DPaint looks rather poor in comparison to an amazing one-man project like Dennis with his Minimig.
-
@Kronos
Chipmem and fastmem are equally fast, but since there's a dedicated time slot for CPU access, undisturbed by chip access, it will always be like fastmem! At least for 68000 - but then again, if there's a bigger CPU to come, it'll bring its own mem and Dennis will probably turn Minimig's RAM completely to chipmem. :-D
Otherwise, if AGA should surface somehow, you'd need a lot more bandwidth for chip access...
-
@Zac67
The question is how close he kept his implementation to the orginal, cos on a real (OCS) Amiga useing more than 4 planes in LowRes or 2 in HighRes demands more bandwith than every 2nd cycle.
Useing 4 bitplanes highres (or 2 superhigres, or 8 superhighres in AGA) will eat up ALL cycles except those in the vertical/horizontal blanks. Add overscan and your miggy will come to a complete crawl (can be reproduced with any FAST-Mem-less miggy).
-
Any plans to add a simple PIO mode IDE port like the Gayle chip does?
Not on this board, not enough room and pins on the FPGA.
After your debug of MiniMig, Will there be any blank PCBs available for purchase so that others can start to try to build up a Minimig as well to experiment with ?
Maybe... (gerbers will be available for download however)
LoL "never give up never surrender !"
Hey, another "Questarian" :lol: :lol:
However, it doesn't look like this has any audio connectors, which could be a problem, as you would need that in games.
Oliver is right, there is a 3.5mm jack on the back.
Frankly, the MiniMig is too important to let it coast, or not see it for the only opportunity the Amiga really has
It will become available to the community in some form, I promise!
if you want, I can to realize the first prototype of this card
Thank you for the offer, but I have access to some fine soldering equipment and building it are the least of my worries right now. (getting the FPGA working on 2 layers and adapting the core to the slower RAM for example...)
Wouldn't that turn the all the FAST-MEM into SLOW/Ranger-MEM ?
No, see Zac67.
The question is how close he kept his implementation to the orginal, cos on a real (OCS) Amiga useing more than 4 planes in LowRes or 2 in HighRes demands more bandwith than every 2nd cycle.
True, but I have more cycles (twice as many), so there are always cycles left for the fast ram, cia's and rom.
Dennis
-
Dennis, if you are going to experiment with trying to get the FPGA to work on two layers, wouldn't it require you to experiment with PCB layouts? Therefore, wouldn't you require more than one Spartan 3, or do you have some method to easily salvage the chip from failed prototypes?
Edit: By the way, I was talking about the audio on the cheapo VGA to SCART lead ;) (see earlier post)
-
* Considering the frequncies on the board have you checked impedance matching?
Ie resistors to rid of any ringing. Maybe DCI could be helpfull ;)
* Some pcb manufactureing houses will offer serious discount for 2L eurocard
100 x 160 mm sized pcbs. Like 33 USD (iirc). Any chance to slightly adjust
the pcb? :)
* Will your schematic and/or pcb layout files be available?, So one could alter
the board for more/less chips etc..
* What power supply chips did you use..?, I presume linear ldo to keep things
simple. Cooling might be an issue.
The fpga might require "soft-start" to avoid hickup at boot.
* Working from your schematic it seems you will go for XC3S400-4PQ208 right?
An XC3S500E might provide more gates for future expansion?, without
sacrificing layout complications.
* Serial or Parallel load of the fpga ..?
* Does the current circuit handle realtime mfm encodeing from disc files present on the flashcard? (it was an issue a while ago).
* I'll see that parallel port and ATA is missing. If a few spare pins are
setup as a highspeed differential channel. They could use another fpga/cpld
to breakout those ports. The prerequisite is that a few i/o's are free.
* You will avoid bga like the plaugue? :-)
* Tip on component source for us RoHS plauged Europeans? :-)
-
* Considering the frequncies on the board have you checked impedance matching?
No, but on the current board I use SLOW 2ma drivers to keep edges as slow as possible. The maximum frequency on the board is just ~7Mhz. What I have now works nice and that is with very long wires in comparison with the new board.
* Will your schematic and/or pcb layout files be available?, So one could alter
Yes
* What power supply chips did you use..?, I presume linear ldo to keep things
Yes, linear LDO's for +1.2V, +2.5V and +3.3V. The 3.3V has a circuit to slow down the startup ramp as recommended by Xilinx. Total power consumption is about 200mA@5V with an optical mouse connected so cooling should not be a problem. (Although that was a concern to me)
* Working from your schematic it seems you will go for XC3S400-4PQ208 right? An XC3S500E might provide more gates for future expansion?, without sacrificing layout complications.
Yes, but where to buy? I'm already having trouble sourcing the XC3S400 (should have bought them when the Xilinx online shop was still in bussiness :-( ) in sample quantities.
* Serial or Parallel load of the fpga ..?
Serial
* Does the current circuit handle realtime mfm encodeing from disc files present on the flashcard? (it was an issue a while ago).
Yes, it is done inside the PIC18 microcontroller
* I'll see that parallel port and ATA is missing. If a few spare pins are
There are 4 spare pins wired to a 10pin header together with +3.3V and plenty of ground pins.
* You will avoid bga like the plaugue?
Yes, that is why I used the PQ208.
Tip on component source for us RoHS plauged Europeans?
I am from Europe, and I have completely ignored ROHS for this project, it is a hobby after all.....
Dennis
-
Yes, but where to buy? I'm already having trouble sourcing the XC3S400 (should have bought them when the Xilinx online shop was still in bussiness ) in sample quantities.
http://www.nuhorizons.com/ has both the XC3S400-4PQ208C and the XC3S500E-4PQG208C in stock, and will sell them in any quantity.
-
Dennis wrote:
* What power supply chips did you use..?, I presume linear ldo to keep things
Yes, linear LDO's for +1.2V, +2.5V and +3.3V. The 3.3V has a circuit to slow down the startup ramp as recommended by Xilinx. Total power consumption is about 200mA @ 5V with an optical mouse connected so cooling should not be a problem. (Although that was a concern to me)
Concern howcome?
Anyway which LDO did you choose?, and esp what ampere rateing? The ampere budget for xilinx fpga is still clouded in mystery ;)
It got to manage the powersurge at boot, not just normal usage.
I noticed many xilinx fpga boards use different sized capacitor to avoid resonance, btw.
* Working from your schematic it seems you will go for XC3S400-4PQ208 right? An XC3S500E might provide more gates for future expansion?, without sacrificing layout complications.
Yes, but where to buy? I'm already having trouble sourcing the XC3S400 (should have bought them when the Xilinx online shop was still in bussiness :-( ) in sample quantities.
Your not alone in cursing Xilinx for this behaviour. Fortunly there are other viable options. Digikey and NuHorizons.
NuHorizons have XC3S500E-4PQG208C for 25.94 USD.
Digikey have some XC3S200 in non-bga. However you can ask them to get non-catalogue items by phone.
There's even chips to be had on ebay.. ;)
Here's another shop selling xilinx fpga:
http://www.kamami.pl/?id_k1=87&id_k2=76&id_k3=20
The way to manage BGA should you ever consider it seems to be to apply solderresin on the pads. And then bake the whole lot in a hotair oven with the specified temperature curve.
I visited a assembly production house and saw the machines etc.. Metalbox with hotairgun controlled with a thermocouple and a laptop should do wonders ;)
The professional way is to have a pastescreen. Smear solderresin evenly (easy). Then put the chip(s) on. Have it travel through the oven. Done.
There's also a new vendor on the block. Actel.com have released their linux tools for free. And they don't have the power surge issue and don't need prom to startup. Something to consider for feature projects.
* Does the current circuit handle realtime mfm encodeing from disc files present on the flashcard? (it was an issue a while ago).
Yes, it is done inside the PIC18 microcontroller
Didn't you write some weeks ago that the PIC18 was too slow to handle the on-fly mfm conversions (from .adf iirc) ..?
I think the quick approach is to let the fpga do it. Ie feed the real data. And then convert it inside fpga. That way the PIC get's less workload.
* I'll see that parallel port and ATA is missing. If a few spare pins are
There are 4 spare pins wired to a 10pin header together with +3.3V and plenty of ground pins.
Expansion port then.. :)
Should work with 600 MB/s speeds or whatever Sp3 can fix =)
Tip on component source for us RoHS plauged Europeans?
I am from Europe, and I have completely ignored ROHS for this project, it is a hobby after all.....
The catch is if you sell it you better be RoHS compliant..
Ofcourse with gerber files and individual orders you can screw Brussels :-)
Board assembly could be tricky though.
Some notes I made on xilinx fpga:
* Outputs will have significant jitter such that PHYceivers etc.. should not be driven from them directly.
* Diode protection against accidental wrong polarity.
* Softstart and powersequencing (avoid surge).
* No unconnected Vcco within same bank.
* Large groundplanes.
* Skipping silkscreen & soldermask cut costs.
* A single 54p TSSOP sdram will give 32 MB (saveing pins on ram).
Do you have any checklist for fpga electricaly wise?
(like 5% regulation, prom setup, Vcco volts etc..)
-
motorollin wrote:
Trust me, price will not be an issue. Come on, I've just spent over £500 building an A1200 system!!!
moto
You dont even want to know how much I have spent on my A4000 recently.
Work it out for yourself with Google calc.
Australian dollars $800
4000 2MB chip, 16MB fast
Cyberstorm 060 (128MB Ram)
Scandoubler
PicassoII
-
Thanks @ mongo and freqmax
Haven't though about nuhorizons (didn't even know they existed!). I will place my order there today, before they're sold out.
The ampere budget for xilinx fpga is still clouded in mystery ;)
Yes, that's what I was concerned about. I ended up using the Xilinx online power estimatator and measuring the power usage of the spartan-2 development board. I use relatively fast 0.5A LDO's (some LDO's are really slow) and plenty of decoupling with 10n ceramic, 100n ceramic and 100u tantalum C's.
The professional way is to have a pastescreen. Smear solderresin evenly (easy). Then put the chip(s) on. Have it travel through the oven. Done.
For the company I work for, I design electronics for measurement instruments and we have a small assembly line with a benchtop vapour phase oven :-D . I was planning on using that to assemble my prototypes. The trouble with BGA is however that you need 4 layers or more and there is no easy way to check if all connections are properly soldered.
Didn't you write some weeks ago that the PIC18 was too slow to handle the on-fly mfm conversions
No, I did MFM encoding on the fly from the beginning. Maybe you where thinking of the guy who is building the MMC card floppy replacement?
Do you have any checklist for fpga electricaly wise?
No, offcourse it will work straight away... ;-) :lol:
Dennis
-
I agree.
AGA would be a dream...
I hope you understand the emotional attachment people have to the philosophy of the Amiga architecture (to say nothing of their fond memories and experiences and years learning about it and money invested in). Clearly these people want to see a real Amiga come back in the 21 century even if it is a bit of a niche. I for one hope that you have an eye towards the future, or if you don't, are willing to work with people who do. For the sake of the Amiga architecture and the people that love it. Even if you don't have the time or interest it would be a shame to see your work and designs not progress to fill the needs of actual Amiga fans.
So true.
Dennis, please, continue the developement...
:-)
-
Dennis wrote:
Total power consumption is about 200mA@5V with an optical mouse connected so cooling should not be a problem. (Although that was a concern to me)
Dennis
Woohoo, just imagine it - battery powered can runs for hours!
-
So Dennis... I think the question on everybody's lips is this: when will we have a chance to get our hands on your amazing invention?
--
moto
-
Dennis wrote:
Thanks @ mongo and freqmax
Haven't though about nuhorizons (didn't even know they existed!). I will place my order there today, before they're sold out.
Therse's also silica.com althought they have a weird ordering.. you phone to order.
The ampere budget for xilinx fpga is still clouded in mystery ;)
Yes, that's what I was concerned about. I ended up using the Xilinx online power estimatator and measuring the power usage of the spartan-2 development board. I use relatively fast 0.5A LDO's (some LDO's are really slow) and plenty of decoupling with 10n ceramic, 100n ceramic and 100u tantalum C's.
Where will you buy the non-fpga stuff..?
Have you measured the ampere usage on each voltage rail?
I suspect the proper way to do this is with a DSO at >10ksps, 0.5V/div. The first second is the most interesting.
One could keep the fpga on hold until at least the caps have loaded up ;)
Softstart issue:
http://www.eetasia.com/ARTICLES/2005JUN/A/2005JUN10_PL_POW_AN.PDF
Anyway just a quick ampere check with a dvm would be nice as a rough guide.
Any ideas on capacitor types..?, or is ceramic the way to go..?
The professional way is to have a pastescreen. Smear solderresin evenly (easy). Then put the chip(s) on. Have it travel through the oven. Done.
For the company I work for, I design electronics for measurement instruments and we have a small assembly line with a benchtop vapour phase oven :-D . I was planning on using that to assemble my prototypes. The trouble with BGA is however that you need 4 layers or more and there is no easy way to check if all connections are properly soldered.
[/quote]
I thought one could make some 2-Layer bga wonder. Maybe even through leaving lot's of pins unconnected..
Didn't you write some weeks ago that the PIC18 was too slow to handle the on-fly mfm conversions
No, I did MFM encoding on the fly from the beginning. Maybe you where thinking of the guy who is building the MMC card floppy replacement?
I hope the PIC18 can read those .adf files now flawlessly on the fly? ;).. and write.
Do you have any checklist for fpga electricaly wise?
No, offcourse it will work straight away... ;-) :lol:
Yeah sure :P :-)
-
@motorollin
I agree.
I can't wait.
Also if Dennis releases the Verilog and PCB sources for free he will not only instantly transform into an Amiga God (Like on the Populous!) it will also lead to a thriving minimig community with it's own Forums, Dev boards and community.
Yeehaaw! :lol:
-
I remember a few months ago reading about emachineshop.com - where you could download free CAD software and have them turn out your design. At the time, there was a place mentioned that would do the same for electronics designs. Not just make the bare board, but place and solder the components as well. I can't find the link right now- moving machines is a pain... but it might be the thing for folks not up for SMT soldering. (I can barely see those chips, let alone solder them. ;-) )
The idea hit me recently... I've seen home built CAM milling machines using steppers controlled by a pc. How about a really small one with a soldering iron instead of a mill?
(Way to Go Dennis!)
-
JimS wrote:
I remember a few months ago reading about emachineshop.com - where you could download free CAD software and have them turn out your design. At the time, there was a place mentioned that would do the same for electronics designs.
ExpressPCB (http://www.expresspcb.com/ExpressPCBHtm/HowWorks.htm) will do the PC boards but not the assembly work. I sure there are more out there.
-
RickParker wrote:
ExpressPCB (http://www.expresspcb.com/ExpressPCBHtm/HowWorks.htm) will do the PC boards but not the assembly work. I sure there are more out there.
Yeah, I've heard of them, and a few others advertised in Nuts & Volts or Circuit Cellar. This place will do the assembly work also. I'll see if I can find the link
-
Yes!!!
-
Wow, I'm surprised that this thread is still going on. I'm glad that some action is going on in the hobby world.
-
JimS wrote:
I remember a few months ago reading about emachineshop.com - where you could download free CAD
The idea hit me recently... I've seen home built CAM milling machines using steppers controlled by a pc. How about a really small one with a soldering iron instead of a mill?
Emachineshop seems great!, I have been thinking on it for a while..
Anyway milling machines would infact be less complicated in some aspect than an assembly/soldering machine. I've seen a 350k USD do this professionally. Moveing 5 cm/s.
Normal procedure is to have a pastemask (metal with holes) laid over pcb. Smear solderresin. Remove mask. Then components are placed. And finally it's run through an oven approx 4 m long with a programmed temperature curve.
So a robotic solution to this for prototypes would need to have X-Y-Z-Rotation.
It would need to drip solderresin on every pad, place component, Solder with a minature iron. A camera/boroscope should enable the user to see what's going on.
Obstacles:
10 mil precision and repeatable precision X-Y-Z arms or movement. Traction like printers may work for first/second axis. But might run into troubles with Z-axis due acceleration issues (you don't wanna smash your $20 fpga :).
Alternatives are linear motors if they can be made good enough and on the cheap. Motor driven screws inteded for this purpose I've seen DRC machines use this approach. And lastly gear driven positioning, which ofcourse requires precision gears made..
Machinery should have some kind of feedback or "Track 0" device to know where it is.
Don't forget that if you move a mass precision will suffer from acceleration/deacceleration mismatch. Also mechanical parts hysteresis comes to play. Thus the need for feedback sensors.
So how to physicaly make the X-Y-Z leds move needs to be solved.
One could use meccano or lego to make something..
Any tip on lowpriced sources for this? :)
-
Hi,
You may want to take a look at this:
http://www.pcbtrain.co.uk/assembly.php
I think, in fact its worth the time to visit the whole web site.
If you know of any other company that does this same thing, please let me know.
-
freqmax wrote:
Normal procedure is to have a pastemask (metal with holes) laid over pcb. Smear solderresin. Remove mask. Then components are placed. And finally it's run through an oven approx 4 m long with a programmed temperature curve.
I saw a web site once (linked from slashdot, I believe) on this very topic. The guy was using this process, but with a toaster oven! Naturally a thriftshop purchase, since you wouldn't want to cook in an oven with lead deposits. Seems he had good results with this method. There was something about a special solder paste, it had to be kept in a refrigerator untill needed... wierd..
So a robotic solution to this for prototypes would need to have X-Y-Z-Rotation.
It would need to drip solderresin on every pad, place component, Solder with a minature iron. A camera/boroscope should enable the user to see what's going on.
I was only thinking about doing the soldering robotically. I figured on doing the placement manually. I like the camera idea, they're small enough these days. I suppose the solder paste is a must. I wonder if surface tension would prevent you from soldering something that small with wire solder. I know... a laser. :-)
Obstacles:
10 mil precision and repeatable precision X-Y-Z arms or movement.
All points well taken. I used to work on disk drives as a service tech. One of the drives used an optical scale to drive the head positioning servos. An active system would probably be better than a 'dead reckoning' stepper based system. Anyway, at this point it would probably be cheaper to hire someone with the right equipment to build your own minimig. :-)
-
"Have a look at this little beauty!", Steve Irwin would say about the minimig :-D
Nah, it's just an excuse to bump this topic up :lol:
-
Dennis!
I saw this in an electronics discussion group:
>> Why are LDOs so picky about output caps? Makes no sense to me.
>Maybe to complement the stability of Windows. Think of ESR like you'd
>think about salt when cooking: No too much, but also not too little. And
>don't expect any hard data on what "too little" means ...
>I don't use LDOs unless there is absolutely no other option. Usually there is.
Maybe you should keep this in mind when designing the power section of the minimig.
Is the verilog source complete as you have started the board development?, if so maybe we could have a look ? :)
-
Chain wrote:
Dennis wrote:
For the company I work for, I design electronics for measurement instruments and we have a small assembly line with a benchtop vapour phase oven . I was planning on using that to assemble my prototypes. The trouble with BGA is however that you need 4 layers or more and there is no easy way to check if all connections are properly soldered.hours!
Just for the record, and I do understand it's not at your shop.. I worked with a contract electronics manu. company which does through and surfacemount.. BGA is "Easy" as long as you have an Xray.. That's how we check for solderability on BGA. If you're looking to get your company up on the latest, a BGA inspection machine (x-ray) can be added to any SMT's manufacturing line. But yeah, personally as a hobby or private endeavour, I wouldn't get into it. TOo hard to fix when there IS a mistake.. Contract manu. simply scraps the board and writes it off, you obviously won't want to be doing that..
On an aside - This is my first post back here in months, maybe the year??? Glad to see you're still working on this project.
-
Just for the record, and I do understand it's not at your shop.. I worked with a contract electronics manu. company which does through and surfacemount.. BGA is "Easy" as long as you have an Xray.. That's how we check for solderability on BGA. If you're looking to get your company up on the latest, a BGA inspection machine (x-ray) can be added to any SMT's manufacturing line. But yeah, personally as a hobby or private endeavour, I wouldn't get into it. TOo hard to fix when there IS a mistake.. Contract manu. simply scraps the board and writes it off, you obviously won't want to be doing that..
Shouldn't it be possible to desolder a bga with an oven..?
As for checking connections, seems any radiotube + >20kV powersource should give xrays to work with. Maybe a ccd is sensitive to this kind of radiation in a meaningful way.
Dram's or flash memories maybe could serve as receiving end of such setup aswell.
Hmm.. AGAIN amiga.org logout just in time for posting.. *GRR* (end-rant :)
-
Is this the longest Thread ever?
can someone in Authority (like Tom on MySpace - LOL) do some sort of summary page.
I have been trying to keeping up but, I was away a few days and missed 10000000 posts....
please don't flame me!
-
Basically Dennis designed an A500 compatible computer implemented on an FPGA. It uses a real 68000 chip and a few other externals. The whole OCS chipset is implemented on the FPGA (ok, it's not gate-for-gate OCS, but Dennis' equivalent based on documentation and observed behaviour.
At the start of the thread you have people calling him a fraud and yelling "hoax". Then Dennis demo'd it at some gatherings in the Netherlands and the photos were posted online. Since then, most of the posts are from people who want one, or suggest improvements (e.g. add AGA support please).
Dennis is now in the process of designing and building a circuit-board for the Minimig instead of using the FPGA starter-kit he developed it on.
I think that's the gist of it. I'm sure others will fill in any gaps.
Hans
-
BTW,
The PCB manufacturer that also does component placement and assembly I mentioned earlier in the thread is called padtopad.com.
-Jim
-
Plus4 wrote:
I was away a few days and missed 10000000 posts....
And at 36296 views (currently), it's got to be the most popular one on A.org...
At least the forum software didn't crap out when the view counter got past 32767!!!
- Ali
-
InTheSand wrote:
Plus4 wrote:
I was away a few days and missed 10000000 posts....
And at 36296 views (currently), it's got to be the most popular one on A.org...
At least the forum software didn't crap out when the view counter got past 32767!!!
- Ali
you know, if that counter was truly a signed short int... then some forum software developer is in a loooot oooof troooouble! ;)
-
I think there are more people who want to build a Minimig than there are people who want to buy a new 'AmigaOne' :lol:
-
I would LOVE to get my hands on a minimig... so yeah, you are probably right.
Hell, I would love to grab an A1200, A3000, or A4000 again. Of course, I'm technically "broke" at the moment, so that isn't gonna happen for a couple months at least.
I have little interest in an A1 because of the 1) excessive prices and 2) doesn't run classic Amiga software without emulation. If I am going to emulate, I'll just use my PC. :)
-
CD32Freak wrote:
I think there are more people who want to build a Minimig than there are people who want to buy a new 'AmigaOne' :lol:
Because most people who want an Amigaone would love to have a Minimig too. Add to that all the existing OS4 users, Morphos users and people interested only in the classic Amiga and you've got a sizeable group. The Minimig appeals to almost everyone here.
Hans
-
I guess it appeals to even more as the code to program the FPGA is loaded at startup from a flash-card. I guess that it's a matter of weeks after this wonderfull thing has been released before the first ECS/AGA implementations will be available and heck... I guess that Atari MegaST and other 68k based systems should be possible (that's my 'non-expert'look at it) on the same hardware as well. Or am I too optimistic about the potential of the hardware?
-
I guess that it's a matter of weeks after this wonderfull thing has been released before the first ECS/AGA implementations will be available
Well, Dennis has said a few times that he isn't interested in AGA. I am sure if he releases the HDL then AGA would follow - however it is not a trivial 'upgrade' and I am sure it would take at least months (not weeks). :cry:
I guess that Atari MegaST and other 68k based systems should be possible (that's my 'non-expert'look at it) on the same hardware as well.
I think these systems are fairly easy to implement as (if I remember right) they just used a small amount of ULA for glue and standard peripheral LSIs for sound, graphics etc. The minimig would have no bearing on these type of projects.
-
Concerning the Atari ST implementation, a german guy has just done it !
check this out:
http://www.experiment-s.de/index.php?id=187&L=1&L=1
And MikeJ from www.fpgaarcade.com has announced that he would implement the Atari Jaguar and Jaguar II.
Any volunteer for the Sinclair QL or the X68000 ? :-D
I agree with Colin_Camper that the AGA is not a simple upgrade. It is also more expensive : just take the video DAC, you will need a real triple 8-bit DAC.
The best dev board I see for that is the Terasic DE2, only $269 if you are a student ($495 otherwise... :-( ).
It has everything you need to make an AGA Amiga on steroid : 10-bit triple video DAC, SD Card, RS232, Ethernet, 24-bit audio, USB, PS2 port, Video in (!) 8 MB SDRAM, 4 MB Flash, 512KB SRAM and the cyclone II has 35000 LEs ! (equivalent to a X3S1500).
-
The Minimig is AWESOME!!! Seen it live, touched with my hands, watched demos and games on it... it's kinda miracle! :D
check some pics here! You can even spot Dennis and Bernd Schmidt (father of UAE) playing Turrican on the Minimig!!! :D
Codex Alpe Adria site (http://0xaa.org)
-
Colin_Camper wrote:
The minimig would have no bearing on these type of projects.
But it'd still be cool to have the Minimig hardware emulate an ST as well as an Amiga - nice to have both!
- Ali
-
Would have been nice to know about the event beforehand so that those in the area could have attended.. Did I miss the post on that, or were we kept in the dark?
-
Rooster wrote:
Would have been nice to know about the event beforehand so that those in the area could have attended..
The show/event was in Italy. A long way from Ohio & Michigan. :-/ But it was posted on the amiga.org homepage.
-
See and I was just in Italy last month. Ok, I wasn't, unless you count video games, lol.. Must'a missed the announcement somehow I guess, true, not that I could've gotten there. :(
-
Codex Alpe Adria site (http://0xaa.org)
is that a prototype minimig board i spy?
-
oh dear...
http://0xaa.org/2006/gallery/gfx/IMG_9250.jpg (http://0xaa.org/2006/gallery/gfx/IMG_9250.jpg)
http://www.0xaa.org/2006/gallery/gfx/IMG_9255.jpg (http://www.0xaa.org/2006/gallery/gfx/IMG_9255.jpg)
http://www.0xaa.org/2006/gallery/gfx/IMG_9454.jpg (http://www.0xaa.org/2006/gallery/gfx/IMG_9454.jpg)
new pics coming soon on the Codex site...
-
But it'd still be cool to have the Minimig hardware emulate an ST as well as an Amiga - nice to have both!
- Ali
the ST's CPU clock speed was a little bit faster at 8MHz, so i suppose it depends on whether the minmig board allows for overclocking (at the very least). i think i read that that might be a bonus feature, but i can't be bothered sifting through 32 pages to check...
and here i go, adding to it. LOL.
-
@discostu28
Here is the quote from Dennis earlier in the thread.
To keep things simple I have merged the Amiga chipram bus and fastram bus into a single synchronous bus running at 7.09379MHz.
I think this means that the 68000 is running from the same clock.
I would imagine (naturally Dennis can correct me!) that it would not be a trivial change to incorporate overclocking into the Minimig. I would have thought that the 'natural Amiga way' using a CPU expansion bus would involve an FPGA with more i/o.
Of course, instead of guessing, it would be nice to have an update from Dennis! :-)
-
Changing the clock speed just one MHz would be the least of all troubles - but you'd have to model the chipset first and that's where the work is.
I followed the link to the ST in FPGA project: that one was made to be compatible at chip level - much more complicate to do, but could probably be used.
-
Some games on the Amiga probably expect that precise clock, so if you go adjusting it, you'll mess up the timing within those games.
-
@ Colin_Camper
cheers, dude. well, we shall see...
@this thread
y'know what would be really cool? a generic hardware platform for synthesising old platforms. there's already the C1 (which has a Z80 core and chipset for the Amstrad CPC) , but what about one that could do everything from the A1200 to System16?
it could be the biggest buzz since emulation! i really feel this has potential, but the level of knowledge, effort. and financial investment required is too high right now. people like dennis and jeri ellsworth are real trailblazers.
a functional 68000 core could set it off. the Z80 cores are here. yeah, and that's all a lot of arcade machines used back in the 80s. MAME 4 real!
-
Have you any idea how much FPGA's that size cost? Or the cost of a low production run multi-layer PCB? It's not economically viable... yet.
Plus the FPGA code is only a re-interpretation of the original hardware and is as much subject to bugs as emulation.
-
Take a look at these folks... www.schmartboard.com. They've got a system of prototyping boards that make it possible to hand solder even those sub-millimeter pitch surface mount chips. It looks interesting. It works by having a high solder mask between the traces. This helps to hold the chip in place while soldering. It also forms channels to flow the solder onto the pins without bridging over to adjacent ones. You just heat the pre-tinned trace and push that solder up to the pin in the channel.
They also say they support BGA chips, but I see their site lists those as "coming soon".
-
@alehx
it's true. you'd need massive FPGAs to do this. 1 or 2 million logic blocks. even one 68000 has 300K gates. a synthesized version would probably be larger, due to redundancies and inefficiencies in the contrived logic. an analyzer could be used on a real 68000 and the core, running in tandem, to identify descrepencies in I/O between the two.
oh, i TOTALLY know what i'm talking about, BTW. :lol:
i've been reading. it's a fascinating subject. i'd say there's a market for a board that satisfies both developers and users, here. maybe xillinx will recognize this and produce something more geared for the hobbyist.
or maybe not xillinx. i suppose what i'm getting at is that the analogue outputs for whatever arcade/computer system aren't really that different. if you could just have one big fat FPGA, a chunk of SRAM and video/audio DACs, you could suck all the sytem specific stuff onto the FPGA. and burble burble burble. okay, i ran out of knowledge here. but if people will shell out 100s for a supergun, this has a market. for sure.
-
discostu28 wrote:
even one 68000 has 300K gates. a synthesized version would probably be larger
Eh? The original 68000 had approx 68k gates, hence why it was called 68000 :-)
You are right a HDL version will take up MUCH more than 68k gates.
-
alexh wrote:
discostu28 wrote:
even one 68000 has 300K gates. a synthesized version would probably be larger
Eh? The original 68000 had approx 68k gates, hence why it was called 68000 :-)
You are right a HDL version will take up MUCH more than 68k gates.
oops, my bad. wow, it could nearly fit on the current spartan with the chipset.
-
These dudes have a comerical 68k core ('instruction set compatible');
http://www.dcd.pl/dcdpdf/xil/d68000_ds.pdf
Their Xilinx implementation doesn't list Spartan though.
I've watched this for a whiles now, thinking to myself how cool it would be to find a freebie 68k core. Google around a bit you'll find a very imcomplete one somewhere (sorry I forget where) and a couple of comerical ones. I recall from a long ago news / forum thread that the 68k was not considered an 'easy' CPU to implement in an FPGA and that the economics of FPGA area meant that other more powerful CPUs could be implemented cheaper. Of course programmable logic technology has come along way in the last few years, so who knows.
Anyway, just my 0.2$. Very - very excited about the Minimig , especially if the source gets released. :-D
-
The free ones are poor and the commercial ones are too expensive for Denis's needs (and are probably also poor).
I came across D68000, Cast C68000 (http://www.cast-inc.com/cores/c68000/index.shtml), m68000 (http://tams-www.informatik.uni-hamburg.de/vhdl/) (very poor), k68 (http://www.opencores.org/projects.cgi/web/k68/overview)
MiniMig's 68k hardcore is a good idea IMHO, why re-invent a complex wheel when they are still available and you can buy them cheap enough.
-
i suppose i should stop posting meanderings as it keeps bumping the thread causing everyone to think there's been an update. LOL.
anyway, i'll shut up after this. there's no good reason not to use a real 68000 in minimig. however, it's a logical progression to synthesise the CPU as well. it could potentially allow the entire hardware to be described for any reconfigurable system. this does raise chicken-and-egg issues, though. like...
why would i waste time engineering a 68000 core when there's no system to run it on?
why would i waste time producing a more powerful reconfigurable computer when there aren't any decent cores to run on it?
which is probably why all the open-cores are crap. i expect somebody would have to feel motivated enough to do both the platform and the core.
-
Hi, my first post here!
Just to say that the 68000 has 68000 transistors, not gates.
-
I completely agree : a 68000 is not an easy CPU for an FPGA.
Actually, it does not worth making a 68000 clone in VHDL when you can make a 68030/68040 clone for the same effort.
The 680x0 ALU is quite simple and can run at 70+ MHz on cylone II / spartan 3 serie with almost all the operations taking 1 cycle (only exception : the divide that takes 8 cycles).
The big difficulties come when you implement the instruction decode with the 12 kinds of effective addresses.
If you really want an efficient CPU, you need a cache and some branch prediction to minimize the pipeline stall.
The pipeline adds also few data hazards.
Anyway, if you still want to make a microprogrammed 68000 instead of a pipelined one, you can have a look at this document, it is a very good introduction :
http://www.cs.tcd.ie/Michael.Manzke/fyp2003-2004/DavidLynch.pdf
Regards,
Frederic
-
Hi, my first post here!
Hi! Welcome to Amiga Org!
Just to say that the 68000 has 68000 transistors, not gates.
I thought 68000 gates was a lot for a late 'seventies processor! 68K trannies rings a bell with me too! :idea:
I guess we're only talking about 8K gates then....I'm not sure what the average trannie count per gate is - I would think about 8 for a NAND?
PS! :-o Have you got an update for us, Dennis?
Is there any chance we could be playing with your invention by the Autumn? :-)
-
I'm not sure what the average trannie count per gate is - I would think about 8 for a NAND?
4 transistors for NAND or NOR. Only 2 for a simple inverter. 20 or so for a D-latch, 28 for a standard full adder (the sort you'll find in textbooks), although you can make one with only 14.
I was wondering if it would be possible to make a Minimig that basically just takes a standard A1200 accelerator board. Or does the Amiga still need the processor on the motherboard even if you're using an accelerator?
-
Awesome work, Dennis.
I have just signed in because I have some points about this project. I am building an Amiga clone too, but my target is to build full-featured AGA machine. Also, as I saw the photographs of minimig PCB, there are some circuit-related problems with interfacing a Spartan-3 FPGA.
To introduce myself, I am a lon-term developer of devices that are using Xilinx FPGAs and Analog Devices DSPs. I am also a fan of Amiga and I started to build a FPGA based clone. For start, I have a Spartan-3 3S1500 1.5MGate device to build full-featured core. I will say something about my project later.
At first, I have to say that you have done awesome work on the core. On the second hand, there are some flaws on PCB from my point of view (and, I cannot see other side of PCB :-) ). At first, there is lot of 5V logic around Spartan-3. PS/2 ports, joystick ports etc. All of this hardware is tied directly to Spartan-3 via single resistor. So you are counting on protection diodes in spartan IOB. But - this is NOT good design practice and with some kinds of hardware it may fail to work and you are risking to fry both FPGA and the peripheral. There should be a logic level converter for this. Next, the video port is build around simple resistor DAC, there should be a video DAC part instead of this because of better conversion and mainly, because there is a separate ground for R-G-B pins to improve image quality and overcome any on-screen artefacts that can be caused by bad grounding. You even have monitor digital ground directly connected with R-G-B analog ground. At least, there should be a ferrite bead. Next, I cannot see all power regulators needed for Spartan-3 device. This FPGA needs three power supplies, 1.2V VCCINT, 2.5V VCCAUX and 3.3V VCCO for 3.3V outer peripherals. A can only see one three terminal regulator (maybe others are on other side ?). But if you have these supplies created by resistor networks or any other circuit, it will not work correctly. There is a need for either a low-dropout regulators or (better but harder to properly create unless complete part is obtained) DC-DC switching circuits.
If you need any help regarding the hardware point of view, just ask me and please try to write about mentioned points.
About my implementation - it is currently in a PCB development (but concepts and parts of circuit were completely tested) and it compromises of two boards. On one board, there are all ports and connectors:
PS/2 keyboard, PS/2 mouse, Amiga joystick, serial port, RGB VGA, composite video, S-Video, floppy connector and CompactFlash card slot (maybe there will be addition of some additional ports)
On this board, there are all circuits regarding to ports, so there is a VGA RGB encoder, RGB-to-composite/svideo encoder, sound codec (96 kHz / 24 bits stereo), logic level converters and glue logic for all ports (my target is QUALITY of analog output signals on both video and audio). Also, maybe there will be a Spartan-3E 500kgate device to operate all this stuff and behave as part of amiga chipset responsible for all this stuff. CompactFlash card is used as a HDD and as a kickstart and .adf loader. Next there will be a system connector to other board, where will be the core of the system. For now, there will be another Spartan-3E device, 68020 processor (but I am working on a softcore, now it is done nearly to 68000 state and lot of work is pending...), 2MBytes of memory and test space for connecting custom chips to ensure full compatibility.
Now, it is in the state of PCB design, and I have not much spare time to do some quick progress, although I have logic analyzer and any comfort a development lab could provide.
Maybe creating an Open-Source project page about Amiga FPGA implementation and creation of these projects in community should help to progress this to the point where WinUAE is now ;-)
-
but my target is to build full-featured AGA machine.
Wow! God, sorry Dennis, now has competition! :-D
Are you sure about AGA?:-o - you're not Mick Tinker are you?
-
No, I am not :-D . The way to full-featured AGA machine is longer than OCS. First FPGA core implementation sure will be OCS/ECS which is partly done and I can post some pictures on verifying parts of core on Virtex-4 ML-401 board (which has RGB VGA out), but if it were possible to emulate AGA in WinUAE then it will be possible to implement it in hardware.
-
Great news then !
It's Amiga time again soon :)
-
@ Lukage
An AGA machine on a PCI card would be nice, not as an active PCI expansion of course, just using a passive slot for power will do.....now put this on an A1 (or whatever new PPC mobo) and a true OS4 machine can sit in the box with a 68K machine in a slot. OH JOY!! :pint:
-
LoL - I threw that idea out there like a year ago - Putting the Ami 'guts' on a PCI card :) Glad to see someone else does think it's a hot idea :)
-
Colin_Camper wrote:
the_woz wrote:
Just to say that the 68000 has 68000 transistors, not gates.
I thought 68000 gates was a lot for a late 'seventies processor! 68K trannies rings a bell with me too!
The name "68000" was selected largely to provide some marketing continuity with the earlier "6800 series". The name was justified by the claim the chip featured about 68,000 transistors, in fact, the count (on the origional version) was closer to 70,000 transistors.
Rather than a 68000 clone in VHDL, wouldn't it be better to implement a "Coldfire co-processor" in VDHL to trap/execute missing 68000 instructions, to use with a ColdFire MCF5475 processor?
-
metalman wrote:
Colin_Camper wrote:
the_woz wrote:
Just to say that the 68000 has 68000 transistors, not gates.
I thought 68000 gates was a lot for a late 'seventies processor! 68K trannies rings a bell with me too!
The name "68000" was selected largely to provide some marketing continuity with the earlier "6800 series". The name was justified by the claim the chip featured about 68,000 transistors, in fact, the count (on the origional version) was closer to 70,000 transistors.
Rather than a 68000 clone in VHDL, wouldn't it be better to implement a "Coldfire co-processor" in VDHL to trap/execute missing 68000 instructions, to use with a ColdFire MCF5475 processor?
It would probably be better just to implement a very simple RISC core, which has built in features to aid running a 68k emulator (Flag calculator, etc..)... Then run a 68k emu on this core.
-
bloodline wrote:
metalman wrote:
Rather than a 68000 clone in VHDL, wouldn't it be better to implement a "Coldfire co-processor" in VDHL to trap/execute missing 68000 instructions, to use with a ColdFire MCF5475 processor?
It would probably be better just to implement a very simple RISC core, which has built in features to aid running a 68k emulator (Flag calculator, etc..)... Then run a 68k emu on this core.
Freescale Semiconductor partner Micro APL has CF68KLib, an emulation library available which provides exception handlers to implement those 680x0 instuctions and addressing modes which are missing in the ColdFire architecture
CF68KLib: 68K Emulation for ColdFire (http://www.microapl.co.uk/Porting/ColdFire/index.html)
CF68KLib is a free download (by agreement with Freescale)
-
Hi Dennis,
I'm Luke of Amigaplatform (www.amigaplatform.com) an italian website dedicated to Amiga.
I need to contact you...please contact us to info at amigaplatform dot com
thank you
Luke
-
If an Arm or StrongArm is used you could run at about 400Mhz - Most Arm/StrongArms are cheap.
You could then run a 68K emulator for the 68K bit (a fast Arm chip could easily emulate an 8Mhz 68K) and have a native Arm mode for faster code.
This way you could have one board that could emulate hundreds of machines - the FPGA doing the custom chipsets, the Arm doing the software CPU.
Another alternative could be the AVR32 - they run at about 200Mhz are (imho) better than the Arm (no swapping between 16 and 32 bit instruction modes) - that are about $15 see Atmels website for more info.
-
metalman wrote:
Freescale Semiconductor partner Micro APL has CF68KLib, an emulation library available which provides exception handlers to implement those 680x0 instuctions and addressing modes which are missing in the ColdFire architecture
CF68KLib: 68K Emulation for ColdFire (http://www.microapl.co.uk/Porting/ColdFire/index.html)
CF68KLib is a free download (by agreement with Freescale)
The CF68KLib is designed to support the execution of 68K object code, nothing is said about the actual performance of it. From what I recall, the emulation by exception overhead for the missing opcodes is extortinate in machine cycle terms under coldfire. Any code that uses a large number of these instructions or uses them in any kind of tight loop is going to take a massive performance hit.
-
Hi,
At first, there is lot of 5V logic around Spartan-3. PS/2 ports, joystick ports etc. All of this hardware is tied directly to Spartan-3 via single resistor. So you are counting on protection diodes in spartan IOB. But - this is NOT good design practice and with some kinds of hardware it may fail to work and you are risking to fry both FPGA and the peripheral.
I know, but this board was designed to be as simple as possible and there is nothing more simple than just a resistor. It is even suggested by this xilinx answer record (http://www.xilinx.com/xlnx/xil_ans_display.jsp?iLanguageID=1&iCountryID=1&getPagePath=19146). Besides, there is never really 5V put into any pin at all! The joysticks are (electrically) just a bunch of switches to gnd, they make the input low when pushed and high (through the FPGA's internal pullup) when released. More or less the same goes for the PS/2 ports, the signal lines are defined as open collector, so they will never go higher then 3.3V. (again pulled up by FPGA internal pullup) The only thing I am wearry about is the high level of the PS/2 lines but as I've understand PS/2 is all TTL level so even that one will go right.
Next, the video port is build around simple resistor DAC, there should be a video DAC part instead of this because of better conversion and mainly, because there is a separate ground for R-G-B pins to improve image quality and overcome any on-screen artefacts that can be caused by bad grounding.
It is a simple resistor DAC, just like in a real A500. I agree that seperate grounds CAN give better results, but only when used wisely. In my current proto, I have excellent picture on a variety of TVs and monitors thus I see no reason to do it otherwise.
Next, I cannot see all power regulators needed for Spartan-3 device. This FPGA needs three power supplies, 1.2V VCCINT, 2.5V VCCAUX and 3.3V VCCO for 3.3V outer peripherals
There are 2 extra LDO's mounted at the underside of the PCB to provide +1.2V and +2.5V. Also, all FPGA, CPU and RAM decoupling caps are also mounted at the underside (I am surprised you didn't asked about the capacitors?).
my target is QUALITY of analog output signals on both video and audio
Why does anybody think that you can't have quality without superduper Analog Devices chips? In the end your audio codec will be just some noise shaping single bit thingy like the minimig dac. AND, to match the codec you will have to do the samplerate conversion right, if you don't.. bye bye quality. Once the new board is ready I will be happy to demonstrate it again in Maarssen so anyone can see the audio and video quality.
Maybe creating an Open-Source project page about Amiga FPGA implementation and creation of these projects in community should help to progress this to the point where WinUAE is no
Open source is definetely the way to go. But to run cores like this, one needs hardware. I am planning to release minimig as opensource but it is unwise to do that now so have postponed it untill this july next year. It would be a bad thing for the community if I do it now and in the end, this should be better.
Good luck with your project,
Dennis
-
You could then run a 68K emulator for the 68K bit (a fast Arm chip could easily emulate an 8Mhz 68K) and have a native Arm mode for faster code.
You know, that's what I've been thinking about too. Run an emulator on some fast, readily available mcu/cpu. But I would use a freescale ppc (like the one ACKcontrols was going to use) as that would also allow PPC software to be run.
Dennis
-
@Dennis
Weren't IBM were making 603 cores with all sorts of additional IO HW on board? Could be fun :-)
-
You know, that's what I've been thinking about too. Run an emulator on some fast, readily available mcu/cpu. But I would use a freescale ppc (like the one ACKcontrols was going to use) as that would also allow PPC software to be run.
:-)
Are you going to update your board design with this possible change?
-
so have postponed it untill this july next year.
Wow! A year! :-o
Will you be doing (or allowing) someone to do a minimig kit in the meantime? :-)
It would be a bad thing for the community if I do it now and in the end, this should be better.
Come on, Dennis! 8-) You said right at the start you have been lurking here, long before you announced minimig. Do you realise how the above statement sound to our suspicious and conspiritorial ears. :crazy:
-
@Dennis - I want a minimig
@lukage - I want the Aga version
:-D
-
Dennis wrote:
You could then run a 68K emulator for the 68K bit (a fast Arm chip could easily emulate an 8Mhz 68K) and have a native Arm mode for faster code.
You know, that's what I've been thinking about too. Run an emulator on some fast, readily available mcu/cpu. But I would use a freescale ppc (like the one ACKcontrols was going to use) as that would also allow PPC software to be run.
Dennis
A good idea, but perhaps 4 years too late be a commercial product I fear. Unless you can build your board and make a profit with a retail price of less than 80 euros (~£50).
Why this price? It is the average price of a second hand XBOX 1.
-
alexh wrote:
Why this price? It is the average price of a second hand XBOX 1.
I'm afraid you are wrong. They were sold new at only 49 Euro at the very last moment (not so long ago).
So they could never be sold at such a high price, unless they experiment a higher demand now they are "classic" hardware.
But I dont think they will never be as popular as the Amiga was, or other classic computers,...
-
But I dont think they will never be as popular as the Amiga was, or other classic computers,...
I think you have misinterpreted what I meant.
You can buy an XBOX1, to run all your favourite emulators, at a very low price. (perhaps as little as 50 euro's)
So if you wanted to make a dedicated emulator console using a PPC as Dennis suggested, I believe it would have to be faster or cheaper than an XBOX1 to be a commercial success.
-
err....have you been reading this thread from the beginning?
motorollin wrote:
Trust me, price will not be an issue. Come on, I've just spent over £500 building an A1200 system!!! I'm sure the cost of manufacturing and distributing Dennis' Minimig wouldn't even come close to that, allowing a healthy profit margin and still sell at a price Amiga enthusiasts would welcome!
--
moto
See my point? We Amiga users sell our kidneys to buy hardware, what the hell makes you think more than fifty quid is too much?
-
alexh wrote:
You can buy an XBOX1, to run all your favourite emulators, at a very low price. (perhaps as little as 50 euro's)
quote]
Plus the chip ;) etc..
-
Tripitaka wrote:
alexh wrote:
You can buy an XBOX1, to run all your favourite emulators, at a very low price. (perhaps as little as 50 euro's)
Plus the chip ;) etc..
...and risking screwing up the xbox
...and voiding any warranty you have left (unless you buy it used, then it doesn't matter)
-
revlayle wrote:
Tripitaka wrote:
alexh wrote:
You can buy an XBOX1, to run all your favourite emulators, at a very low price. (perhaps as little as 50 euro's)
Plus the chip ;) etc..
...and risking screwing up the xbox
...and voiding any warranty you have left (unless you buy it used, then it doesn't matter)
I bought mine on launch, after the warranty was over I then added my executor chip....then a 200GB HDD...then a nice blue transparent case....Lights....etc..etc...AND I DIDN'T BREAK ANYTHING...well, maybe my bank account. :-D
-
revlayle wrote:
Tripitaka wrote:
alexh wrote:
You can buy an XBOX1, to run all your favourite emulators, at a very low price. (perhaps as little as 50 euro's)
Plus the chip ;) etc..
...and risking screwing up the xbox
...and voiding any warranty you have left (unless you buy it used, then it doesn't matter)
oh no...what a shame that would be lol
-
Tripitaka wrote:
err....have you been reading this thread from the beginning?
motorollin wrote:
Trust me, price will not be an issue. Come on, I've just spent over £500 building an A1200 system!!! I'm sure the cost of manufacturing and distributing Dennis' Minimig wouldn't even come close to that, allowing a healthy profit margin and still sell at a price Amiga enthusiasts would welcome!
--
moto
See my point? We Amiga users sell our kidneys to buy hardware, what the hell makes you think more than fifty quid is too much?
because we shouldn't have to
and if it is cheap enough, you could get other people back onto the amiga
-
sir_inferno wrote:
and if it is cheap enough, you could get other people back onto the amiga
HA HA HA.... On the one hand I agree we SHOULDN'T have to .....but the reality is the Amiga is such a minority platform now that we DO HAVE to...and yes, a greater user base should make life cheaper for all of us, however, minimig or similar will NOT do this. Only new modern hardware (I'm not going to get sidetracked into the OS4 on x86 etc. argument here) with a proper reason to exist can expand the user base.
As to what the "proper reason" for a new Amiga to exist in the mainstream would be......that's a whole new world of opinion.
-
Jeesh, can we all just let the poor fella get on with his project without complaining about how much it should cost or how (un)commercialisable it might be?
-
For something to be mainstream it have to fulfill something desired by a large or influential portion of the marketsegment.
Anyway I think the main concept behind Amiga/Commodore etc.. is to push hw limits and fit as much good hardware in a price that most people can afford. Like how much integrated hw can you stuff on a board below 500 EUR?
AND have a good software base. Games non-computer people can appriciate. Programs that computer people can make use of.
A speculation.. maybe a Mips/Arm cpu with some powerfull graphics, mixable n-channel 192 kHz sound, sockets for
cheap dram, PCI-expresss, S-Ata, ethernet/usb/firewire/dvi
connectors.
Due that both ATI & NVidia seems to gone the "keep-documents-from-developers" path. None of them is a good candidate however.
You can get far with a fpga except for the real power sections like cpu, and very fast I/O buses.
An XBOX maybe can do many things, but what any cpu based emulation won't do is handle latency/parallelism gracefully. Because an fpga can manage nanosecond delays.
-
Yay Dennis ! :banana: :banana::banana::banana::banana::banana::banana:
Excellent project, enjoy life ! :pint:
-
Yeah! and a :pint: for Lukage too! :-D
-
HA HA HA.... On the one hand I agree we SHOULDN'T have to .....but the reality is the Amiga is such a minority platform now that we DO HAVE to...and yes, a greater user base should make life cheaper for all of us, however, minimig or similar will NOT do this. Only new modern hardware (I'm not going to get sidetracked into the OS4 on x86 etc. argument here) with a proper reason to exist can expand the user base.
There is nothing wrong with the Amiga. I recon that you could take the Minimig project, and beef up certain aspects of it to make a it a viable modern game platform while maintaining backwards compatibility with old Amiga games.
-
Dessimat0r wrote:
There is nothing wrong with the Amiga. I recon that you could take the Minimig project, and beef up certain aspects of it to make a it a viable modern game platform while maintaining backwards compatibility with old Amiga games.
At what level? If you mean a GP32 kind of thing that's one market........or perhaps an Amiga to compete with Ninty's DS...or PS3...or...
So what kind of "viable modern game platform " are you percieving for a beefed up Minimig????
-
At what level? If you mean a GP32 kind of thing that's one market........or perhaps an Amiga to compete with Ninty's DS...or PS3...or...
Perhaps not necessarily 3D games, but innovative 2D ones, developed in the same manner as classic Amiga games, with more memory, bandwidth and processing power available.
-
Plus the chip ;) etc..
...and risking screwing up the xbox
...and voiding any warranty you have left (unless you buy it used, then it doesn't matter)
Erm, XBOX's havent needed mod chips for a very, very long time. But we are going very much off topic.
I have my fingers crossed that MiniMig comes out as an open source project and becomes widely used, improved etc.
-
Dennis, you have captured the imaginations and hopes of so many Amiga users with your work. Keep up the great work and know that there are many of us who appreciate your work, just as you have described it, without AGA. I am very happy to read that you are planning to release Minimig as open source and look forward to its release next July.
Take care,
David
-
Karlos wrote:
metalman wrote:
CF68KLib, is an emulation library available which provides exception handlers to implement those 680x0 instuctions and addressing modes which are missing in the ColdFire architecture
CF68KLib: 68K Emulation for ColdFire (http://www.microapl.co.uk/Porting/ColdFire/index.html)
CF68KLib is a free download (by agreement with Freescale)
The CF68KLib is designed to support the execution of 68K object code, nothing is said about the actual performance of it. From what I recall, the emulation by exception overhead for the missing opcodes is extortinate in machine cycle terms under coldfire. Any code that uses a large number of these instructions or uses them in any kind of tight loop is going to take a massive performance hit.
A few opcodes massivly slower, most much much faster, the result would probably be faster than an actual 680x0, hopefully.
The benefit of using a Coldfire (MCF5475 or MCF5485) is it's relatively cheap, and has built in support for 133 DDR/SDR-SDRAM, Ethernet Controllers, USB, 33/66 PCI interface, ect.
Someday........
....... Freescale may even release the V5 core chip. :-o ?!
that was announced in 2003 ......
a faster more modern version of a A3000 (25mHz)(ECS) machine, that would be easy to interface and a minimal chips to build is the idea.
When :bow: Dennis :bow: releases minimig open-source it will spawn a lot of creativity in this area.
-
Karlos wrote:
@Dennis
Weren't IBM were making 603 cores with all sorts of additional IO HW on board? Could be fun :-)
well, xilinx make these:
http://www.xilinx.com/products/silicon_solutions/fpgas/virtex/virtex4/capabilities/powerpc.htm
-
@Metalman
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
As you say, most operations will obviously run faster but the one's that are emulated are the problem. Don't underestimate the cost of exception processing - you could possibly execute dozens of instructions in the time it takes to fully trap and emulate just one.
The proportion of emulated opcodes to nonemulated ones in the machine spec is not the issue either, it's the proportion of emulated ones present in the object code and where they occur that (at least historically) is the issue...
-
Karlos wrote:
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
Would it be possible to write a program that parses through the 68K object code looking for coldfire illegal instructions & modes, replacing them with coldfire instructions.
-
JimS wrote:
Karlos wrote:
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
Would it be possible to write a program that parses through the 68K object code looking for coldfire illegal instructions & modes, replacing them with coldfire instructions.
It isn't always that simple. It's unlikely that the replacement code will fit in the space where the offending instruction would fit. That means appending code and adding branches and so on, which is pretty hacky to do statically. Also, do you really want some program walking all over your executable files, libraries etc and patching them like this?
However, provided you can definately trap all the unimplemented opcodes at runtime (and this was certainly an issue on older coldfires) then you could conceivably create an OxyPatcher type utility that does this on the fly as the code executes. You only take the exception hit once, the first time the instruction is encountered.
-
@Karlos
Lets not forget that pretty much ALL Amiga executables have data in code segments and/or code in data segments, so with file patching theres no way to be sure the "instruction" you just detected an patched was actually an instruction, and not some critical data, so either a JIT like Oxypatcher, or just a full on 68k emu launched at power-on are the only two workable solutions.
And the former isn't definately 100% possible, since some opcodes on 68k aren't just missing on CF, but actually different and so don't trigger any kind of exception.
-
@Xeron
Quite - that's what I meant by "hacky". As for mixing the code and data sections, that always struck me as a bad idea ;-)
Regarding the trap problems, I know that some opcodes didn't trigger an exception per se, but as I understood it, some form of runtime detection of unsupported/differently implemented opcodes exists in the more recent cores in order to implement the CF68KLib. Don't quote me on it though, it's been some time since I looked.
-
Good... so we're all agreed that the coldfire isn't the way to go! :-D
-
JimS wrote:
Karlos wrote:
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
Would it be possible to write a program that parses through the 68K object code looking for coldfire illegal instructions & modes, replacing them with coldfire instructions.
Thats what the old Decigel and Fix68010 programs did to fix program sources when people were swapping 68010's for 68000's for the faster multiply/Divide on the 68010's. That was one of my first upgrades to my old A1000.
The other solution was to run a wedge program to trap the priviledged instructions at runtime.
-
Karlos wrote:
@Metalman
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
As you say, most operations will obviously run faster but the one's that are emulated are the problem. Don't underestimate the cost of exception processing - you could possibly execute dozens of instructions in the time it takes to fully trap and emulate just one.
The proportion of emulated opcodes to nonemulated ones in the machine spec is not the issue either, it's the proportion of emulated ones present in the object code and where they occur that (at least historically) is the issue...
Use a Coprocessor to that executes the missing instructions, that eleminates emulation.
Implement a "Coldfire hardware instruction trap" to trap missing 680x0 instructions, sending missing instructions to a coprocessor ( a 68020, 68030, or VDHL ??) for execution, send regular instructions to the ColdFire MCF5475 or MCF5485 main processor. Similar to how the MC68882 floating-point coprocessor executes Floating-Point Arithmetic instructions.
-
I can see that actually being slower, tbh...
Think about it for a moment. How do you even interface it? The 6888x used a defined set of opcode 'lines' for the fpu instructions and the 680x0 were designed to work with it.
-
metalman wrote:
Karlos wrote:
@Metalman
Don't get me wrong, the coldfires are cool and ideal in terms of what they support 'out of the box' hardware wise. I'm just not sure about the CF68KLib with normal amigaos object code.
As you say, most operations will obviously run faster but the one's that are emulated are the problem. Don't underestimate the cost of exception processing - you could possibly execute dozens of instructions in the time it takes to fully trap and emulate just one.
The proportion of emulated opcodes to nonemulated ones in the machine spec is not the issue either, it's the proportion of emulated ones present in the object code and where they occur that (at least historically) is the issue...
Use a Coprocessor to that executes the missing instructions, that eleminates emulation.
Implement a "Coldfire hardware instruction trap" to trap missing 680x0 instructions, sending missing instructions to a coprocessor ( a 68020, 68030, or VDHL ??) for execution, send regular instructions to the ColdFire MCF5475 or MCF5485 main processor. Similar to how the MC68882 floating-point coprocessor executes Floating-Point Arithmetic instructions.
It's not just missing instructions with the ColdFire... there are some instructions which actually function differently to the 68k specification... the multiply is the oft quoth example, which doesn't handle the flag calculation as 68k software would expect.
-
Hi Dennis,
your is a great job...
Why not use your ability to clone a A1200 or a A4000?
However...very very very good project.
Luke
-
I would rather that he cloned an Amiga 500, and saw the project to completion. If he wants to do A1200, My guess is that he will release the core under a license and work on AGA support with the community as a joint effort.
-
Dessimat0r wrote:
I would rather that he cloned an Amiga 500, and saw the project to completion.
Oh yea!
One of the problems since the diminished of C= Amiga has been grand ideas that never got finished.
A project, regardless of now grand, is worthless unfinished.
It's much better to pick a project big enough to make a difference but small enough to accomplish.
Guys how about laying off the AGA stuff for now. AGA is to big of a project right now.
Dennis has his project sized large enough to make a difference small enough to get it done.
-
RickParker wrote:
Guys how about laying off the AGA stuff for now. AGA is to big of a project right now.
Dennis has his project sized large enough to make a difference small enough to get it done.
I agree. Just thought I'd add the following:
Dennis is implementing the OCS chipset in an FPGA. FPGA's are Field Programmable, that it, you can completely reprogram the logic circuits whenever you feel like it. If Dennis' design uses an FPGA with enough unused logic cells then one could always work on developing an AGA implementation using exactly the same hardware. If he provides a JTAG port so that you can connect a programing cable, this will be really easy. At the most, you might want to change connect a 68020 processor or higher to the FPGA instead of the 68000.
Hans
-
Dennis, my points was not meant to insult you in any way, as I said, I don't have complete schematics and only one side of the board is shown... You have done great work and I hope to see it finished smoothly. Well done.
I have the video & audio & ports board almost done and have all electronic parts for this board. But I am stuck with CPU & FPGA board for a simple reason: It will be best to use a PPC processor instead of 68020 or a softcore, because higher speeds of CPU and possible movement to PPC cannot be obtained by PPC CPU softcore or hardware redesign. At last, I can waste a lot of work simply because I have the cycle-accurate 68020 core 50 percent done. But I would like to not be stuck with 68020 machine and further development in softcore will not result in much speedup because complexity rises and FPGA softcore CPU will not have much high clock speeds.
Now, I can use a normal PPC processor or I am thinking to use a Virtex-4 FPGA device with embedded PPC processor. I have not yet checked opcodes and possibilities of 68k emulation so this is for further evaluation.
Cheers,
Lukage
-
Hello a first post here too!
Thank you Dennis for sharing your thoughts and updates with us. Thank you all others for your respective contributions to the group. I'm a software engineer with a hobby interest in FPGA's last 2 years. I bought the original Spartan-3 Starter Kit (200k) right away when it was available and has been fooling around with it since day 1. Very fun!
My main projects started (not completed) is:
1) VGA output with 4096 colors without extra hardware on the original Spartan 3 Starter Kit (from now on called S3SK)
2) The start of a vector-based drawing engine that for now only draws circles/arcs, dots and lines.. but fun anyway.
3) A MIDI-Synth (very simple) that actually can be used directly from f.e. Cakewalk. Monophonic, only sinus-waveforms, only 1 osc. today. I have the start of ADSR-handling at least.. The only hardware, beside the S3SK is 2 resistors, 2 capacitors and 2 RCA-plugs! All these are used for simple D/A sound-output directly from the FPGA-pin.
4) The start of a 68k IP, with the goal for minimal implementing gate-equivalent count.. OF course inspired by the talk/work done here.. And bloodline, yes I've used the same thoughts you presented here. My first goal is NOT to get it cycle-accurate, but only behavioural eqivalence. Today I'm guessing 40% 68'000 ready and < 10% used in the Spartan-3 200k circuit.. The goal is of course at least 8Mhz Freq. MAX but today it seams very tight..
OK hopefully some of this was interesting to the group al though I didn't mention the A***A anytime... :-)
Good luck everybody and if anybody starts a website for these projects I maybe could share my verilog code...
Best Regards
Magnus
-
1) VGA output with 4096 colors without extra hardware on the original Spartan 3 Starter Kit (from now on called S3SK)
I thought that the Spartan 3's VGA port had only one-bit per channel? So how do you add the extra colours? Are you using Pulse Width Modulation (PWM) of some sort? If so, does that work on LCD monitors too? I have a feeling that, due to the sampling nature of LCD monitor inputs, PWM wouldn't work very well.
Hans
-
You're right on the spot there, Hans! :-D Yes I'm kind of PWM'ing the output and, yes LCD's don't like the signal but it still looks OK on my old CRT.
And for you to be such a good guesser, maybe you also could help me specifying a passive output-filter that would make it work better? :-) And why would I waana do that? Of course I'm trying to build minimalistic just for the fun of it!
Nice to see that someone else is listening..
/M
-
magnus wrote:
You're right on the spot there, Hans! :-D Yes I'm kind of PWM'ing the output and, yes LCD's don't like the signal but it still looks OK on my old CRT.
:-D
Well I couldn't think of any other method of getting 4096 colours from a 3-bit VGA connector.
And for you to be such a good guesser, maybe you also could help me specifying a passive output-filter that would make it work better? :-) And why would I waana do that? Of course I'm trying to build minimalistic just for the fun of it!
A simple RC filter should be enough. IIRC Xilinx provides the schematic for the starter kit you're using. Find out what the resistor connected to the VGA output pins are. You can then calculate what capacitor needs to be put across the output. You can use this (http://ww1.microchip.com/downloads/en/AppNotes/00538c.pdf) document from Microchip to do the calculations. Make sure that you use the same PWM frequency for all resolutions, or you'll need a different RC filter for each resolution.
You can skip the OpAmp that they have after the RC filter in the Microchip document I referenced above, as a computer monitor's inputs usually are pretty high-impedance (yes, I know that this is bad practise, but you're trying to keep it simple) and the connected monitor shouldn't cause the filter response to change by much.
Note that, due to the higher frequencies involved, you have to choose capacitors that operate well at these frequencies. Ceramic and MFT capacitors should be ok. Electrolytic capacitors are definitely not as they cease behaving like capacitors at high-frequencies. If the capacitance is too large, add extra resistance. Keep all wiring as small as possible or the image quality will suffer.
Make sure you use components with small tolerances. Otherwise the frequency response will be wrong due to the actual component's value being off from the desired value.
Finally, if this doesn't work well enough, have a look at R-2R digital to analog converters. This is the method that Dennis is using and he says that the image quality is good. It's just a series of resistors connected to different pins on the FPGA, and to the VGA connector. It's pretty simple too and avoids the filtering issue.
Let me know if the RC filter works ok.
Hans
-
Thanks Hans!
I'll give it ago after I've set some more features on my synth-circuit! And yes, I'll be sure to post the outcome/result of it here!
The R2R-ladder is a cheap, easy and good solution, but you know the old saying, NIH... ;-) It is already used on most simple/cheap FPGA-development boards (that doesn't include RAMDAC) so... I like new, original ideas to be tested!
Thanks for the info!
-
Here is a useful link for getting 4096 colours out of a Spartan dev board:
http://home.freeuk.com/fpgaarcade/displaytest.htm
-
Hi Dennis I also thought about doing some little FPGA work. So my questions:
1. do you know if it would be possible to make a 32Bit CPU out of a Spartan 3 FPGA?
2. What´s better? An Altera Cyclone 2 or a Xilinx Sparta 3?
3. Would it be possible to interface an FPGA with a PC via PCI bus and then using the FPGA as a FPU/Co-Processor?
-
@Blinx123
3. FpgaForFun - PCI (http://www.fpga4fun.com/PCI.html) complete with source code and stuff
-
Oh thank you. I saw this site just before but thought the project would be about a PCI interface on a self build motherboard.
-
Thanks. Just for the record R2R ladders are a better solution than my fpgaarcade mod - but would require more soldering.
/Mike
-
Hi Dennis,
very cool project. Have you got some news?
Thanks!
-
Hey Dennis,
There's a thread on Amigaworld mentioning the following.
PICO E-12 (CompactFlash Format) (http://www.picocomputing.com/products/cards/e12ep.php)
* Virtex-4 FPGA
* 128 MB RAM
* 64 MB Flash ROM
* Gigabit Ethernet
* JTAG Hardware / Software Debugging
* 44-bit High Speed Digital Bus
* PowerPC 405 Processor (EP Model)
PICO E-14 (CardBus Format) (http://www.picocomputing.com/products/cards/e14ep.php)
* Virtex-4 FPGA
* 256 MB RAM
* 64 MB Flash ROM
* High Speed Analog Converters
* Gigabit Ethernet
* RS-232 Serial Port
* JTAG Hardware / Software Debugging
* 54-bit High Speed Digital Bus
* PowerPC 405 Processor (450 mghz.)
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=20030&forum=33 (http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=20030&forum=33)
I think someone mentioned virtex in this thread.
Have you considered this as a potential platform?
And of course the PPC would give you the possibility of loading PPC based Amiga OSes that could then hit the good old Amiga custom chips. That's really what an Amigan would want anyway.
Ok. Going back to my happy place now.
-
I think perhaps it would be more cost effective to pay for the PPC on another FPGA and maybe between two there would be more room for I/O lines and expansion.
btw. I, also, am very much looking forward to an update from :bow: Dennis :bow:
-
Dennis, any new updates? I'm learning electrical engineering just so that I can have a chance to build the Minimig when it comes ;)
-
@Dessimat0r:
Guess a very small and quick: 'Working on it, goes well' from Dennis would already be more than welcome.
-
Just to keep this thread up to date, Dennis has confirmed that he will be exhibiting MiniMig at Big Bash 4 in Peterborough, UK on the 16th September.
See also: http://www.amiga.org/forums/showthread.php?t=24417
-
Awesome -- this seems to be a good chance to see the latest developments regarding Minimig.
-
To me that anouncement sounds prettymuch like: 'Still working on it, progressing steadily' :-) Wish I could be there.
Hmmm... hope there's no voltage difference between the continent of Europe and England / UK.
-
Blinx123 wrote:
1. do you know if it would be possible to make a 32Bit CPU out of a Spartan 3 FPGA?
Yes, but it depend on how many "gates" you need. Virtex adds significantly to the cost. But you could manage with more than one spartan-3.
Btw, Actel is on the move with free tools aswell.
2. What´s better? An Altera Cyclone 2 or a Xilinx Sparta 3?
Altera lacks any linux version of their programming tools.
The price picture may differ.
3. Would it be possible to interface an FPGA with a PC via PCI bus and then using the FPGA as a FPU/Co-Processor?
Yes it's possible. Look at:
http://www.enterpoint.co.uk/moelbryn/raggedstone1.html
98 EUR
-
amiga_3k wrote:
To me that anouncement sounds prettymuch like: 'Still working on it, progressing steadily' :-) Wish I could be there.
Hmmm... hope there's no voltage difference between the continent of Europe and England / UK.
It's 240V in UK, vs 230V in the rest of Europe asfaik.
So yes, and it's not likely to make any significant difference. Particulary in this SMPS world of powersupply..
-
AFAIK it's 230V in the UK. I think it used to be 220V for the rest of Europe and 240V for the UK, so they picked 230V as the standard.
-
maffoo wrote:
AFAIK it's 230V in the UK. I think it used to be 220V for the rest of Europe and 240V for the UK, so they picked 230V as the standard.
Ah... that's no problem then. Would have been a serious silly situation if Dennis showed up and not having the right power to feed the machine with.
Over here, the Netherlands, I believe it's 220V garantueed with a maximum towards the 240/250 ish Volts.
-
Hi All,
The last couple of months have been very busy for me. I spent most of the time painting, putting wallpaper on the wall ("behangen" in dutch), laying floors etc. In one word, I have been decorating my new home :banana: . In the meantime I did not have much time to work on Minimig. I have done some small things but the new board still has not been assembled. Now that I have more or less settled in my new home I want to pick up on the project. I now have a seperate hobby room where I can just leave a mess and shut the door :-D .
I am looking forward to go to BB4, it is a nice way to spend the weekend. I will bring Minimig with me to demo it just like I did in Italy. BTW, no worries about the voltage. All power supplies I have are SMPS (switch mode power supply) and accept a range of voltages. Also, as far as I know both GB and the Netherlands have 230V. I only need to bring a suitable adapter as the dutch power plugs are different than the british ones.
See you at BB4!
-
@Dennis
I understood from what you said in Italy that your next move was to reduce everythign down to a 12x12 cm/inch board.
how far away do you think you are form this?
-
I understood from what you said in Italy that your next move was to reduce everythign down to a 12x12 cm/inch board.
I am actually very close, the board schematics and artwork are done. (see this photo (http://www.amiga.org/gallery/index.php?n=1470=13)). Also, all parts have been bought to assemble some prototypes. The only problem is time....
Dennis
-
Next time you move into a new house, just inform us and we all come to help you putting on the wallpaper and your floor so you can spent more time on the minimig :lol:
-
Dunno if this has been posted before, anyway here it is:
0xAA podcasts (http://amigaforever.com/news-events/20060701codex_alpe_adria/)
there's the seminar Dennis held at Codex Alpe Adria there. Enjoy!
@Dennis
you did not reply to my email! :)
-
Very good work: I am impressed!
I am actually very close, the board schematics and artwork are done. (see this photo (http://www.amiga.org/gallery/index.php?n=1470=13)). Also, all parts have been bought to assemble some prototypes. The only problem is time....
-
voyager wrote:
Next time you move into a new house, just inform us and we all come to help you putting on the wallpaper and your floor so you can spent more time on the minimig :lol:
Hehehe... I just get a vision of half a dozen Amiga-users discussing wether to use a red-white-tile pattern wallpaper or a coloured Amiga-checkmark pattern. In an another room again half a dozen Amiga-users discussing wether to use a power-screwdriver to fix a bookshelf or just a good old hand screwdriver... The bathroom is filled with a couple of Amiga-users discussing how easy glue-ing PVC piping is, compared to soldering copper piping, but regret the fact that the glue is there, but no tubing due to delivery problems with the supplier.
Ah well.. Dennis, good luck in your new home and keep up the good work on Minimig.
-
Dennis,
I for one am really impressed and hope to see the reports of how well your minimig works.
I am not sure why the emphasis on making everything smaller but it is pervasise.
Anyway by emulating the functions of the "unobtanium" chips and using industry standard parts for all else you(and others) will preserve and enhance the Amiga experience.
We ESPECIALLY need the 8(oh heck,go for 64) meg chip memory as per other threads :lol:
-
Dennis, I asked this before in the thread, but I don't think you saw it. Would it be possible to 'pimp' Minimig up with stuff which is inaccessible by original games, but which new games could use, if they were programmed specifically for the Minimig (and the relevant development tools were available).
For example, you could add way more memory which could be used, or you could allow more sprites to be created, thus leading to new games which look much more modern.
-
@Skan:
you did not reply to my email! :)
Yeah, I'm sorry. The problem is I don't have internet access at home right now. If everything goes well my phone line should be activated today. The ADSL line should be activated in about 2 weeks. I will answer you through webmail today.
Dennis
-
Dennis, I asked this before in the thread, but I don't think you saw it. Would it be possible to 'pimp' Minimig up with stuff which is inaccessible by original games, but which new games could use, if they were programmed specifically for the Minimig (and the relevant development tools were available).
That should be possible. What I would especially like is a chunky mode display and a 16bit audio mode. The latter could be realized by putting 2 channels in parallel, giving you 16bits of audio.
Dennis
-
@Dennis
I sure hope you didn't use HetNet for your internet, I can give you all the excuses they use straight away before they actually get you connected:
1: Are you sure your computer is working ok ?
2: We're switching to a new internal system
3: You're connected to the wrong connection
4: We're very busy and do all we can
5: It should work within 2 weeks, 14 working days,= 3 weeks
6: We're very busy waiting!
blablabla
:)
-
HellCoder wrote:
I sure hope you didn't use HetNet for your internet
Bah. That name just gives me nightmares at broad daylight. Although no company is really without trouble, I can't say Tiscali has give me any trouble.
:-D
-
@amiga_3k
:roflmao: I thought I was going to wet myself!
-
Dennis: What's a 'chunky mode display', by the way? It sounds interesting :)
-
Here's the technical stuff... (dry, humorless, etc.)
The original display was using one bitmap (storage area) per bit of the color. This means you can have from 2^1=2 to 2^6=64 (OCS/ECS) or 2^8=256 (AGA) colors.
Chunky means: instead of one bit/plane per pixel, you use half a byte or one byte from memory for the color information of one pixel.
In theory this means one memory byte can represent 8 two color pixels, 4 four color pixels, 2 sixteen colour pixels or one 256 color pixel.
Since the byte needs to be split, you loose the possibility of having bitmaps with 8, 32, 64, and 128 colors in chunky mode.
Also, the entire bitmap needs to be one coninuous memory area, whereas with bitplanes, you can allocate them anywhere in CHIP RAM as long as they are the same size.
Hope this helps ;-)
What I would like to see is a difference between the viewing area (monitor resolution), a virtual desktop area with a default background color - no data bitmaps for that - of limited resolution (e.g. 16Kx16K), and the allocated bitmaps (chunky or not) representing possibly overlapping windows in that area, i.e. hardware windowing.
The hardware windows could even lie partially outside the monitor viewarea as long as they lie within the virtual desktop area, and you could not only move windows, but change the viewarea relative location as if you were panning.
This implementation would need more memory than programmatically redrawing, but also less processor power from the CPU (slightly more from the GPU) on redraws.
Of course, doing all this in 3D space would be even more fantastic... :-P
Cheers,
Olaf.
-
I see, so you can have bigger pixels, but with fewer colours.
-
@Dessimator
No, its nothing like that. Both planar (classic amiga) and chunky display modes can display 256 colours in as much resolution as the display hardware can manage.
These links may help:
http://en.wikipedia.org/wiki/Planar
http://en.wikipedia.org/wiki/Chunky
-
One point of interest is that neither system is "better" than the other, but they are better suited to different tasks. Planar graphics suited days where memory/bandwidth were at a premium. Why use 256 colours per pixel if you only happened to need say 8? Other effects such as scrolling, playfields etc are relatively easy to accomplish on a planar graphics system.
Chunky pixels lend themselves better to drawing; especially individual pixels. On a 256 colour planar display, if you wanted to plot a single pixel (and admittedly this isn't the most efficient thing to do in any event), you might make up to 8 reads and writes to the display memory. The reason for the writes is obvious - each bit in your colour value needs to be turned on or off individually for each of the 8 planes. The reads come from the basic fact that there is no way to address an individual bit. In realiy, you'd probably end up reading at least a byte from each plane (containing your target bit), modifying your bit then writing the whole byte back where it came from.
The same operation on a chunky display simply involves writing your byte directly to wherever the single pixel lives.
-
I'm sure it's already been mentioned, but the cd32 akiko chip included a chunky to planar conversion function to apparently assist in porting titles to the platform...
-
And I bet the factory(ies) in China have the plans and could start turning out CD32s next week if there was a demonstrated market.
If someone would design an improved and updated SX32 they might sell hundreds? of them!:roll:
Matter of fact wasn't there some deal where they got the rights to sell Amiga inside China? Of course they are supposedly too busy cranking out "pirated" copies of Windows and Hoolywood movies ....I wonder if there is really that much? :roll:
-
From the Amiga Inc Q&A
Essentially, since you guys have one of the most sought after intangibles assets around, namely a HUGE brand, is it time to cash in, either by selling the Amiga brand to say, Commodore or partnering with another forward thinking organisation or individual, like the MiniMig creator; in a hardware sense. Maybe ala the C64 Joystick. Does Amiga, Inc. have plans to capitalize on this large and growing market?"
Bill McEwen -- Yes, Amiga actually entered into a contract in 2005 with a firm to develop some products based on the Amiga classic technologies and we are looking forward to seeing those products. We have also been in discussions with the creator of the MiniMig, but he decided to go another direction.
so what did they proposed to the creator of minimig and which is the other direction he chose to go? Creator (http://www.adslgr.com/forum/images/smilies/worthy.gif), can you give us these infos? :-)
-
re: bitplanes
IMHO, they're too much of a pain if you have to work with more than a few of them. I think the way to go would have been to have a chunky layer plus a bitplane which could be used for text or optionally as another bit of color depth. Then there could also be the possibility for having high-res text displayed on top while setting the chunky layer to low-res for more speed or colors.
As for adding new features to a modern reincarnation of the old chipsets such as minimig... it would be nice to have some chunky mode in there. Maybe add a new register bit that causes one of the old bitplane pointers to point to a 8/16bpp plane instead. Or if the minimig already has a framebuffer for deinterlacing (I can't remember if this is the case) then just give the CPU access to that.
-
so what did they proposed to the creator of minimig and which is the other direction he chose to go?
I watched the video feed of the q & a with Dennis and I am sure I heard him (heroically) say that he decided against a deal with Amiga.Inc because it could interfere with his goal of making minimig open source.
However it was noisy.... I may have got the wrong end of the stick.
I am amazed that Dennis is going down this road - I am sure he could have made a fair bit of money. Now that would make him a wealthy, good looking genius - maybe he was scared that people would hate him. :lol:
Maybe he could make a compromise, let Amiga give its name to an A500 in a joystick, while letting Dennis allow the supply of minimig kits to the hobbyist market. Maybe the thing could be time limited to 2 years - after that minimig would go open source. Something along those lines. Let's face it, if it goes open source - there's nothing stopping Amiga Inc collaborating with Commodore or others to release the joystick. Any one could make it - but it would only sell a million if it had the name Amiga! :-)
-
Eh, there's a video feed? I only heard the audio feed.
-
Eh, there's a video feed? I only heard the audio feed.
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=20535&forum=16
But be warned, it's not that nice to watch.
-
You realize he could have his cake and eat it, too.. You can license your technology to someone (IE Amiga Inc.) whilst still maintaining control over what you do personally. So he could sell the 'use' of his product to them, to badge as whatever they want, and still run his operation personally, open source, or whatever..
Just like Gateway/Dell (as a basic example) buy printers from HP and label them as their own product, with it's own model number etc., even though it looks JUST LIKE the HP unit. Yea, the example is a little different, but still the same :)
-
EXCELLENT suggestion, if amenable to both Dennis and Amiga,Inc.
After RedHat and other vendors make money from packaging and supporting a free operating system.
Hate to see any viable Amiga platform ignored.
Sell 'em a license,Dennis.
-
Colin_Camper wrote:
I am amazed that Dennis is going down this road - I am sure he could have made a fair bit of money.
With the shenanigans going on with Amiga Inc and related companies over the years, keeping away from them is probably the best thing Dennis could do..
-
Rooster wrote:
You realize he could have his cake and eat it, too.. You can license your technology to someone (IE Amiga Inc.) whilst still maintaining control over what you do personally. So he could sell the 'use' of his product to them, to badge as whatever they want, and still run his operation personally, open source, or whatever..
That would be great in this case - however, you never know what malicious licensing schemes Amiga was willing to offer and accept. :crazy:
-
Hi Dennis,
Many thanks for the description of Chunky and Planer displays, it always bugged me as to what they where.
Also, It was great to see you at BB4 and play with the MiniMig :)
Like I said to you at the time, it so reminded me of my A1000 when I first booted it up last week.
Kind Regards, Michael
aka rockape
The guy with the Guru Meditation Tee Shirt.
-
Why would anyone want to do business with Amiga Inc ?
I very much doubt any deal with them would have any positive effect. You will just get burnt. I find it completely mad that after ALL these years a few words from Amiga Inc and people are going "oh.. not so bad". Have you been asleep ?
-
It's better than nothing.
-
Why would anyone want to do business with Amiga Inc ?
Minimig without the Amiga Name - 10,000 units
Minimig with the Commodore name - 50,000 units
Minimig with the Amiga (and Commodore) name - 1M + units
There you go, Dennis could get rich, Amiga Inc could get some additional funding and marketing kudos, Commodore likewise and Dennis could still retain control of the IP in minimig.
To still be here, Amiga Inc and it's personnel must have totally deconstructed and reconstructed. I would advise you not to underestimate them or assume they are the same company as in 2000-2003. :-)
-
Minimig without the Amiga Name - 10,000 units
Minimig with the Commodore name - 50,000 units
Minimig with the Amiga (and Commodore) name - 1M + units
There you go, Dennis could get rich, Amiga Inc could get some additional funding and marketing kudos, Commodore likewise and Dennis could still retain control of the IP in minimig.
To still be here, Amiga Inc and it's personnel must have totally deconstructed and reconstructed. I would advise you not to underestimate them or assume they are the same company as in 2000-2003. :-)
Depends. I suppose an Amiga joystick type thing might sell. The C64 one sold really well. I'm sure Jeri is a millionaire now (not).
I think you live in dreamland though. They are not the same company? No, they are the all new great Amiga Inc right? Please point to some evidence that anything has really changed (for the better).
There is no "future" in Amiga software/hardware. It's a hobbyist market. Like with the c64. There will be no huge "comeback" of Amiga powered devices. Wake up!
-
buzz wrote:
Minimig without the Amiga Name - 10,000 units
Minimig with the Commodore name - 50,000 units
Minimig with the Amiga (and Commodore) name - 1M + units
There you go, Dennis could get rich, Amiga Inc could get some additional funding and marketing kudos, Commodore likewise and Dennis could still retain control of the IP in minimig.
To still be here, Amiga Inc and it's personnel must have totally deconstructed and reconstructed. I would advise you not to underestimate them or assume they are the same company as in 2000-2003. :-)
Depends. I suppose an Amiga joystick type thing might sell. The C64 one sold really well. I'm sure Jeri is a millionaire now (not).
I think you live in dreamland though. They are not the same company? No, they are the all new great Amiga Inc right? Please point to some evidence that anything has really changed (for the better).
There is no "future" in Amiga software/hardware. It's a hobbyist market. Like with the c64. There will be no huge "comeback" of Amiga powered devices. Wake up!
A tough man with tough words.
But you are dispelling the worth of the retro games market. Here's a hard fact:
http://www.allbusiness.com/north-america/united-states-california-metro-areas/488532-1.html
"Intellivision games generated $25 million in worldwide sales last year, through $20 CDs that run on the PC, Xbox and Playstation 2 formats, and the main product, a plug-and-play device that connects to a TV set."
That's right, out of date concept with out of date titles... a shoddy system... admirable for only one thing : Scamming 25 million bucks from the general public. It didn't take much googling to find a story like this, I am sure there are plenty more.
Whether it be WalMart stores or global telemarketing, there are certainly too many ways to make money out of retro devices. Most of the devices are shoddy, buggy and have limited games and controls that break in minutes. But hell, there's money to be made.
Jarrod.
-
My point of view is the following :
Although the Minimig released as opensource (or whatever called to hardware stuff) is a good thing, because it can be further developed, it may lead to a zillion of variants : with AGA support, with chucky support but no AGA, with chucky support and AGA, with chucky support but no AGA and 16 bit sound, etc, etc, ...
So, if some homebrew software developed to run on it (i'm not seing commercial interest ... maybe only a bunch of small companies exploring the niche) i might happen that a title will only run on vernsion X, but not on Y, maybe on W, and who know, on Z ... That is, the fragmentation of something great.
-
That is, the fragmentation of something great.
So long as Dennis keeps a close reign on the "core" MiniMig then it won't be too much of a drama. Look at how Linus works with the Linux community as an example.
-
There will be no huge "comeback" of Amiga powered devices. Wake up!
It's not likely. It's not impossible! :-)
-
Hi Dennis,
Very cool project you have here; can’t wait to see what happens with it in the future.
I just watched the MiniMig Q&A video posted by Oli_hd on amigaworld.net and there’s one thing I was curious about that wasn’t asked in the video (and doesn’t appear to be in this thread) you mention that you added PS/2 mouse support to the MiniMig but you didn’t say if the MiniMig could support more than one PS/2 mouse at a time.
I mention this because one of the things I miss most when using WinUAE is not having dual mouse support when I’m playing games like Marble Madness and Lemmings (two player mode in Lemmings was always a blast :-) ) Is this possible on the MiniMig with 2 PS/2 Mice? Or would you need to connect one PS/2 Mouse and one Amiga mouse?
Thanks for your answer and for all your hard work :-)
Techie
-
buzz wrote:
I'm sure Jeri is a millionaire now (not).
Uhhhh, if she is, I wouldn't mind a smart, attractive, sugar-mama! :hat:
-
I don't think Amiga will ever make a huge comeback to put MS in their place. However, to thrive, grow at a slow but nonetheless steady pace is realistic. Today's market is huge compared to 15, 20 years ago. There's more places to sell-> phones, pda's, internet, retro devices, smart devices etc.
Amiga doesn't have to be the next Atari 2600, iPod or Motorola Razor to do well, they just got to intelligently work their arses off to well enough. (market & advertise like insane greedy fiends... errum, like MS did years ago)
-
techie wrote:
...
I just watched the MiniMig Q&A video posted by Oli_hd on amigaworld.net and there’s one thing I was curious about that wasn’t asked in the video (and doesn’t appear to be in this thread) you mention that you added PS/2 mouse support to the MiniMig but you didn’t say if the MiniMig could support more than one PS/2 mouse at a time.
...
The video i seen from BB4, he did talk about this!
The development board he has now only support one PS2 mouse!
The new smaller baord he is working on will support 2 PS2 mice.
(i think thats what he said) :-D
-
I don't think Amiga will ever make a huge comeback to put MS in their place.
Who cares? The Amiga community is not a company trying to over throw Microsoft.
On the subject of Chunky vs Planer.
For those of you who are still trying to understand it you might try Section 3.1 of this page. (http://home.quicknet.nl/qn/prive/rhm.herold/computing/amiga/amigafaq.htm)
Minimig is copyrightable software not patentable hardware.
You can open source software and put restrictions on the copyrights.
Restriction that keep others from making a buck off it without paying the copyright holder.
-
RickParker wrote:
I don't think Amiga will ever make a huge comeback to put MS in their place.
Who cares? The Amiga community is not a company trying to over throw Microsoft.
Actually, this is the secret plan. But do not talk much about it. As I sad, secret.
:ponder:
-
The video i seen from BB4, he did talk about this! The development board he has now only support one PS2 mouse! The new smaller baord he is working on will support 2 PS2 mice. (i think thats what he said)
Yeah, the question was actually about having two joysticks for multi-player games, and he said that this board only has one standard joystick port and the on-board PS2 mouse port but the new PCB will have two Amiga standard ports.
which I guess could use two mice as well as two joysticks.
-
EvilGuy wrote:
That is, the fragmentation of something great.
So long as Dennis keeps a close reign on the "core" MiniMig then it won't be too much of a drama. Look at how Linus works with the Linux community as an example.
I'm not sure if Dennis wants to take such a dedicated watch over the development of his work by others. Has far I understood, he doesn't for professional reasons and interest, as he said on this thread.
-
Yeah, the question was actually about having two joysticks for multi-player games, and he said that this board only has one standard joystick port and the on-board PS2 mouse port but the new PCB will have two Amiga standard ports.
which I guess could use two mice as well as two joysticks.
@mihcael & @Oli_hd
Yeah, I remember him saying that there is currently only one joystick port and that he planned to have two for the final design but I don’t think he specifically mentioned having two mice working at the same time so rather than assume that they would I just thought I’d ask :-)
I was also just kind of curious about having two PS/2 mice as PS/2 mice are a lot easier (not to mention cheaper) to get than Amiga mice these days.
-
I was also just kind of curious about having two PS/2 mice as PS/2 mice are a lot easier (not to mention cheaper) to get than Amiga mice these days.
One question though, why do you want two mice?
I cant say I know of a game that can use two mice, settlers maybe?
Anyway if he doesnt you could when the source is released next year. :-)
PS: Anyone who wants to see the BB4 vid I have now deleted the vid from my server but on the up side its now on intuitionbase for all to see thanks to Ryu. :)
http://www.intuitionbase.com/multimedia/minimig.mp4 (http://www.intuitionbase.com/multimedia/minimig.mp4)
-
@Oli_hd:
2-player Lemmings used two mice.
-
The full version (115 MB!) of Dennis's presentation at BB4 is also available at Rapidshare: http://rapidshare.de/files/33591202/VCLP0017.mp4.html
Thanks to Oli_hd for filming and Dennis for creating the Minimig :-D
-
wow. rapidshare is really sh*t.
Got halfway and timed out. Now it tells me I have to wait 51 minutes to try again.
Could you upload this file to ftp://www.exotica.org.uk/incoming/
(And I will share it)
-
I've put it up on eMule: ed2k (http://ed2k://|file|Dennis'%20Minimig%20at%20BB4.mp4|121159680|103F89F3872B3D7F6F283333918A9CFE|h=TVQZ4KDZRGOVR5QISADYMYQRYVJUYXTT|/|sources,84.149.129.103:4662|/)
Hmm, watching that video brings up memories of slowwww floppy disks - I think the first improvement should be a way to make the 68k talk directly to the flash ROM to use as a hard disc. Can't be too hard to implement and to program a driver (or boot ROM) for... :-D
-
edonkey ? Do you not think it would just be easier to put it on my ftp, so I can make it so everyone can download from http ?
oh well..
-
edonkey ? Do you not think it would just be easier to put it on my ftp, so I can make it so everyone can download from http ?
Err guys, its on intuitionbase, see my post above :-D
-
You mean that 1.7mb file which downloads ?
-
Just watching the video of the presentation. Really an impressive project. 8-)
Keep up the good work!
-
I've put the video online here for download. Thanks to Zac67 for uploading to me.
http://totem.fix.no/~buzz/amiga.org/
-
i'm not sure if anyone posted this already, but it's pretty amazing:
http://www.seattlerobotics.org/encoder/200006/oven_art.htm
so, dumb idea or punk-rock engineering?
-
@discostu28
And thus the Video Toaster was born.
;-) :lol:
-
450 Degree ?
Does a board (and CPU) handle this ?!
wow
-
I guess this is °F - not °C...
-
I watched the video, looks like a solid clone. Really looking forward to July 2007...
-
I guess this is °F - not °C...
:) ok
-
w00t! Dennis has once again announced that MiniMig will be open-source which is obviously GREAT for the community.
I am sure that most of you have read the news post on Jens Schoenfeldt's newest project in progress; the "Clone-A". I have been wanting to pick your brains on the matter of MiniMig vs Clone-A. Because the MiniMig is open-source and the Clone-A isn't, it would lead me to think that MiniMig just might be the better option for the Amiga community. This is because we could be sure that the community would take interest in further developing the MiniMig in terms of both software as well as hardware. In that case there would be no shortage of free in-depth support and the number of possible 'hacks' would be endless. I also believe that the price should ultimately be lower than that of the Clone-A. What do you all think?
Please don't hesitate to post you thoughts or even correct me, as I might very well be plain shortsighted and missing the point! I welcome discussion!
Regards,
jen-ss (aka Sander)
-
@jen-ss
Hi jen-ss,
Personally I don't see why there should be a "MiniMig vs Clone-A" type situation, in fact I think the two projects complement each other very nicely.
As Mr.Schoenfeld said in his post
[in reference to Dennis and the MiniMig] We're not competitors - we will work together like you would expect members of the Amiga community to do it. As soon as my ITX board is out, I will open up enough documentation for everyone (including Dennis) to port the Minimig source to the Clone-A board. That way, Minimig will have a platform to run on, and everyone will have the chance to compare two Amiga-Clones on the same mainboard.
So the way I see it is this; the MiniMig project will provide people who are interested in FPGA development with the opportunity to improve upon Dennis already amazing project while the Clone-A project will provide people who are not interested in development (or who don't feel like building their own FPGA system) the opportunity to buy hardware that will be capable of running either projects code (again as suggested by Mr.Schoenfeld). So really there is no need for competition at all.
Just my two cents anyway :-)
-
I really like Jens' description of a possible Amiga ITX motherboard. It sounds like a cross between an A3000 and the C-One.
-
Dennis thank you sooooooo much for contributing this project as Open Source to the Amiga community it is really going to fire things up I suspect! THANK YOU
-
So the way I see it is this; the MiniMig project will provide people who are interested in FPGA development with the opportunity to improve upon Dennis already amazing project while the Clone-A project will provide people who are not interested in development (or who don't feel like building their own FPGA system) the opportunity to buy hardware that will be capable of running either projects code (again as suggested by Mr.Schoenfeld). So really there is no need for competition at all.
@techie:
Interesting. But what if there was actually someone interested in manufacturing D's hardware, ready to go... Wouldn't that spell competition? If there is really money to be made, people are interested.
Dennis thank you sooooooo much for contributing this project as Open Source to the Amiga community it is really going to fire things up I suspect!
@Dingo_aus:
Let's just hope that there are developers out there that like to get their hands dirty!
@World:
Are there developers out there that want to and actually will get their hands dirty?
-
This was in the interview with the guy who is in charge of the Clone-A:
Could you elaborate a bit about the different approaches
between “Clone-A” and the Minimig?
Dennis took the UAE source and the description of the chipset registers to make a forwardengineered
Amiga-compatible chipset. This is not necessarily a
bad move, but like I said, our approach leaves no room for
mistakes, while he has much more freedom for a similar, but
not exactly-the-same implementation. Before we take
the next step of implementing a part of a chip, we do extensive reverse-engineering, for example by writing test-programs and doing measurements on the chips with a logic analyzer while the programs are running. We’re
also doing really strange things to find out what the inner
workings of the chips are, for example programming them in a
way that you’re not supposed to.
This is a bit incorrect -- he didn't just rely on WinUAE as the article suggests. When Dennis was demonstrating the Minimig in the UK, he clearly said that he slowed the clock right down in order to use a logic analyser to make sure that for the given inputs, the correct outputs were produced that were identical to the outputs produced by the Amiga 500.
-
I think the point they are making is that they entirely reverse-engineered theirs while Dennis used UAE source to effectively skip a lot of that work. Like they said, that can hardly be a bad thing. Surely as UAE is an attempt to get it emulated bang-on-accurate anyway it makes sense to use it rather then re-inventing the wheel.
UAE has been developed for years so you are effectively shaving years of work off by using that as a starting point and then just fixing any problems it has.
If it was THAT easy to reverse engineer then UAE would be perfect by now, its good but of course it aint perfect. Clone-A could be making a rod for its own back if they are relying entirely on doing it "the hard way".
-
The Real Amiga chipset was several discrete chips, each with buses and timing, handshaking etc.
Minimig encompases all the chips together. Lots of "backbone" hardware that was irrelevant to making minimig run Amiga software was cut out (or rather never put in).
Just as it was never put into UAE.
Minimig and UAE are a re-implementation of "the system" where as Clone A is re-implementation of each chip.
I still feel that they will be lucky if they get 99% compatibility.
What they need are the original schematics / netlists from Commodore / Amiga.
-
alexh wrote:
The Real Amiga chipset was several discrete chips, each with buses and timing, handshaking etc.
I remember reading many places that the only reason there were three chips, was that the mid 80's technology didn't allow something that complex to fit on one chip. Since it can be done now, so much the better. :-) Unless you absofragginlutely have to run ever demo ever coded. :-) When the Boing & Robocity (pre OS1) can still run under UAE, that's pretty amazing.
-
alexh wrote:
What they need are the original schematics / netlists from Commodore / Amiga.
didn't they lose those? IIRC, commodore had difficulty making AGA completely compatible because they lost the plans or something.
-
Sounds like an Urban Myth. The Amiga IP was the corner stone of commodores company. At the time AGA was written I am sure they hadnt "lost" the OCS/ECS schematics and/or netlists.
They carried on making the OCS/ECS chips right upto 1993! They had different production runs with different mask sets, different technologies, different FAB's, each requiring tweaks of the schematics/netlists.
If we could get the original netlists... or even the MASK streamfiles we could convert back to a form for FPGA.
If any of the boards from the lorraine prototype (http://www.amigahistory.co.uk/prototypes/cbm-lorraine-daphne-agnus.jpg) still existed, and someone had the patience, it would be possible to extract information from that too.
-
alexh wrote:
Sounds like an Urban Myth. The Amiga IP was the corner stone of commodores company. At the time AGA was written I am sure they hadnt "lost" the OCS/ECS schematics and/or netlists.
I had the theory that the AGA engineering data isn't lost it right in the hands of Escoms creditors where should be. I not expert enough on the details of Amiga technology to prove or disprove this. Maybe the experts here can.
Go to the U.S. Patent web site. (http://www.uspto.gov/) Do a search with Escom as the assignee. Look through the Escom patents and determine if Escom was assigned the patent that covers AGA technology.
If it is determined that Escom is the assignee for the AGA technology. We know Gateway bought Commodore Amiga IP but did they ever buy any Escom IP?
-
alexh wrote:
If any of the boards from the lorraine prototype (http://www.amigahistory.co.uk/prototypes/cbm-lorraine-daphne-agnus.jpg) still existed, and someone had the patience, it would be possible to extract information from that too.
All those ICs on the lorraine prototype boards; were they off the shelf parts or were they in themselves little custom chips? Would it be possible (if rather expensive) to make the full custom chips like that again? That would be a cool thing to do just for geek value :-D .
-
I'm fairly sure the prototype Lorraine boards used off the shelf chips. But I doubt that having them would be all that usefull, inspite of their 'geek history' value. I mean, once the first silicon prototypes were made, future incremental changes wouln't have been backwired into the wire-wrap boards. It's also probable that the silicon chips didn't even work exactly like the prototypes.
You probably could replicate the chips on protoboards... cost some $$$ I bet. :-) Back in the day, I built a wire-wrapped Z80 computer that looked a bit like the Lorraine prototype... but much less complex. It even worked, after a fashion. ;-)
-
@jims
Yes the Lorraine used all off the shelf components, however which chips they were and the way in which the were wired up is the important bit. That wiring would be useful as they would offer a faithful representation of the logic inside the chips. Yes you are right that there is probably a difference between the lorraine prototypes and final silicon, but there is almost certainly a BIGGER difference between the MiniMig or Clone-A and the lorraine prototypes.
You probably could replicate the chips on protoboards... cost some $$$ I bet.
What?? I dont understand what you mean? The chips on the prototype boards are just gates (AND, OR, XOR, NAND etc.) the gates and the wiring are replicated inside the FPGA!
-
alexh wrote:
@jims
Yes you are right that there is probably a difference between the lorraine prototypes and final silicon, but there is almost certainly a BIGGER difference between the MiniMig or Clone-A and the lorraine prototypes.
What I'm saying is that even if you made a faithfull copy of the Lorraine prototype boards, if might not work the same as the actual chips. Even the original prototype probably behaved differently than the real chips, due to the process of converting from a hardwired mass of chips into an ASIC. The Minimig & clone-A can be reprogrammed to reflect "observed behavior" ... by someone who, unlike me, knows what they're doing. :-)
What?? I dont understand what you mean? The chips on the prototype boards are just gates (AND, OR, XOR, NAND etc.) the gates and the wiring are replicated inside the FPGA!
Yes, I understand that. I took Pedro7's comment about reproducing the prototype Lorraine boards literally. That he meant building clones of the custom chips on wire-wrapped prototyping boards... for the Geek Value... I have seen some folk build a whole CPU that way.
For me, I'd go the FPGA route... just as Geekish, and gives you marketable experience. :-) That $150 Xilinix starter board is getting more tempting all the time.
-Jim
-
That he meant building clones of the custom chips on wire-wrapped prototyping boards...
AFAIK, there are no "custom chips" on those boards. They are off the shelf bog standard 74LS logic chips with nothing more than a NAND gate (or whatever) in them.
I wish Jens the best of luck with his work. I just hope that he doesnt neglect the rest of his market :(
-
Don't bother replicating the Lorraine....
take the Amiga Nyx apart for the good stuff...
:lol:
-
Dennis!
Any updates?
Are we going to see a minimig product (or HOWTO build a Minimig!) before Jens releases his product?
What are your thoughts! :-)
-
And even more important: will we be able to see the Minimig in Maarssen on december 16th? :-D
-
Hi,
December the 16th yes? Hmmm, that should be possible. I have finally received the new boards and I am busy porting the core (new UCF file and some timing changes due to slower RAM). I hope to have it working in a couple of weeks. I will post a picture of the new board soon!
Dennis
-
@Dennis
Have you given any thought to the possibilities of integrating Minimig on the FPGA on the proposed Samantha board? They would seem a perfect fit if the board ever goes into production.
Ok. Going back to my happy place now. (http://64.33.47.100/images/a1000anim.gif)
-
Dennis, just an idea from a FPGA-evangelist and Spartan-3 Starter Kit 200K owner.
If you still have your "old" Starter Kit-based Minimig lying around collecting dust :lol: :lol:
I've seen on your board that you built a simple customized Clock-circuit to replace the standard 50Mhz Oscillator. Couldn't you try this idea if your familiar with the Spartan-3 's feature.
Using the standard 50Mhz clock and connecting 2 DCM between the oscillator and the minimig.
The settings could for example be:
DCM1 DCM2
50Mhz => 2/19 => 31/23 => Minimig System Clock
I would be intereting to see if this is stable enough?
Will there be a release of material Spartan-3-based Minimig version with it's extra boards too? BOM, PCB-layout would be.... NICE! If so maybe the above solution makes it a little bit easier to build for us Starter-Kit Fans!
Best Regards
Magnus
-
Using the standard 50Mhz clock and connecting 2 DCM between the oscillator and the minimig.
I am now using a single DCM as follows:
4.433619Mhz (PAL color burst crystal)=> 32/5 => 28.375162.
Which is the frequency of the crystal in an original A500.
I use that clock as the VGA pixel clock for the scandoubler. dividing the clock by four gives the system clock.
Your setup would give a max clock frequency of 7.09382 Mhz
If I modify your setup to get a 4 times higher clock:
50Mhz => 8/19 => 31/23 => 28.375286 you'd also have a vga clock which is just 4ppm off from the original clock.
No problem whatsoever :-)
(Good tip, would have saved me some soldering.... :-P )
Dennis
-
Will there be a release of material Spartan-3-based Minimig version with it's extra boards too? BOM, PCB-layout would be.... NICE! If so maybe the above solution makes it a little bit easier to build for us Starter-Kit Fans!
Yes.
Dennis
-
Hi All,
Minimig is progressing fast now, so a quick update:
Here (http://www.amiga.org/gallery/index.php?n=1695=13) is the Minimig rev 1.0 board!
Have you given any thought to the possibilities of integrating Minimig on the FPGA on the proposed Samantha board? They would seem a perfect fit if the board ever goes into production.
Yes, Samantha is definately one of *the* boards to run Minimig on.
Dennis
-
congratulations, Dennis! Great job.
-
Dennis wrote:
Hi All,
Minimig is progressing fast now, so a quick update:
Here (http://www.amiga.org/gallery/index.php?n=1695=13) is the Minimig rev 1.0 board!
This looks fantastic. I just wish I could ever get it into my hands. :-D
So you already made a deal with the Italians?
-
Impressive! I need one! :-D
-
pretty! nice work, dude. :-D
-
Totally awesome Dennis!
-
BigBenAussie wrote:
integrating Minimig on the FPGA on the proposed Samantha board? They would seem a perfect fit if the board ever goes into production.
You would need to order a production run of Samantha with large enough Lattice FPGA. I googled this quickly: http://www.ppcnux.com/modules.php?name=News&file=article&sid=6665
so it seems that the producers can offer the board with different sizes of FPGA.
-
Dennis, great work!!!
That looks like a finished product for sale to hobbyists, not some one of a kind project just for your own amusement.
Care to tell us more about your plans?
What are the dimensions of this board, it looks quite small.
Again, congratulations on your work, beautiful design and the huge amount of interest you have generated.
We are all waiting for news of where this will lead.
-
You sir, are my new idol.
-
Impressive! Very impressive! Even more impressive than "OMGWTFROTFLBBQ with gravy" impressive!
Any scheduled date for mass production? :lol:
-
Respect :bow:
Now, what kind of new products are you expecting?
I mean is there a chance to move forward and have cutting edge products?
-
I mean is there a chance to move forward and have cutting edge products?
Something similar to this (MSX on a chip) (http://www.bazix.nl/onechipmsx.html) would be cool. :-D
-
Dennis wrote:
Hi All,
Minimig is progressing fast now, so a quick update:
Here (http://www.amiga.org/gallery/index.php?n=1695=13) is the Minimig rev 1.0 board!
How much do you want for it? :-D
-
@Dennis
@Thread
This looks very promising!
-
Sorry I'm a bit behind with amiga news, so what exactly does this mean, I've read the 1st page and the last two but dont really want to read another 44 pages. Does this mean that we could have new amiga 500 clones with up-to-date hardware?
-
Does this mean that we could have new amiga 500 clones with up-to-date hardware?
Yes.
-
I want one. I want 2.
This project so close to being a dream piece of hardware for any Amigan. Next step AGA. Then. Who knows.
Well done. A spectacular piece of hardware.
-
*drool*
-
Next step AGA. Then. Who knows.
AAA chipset?
-
I don't if it has AAA, AGA, ECS or OCS... I still want a few of them
-
@ odin
Dude I just checked the links in your sig :-D I've gotta buy them games :-D
@ Dennis
Nice job, you make Amigans happy. :-D
-
Hi there,
I am excited about this project. It provides the possibility of new hardware compared to our aging vintage machines.
Bravo on all your hard work and successes so far.
I'd be more than willing to purchase the Minimig as soon as it is made available.
-
I want one!!!!!!!!!
-
humppa wrote:
Something similar to this (MSX on a chip) (http://www.bazix.nl/onechipmsx.html) would be cool. :-D
Very cool!
Was almost tempted to pre-order one of those, but I'd rather have a Minimig!
- Ali
-
humppa wrote:
Something similar to this (MSX on a chip) (http://www.bazix.nl/onechipmsx.html) would be cool. :-D
Yes indeed. Glad you shared the link!
Fester
-
900th post
The v1.0 board looks awesome. How much? :-D
-
just 3 words...
WOW!!!, how much? :-o :-o :-o :-P
-
@ Dennis,
I expected to see some kind of JTAC port on the minimig, how do you update the code in de minimig?
-
@voyager
The JTAG port is on the right side of the board, next to the Joystick ports.
-
That is one pretty pcb. Much respect for Dennis!
-
I have the minimig on my desktop for days, overlookt the JTAG connector :lol:
-
Dennis, this is awesome news!
Looking forward to 'build your own Minimig' kits :D
-
@Dennis
Everytime you post something on here i get more and more impressed with what you've achieved.
Can't wait for you to release MiniMig on the world :-D
Andy
-
First person/company to market with a completely Amiga AGA and ECS/OCS compatible (not emulated in software with E-UAE, but in hardware like the Minimig) PSP or Gameboy like device is likely to make some decent money selling them. With thousands of games and other software available for it, it should be an instant success. Yes? Or No?
-
amigadave wrote:
First person/company to market with a completely Amiga AGA and ECS/OCS compatible (not emulated in software with E-UAE, but in hardware like the Minimig) PSP or Gameboy like device is likely to make some decent money selling them. With thousands of games and other software available for it, it should be an instant success. Yes? Or No?
You missed "Maybe" :-D
I think that Dennis really deserves to make something out of this because he's put so much time and effort into getting it this far and that requires a lot of dedication.
It depends on the device in the end, if it's something akin to the C-one (Is that it's name? The C64 one? I keep forgetting!) then it suits several markets and appeals to a wide range of people from nostalgia through hobbyist right to the hackcore hardware hacker that wants a platform to build upon so the market could be reasonably sized and quite lucrative. I dont think anyone will become a millionaire from it but it could supply a decent living for a while.
I'm personally more interested in what it could do with a working community around it in terms of expansion, modification etc. MiniMig could really open up people to developing for an Amiga chipset that, in theory, could begin to expand and change into something new without throwing out the old.
Andy
[edit] corrected typos
-
amigadave wrote:
First person/company to market with a completely Amiga AGA and ECS/OCS compatible (not emulated in software with E-UAE, but in hardware like the Minimig) PSP or Gameboy like device is likely to make some decent money selling them. With thousands of games and other software available for it, it should be an instant success. Yes? Or No?
No
-
alexh wrote:
amigadave wrote:
First person/company to market with a completely Amiga AGA and ECS/OCS compatible (not emulated in software with E-UAE, but in hardware like the Minimig) PSP or Gameboy like device is likely to make some decent money selling them. With thousands of games and other software available for it, it should be an instant success. Yes? Or No?
No
No? are you sure? the one chip msx (http://www.bazix.nl/onechipmsx.html) has sold out the first batch of 5000 units and that is for and old 8 bit computer, imagine how many one chip Amiga devices could you sell :-o :afro: :banana:
-
@alexh,
I could not find the 5000 unit number that fransexy quoted, but if accurate, times $300+ USD = $1,500,000 USD. Not a fortune, but should be able to recover development costs (I hope). I believe that a Minimig w/LCD at a price similar, or a little higher, could easily sell 2 to 10 times that number of units in it's first production run. That in turn could fund adding AGA (or better, but with full backward compatibility) to the Minimig II. I still stand by my statement that someone, or some company could make a worthwhile amount of money from such products. So I counter with "YES"!
-
The Minimig is a niche product that could be top or flop. I guess quite some people on this forum would be happy to fork out some money to get one (there are quite some collectors out here, judging from the signatures).
I've already thought up a few good reasons/excuses to get me one: The main is being able to run games from ADF files without the hassle to use all kinds of patches to get them to run (or am I too optimistic in this point?).
Support excuses are: 'Sure honey, this little baby consumes much less power so it's actually saving us money on electric-power.', 'Its very nature is like a chameleon so it could change in next to everything I want, so no MORE boxes', 'It's true value for money!' 'Honey, we met through computing, if it wasn't for the Amiga I'd probably never got into computing so we would never have met!' ;-) 'Honey, with this baby I don't need a scandoubler.'
What I'm also interested in is what kind of community this little baby could start. I can imagine that even software-houses could start develop low-level-entry games that bash the hardware, probably even loading a fully adapted 'custom chip-set' into FPGA. Heheh... while thinking about it... Dennis could modify the old Kick 1.x boot screen, replacing the floppy in the hand by a flash-card. :-)
Ah well, I should stop dreaming and get real again. Go back into my 'wait and see' house.
-
clone-a was demoed at amiwest, apparently. so, hopefully, that will result in some kind of commercial product.
if and when minimig goes GPL, you can bet some enterprising fellow will build and sell 'em on ebay.
personally, i feel challenged with the prospect of soldering a MOD chip in my sega saturn, so i'm kind of banking on it :-D
-
The One Chip MSX is
1) a COMPLETE system, no strange/old/rare hardware is needed to use it (although old cartridges and control pads CAN be used)
2) cost effective when compared with a classic MSX computer upgraded to similar specs
IIRC, the price in Japan is ~$180, but from Bazix in the Netherlands it's more like ~$245
If the same thing can be done with Amiga that would be great.
Also, if someone were to put a 68020 or better CPU on an MSX cartridge, plug it into the OCM, and run an Amiga FPGA core in there, that would be cool too.
-
A song for Dennis.
Minimig christmas (sung to White Christmas)
I'm, dreaming, of a laptop
Thats run on FPGA chips
Gonna kill the Lemmings
And F A 18
And get lost in Fairie Tale
lost again.
I'm dreaming, of Amigas
Smaller then my floppy drive
Put my disk collection
On tiny flash cards
And make all, my, memories be bright.
ok, ok, the poetry is poor. But its something I just threw together quickly.
-
@Amiduffer:
:-D To remember Freddy Mercury I was thinking somewhere in the lines of:
'We want, we want MINIMIG! MINIMIG!'
or The Hooters:
'.... and if you've got an Arcade-stick, God's gonna buy you a Minimig!'
Ah well, Dennis, keep on going! How's the first batch performing btw?
-
I could not find the 5000 unit number that fransexy quoted, but if accurate, times $300+ USD = $1,500,000 USD. Not a fortune, but should be able to recover development costs (I hope).
After everyone takes their cut I doubt they will make back their NRE. Have you got any idea how much a MASK set (used for making chips) costs these days? If you use an FPGA you reduce the NRE but double (or more) the cost per unit. Add in PCB design & manufacturing, testing, wastage, VAT / sales tax. If they are selling them at $300 I'd say the first 5000 are going to cost at least $150 each to make. Then add all the costs of selling to the public, taxes, liability insurance etc. If all the developers worked for free (i.e. no salary) then sales _may_ cover the NRE and give them all some beer money.
But at $300 it is a NICHE product, get any of your sums wrong and you could be left with a big bill and / or a lot of units you cannot sell.
To make serious money with Direct2TV products you need to go for the $40 to $90 for the "impulse buy" market and sell through the supermarket chains.
The C64DTV for example: manufacture (including all packaging etc) and shipping to stores must have cost less than $10 each to be able to sell at the price it did.
-
I Have been reading all 46 pages :crazy:
The Amiga has been reborn.
This is the best news sins C= released amiga computers in the 80's.
Not Bad for a fellow dutch guy, (goed gedaan joggie :lol: )
Is this new pcb design also working okey now?
Dennis your the man that's gona save the amiga comunity, finaly some original 68k hardware.
Count me in, I want to buy two kits....
-
I have been working on the new board the last couple of weeks. I have ported the core from the old prototype to the rev 1.0 board and adapted the core so it supports the extra connectors and ram.
I have set up the 2Mbyte of ram as follows:
0.5Mbyte for kickstart
1.0Mbyte for chip ram
0.5Mbyte for fast ram (at $c00000)
The $c00000 is normally reserved for so-called slow ram (not accessible by agnus but on the same bus so not really fast) but on the Minimig it is real fast ram.
I have also added keyboard support. The keyboard module is very nice, the num-lock and scroll-lock leds are used for power and drive leds. The Amiga keys are mapped to the gui/windows keys. You can even do ctrl-windows-windows to reset! The page-up/page-down/home/end keys will be used to control the on-screen-display to change the floppy image.
I am now busy cleaning up the code a bit and optimizing some parts that do not synthesize well. (techtalk: the keyboard lookup table should be implemented using blockram, not general logic slices, saves about 60 slices)
After that I can finally implement the on-screen-display.
Some clarifications about the board:
PIC socket:
This is where the pic18lf252 (http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=1335&dDocName=en010276) microcontroller goes. It is a dil chip so it can be programmed outside of the minimig. (pic's are always a little bit troublesome when using icsp at 3.3V)
You could see the pic as the "bios" which can be replaced/upgraded by replacing the pic.
JTAG,
There is a jtag connector on the board but it is only there for debugging/testing purposes. The core and kickstart are loaded from the mmc card by the pic controller.
VGA:
The vga connector is wired as a standard vga connector, giving out a 50Hz/31Khz signal. However, it can also give out RGB-scart compatible 15Khz with composite sync. A special cable connects the minimig to a TV in that case.
Dennis
-
All I say is: RAWR! Keep it going baby! :-D :-D
Any certainty yet if we can see your little baby live on december 16th? It's been a while since I've stood in awe in front of a piece of hardware :lol:
-
great news on your project Dennis. Your minimig really looks tempting. when can i buy one :rtfm: .. :-D
-
Ahhhhh, this is like carrot dangling and I start going nuts. :crazy:
:lol:
-
Can somebody make a short summary on what the minimig actually does (or is there a dedicated website)? As far as i know, it's a re-implementation of the original A500 chipset, right?
However, does this mean that i could combine this with modern hardware (i'm dreaming of a large harddrive with all pre-AGA demoscene material here...) or am i missing something? And what about the OS? Is it going to be emulation?
What can we expect from running demos on this baby?
From what i know so far, i really share the enthusiasm of others. Having new A500 hardware would be fab! Even more so combined with modern storage methods (all my disks are nearing complete corruption).
-
when can i buy one ..
You can already preorder. However, be quick! There are only 3 (three!) Minimigs left to order for the first production run in January.
***Preorder here*** (http://www.webgoddess1.com/images/kidding.jpg)
-
You can already preorder. However, be quick! There are only 3 (three!) Minimigs left to order for the first production run in January.
***Preorder here***
:lol:
I almost lost my breath for a second there, trying to get my mouse over that link as fast as possible.
Fester
-
@z5:
There is a Wiki (http://en.wikipedia.org/wiki/Minimig) on the Minimig.
-
That isn't a wiki on the Minimig, it is just a Wikipedia article :)
-
Dennis wrote:
Some clarifications about the board:
PIC socket:
This is where the pic18lf252 (http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=1335&dDocName=en010276) microcontroller goes. It is a dil chip so it can be programmed outside of the minimig. (pic's are always a little bit troublesome when using icsp at 3.3V)
You could see the pic as the "bios" which can be replaced/upgraded by replacing the pic.
Hi Dennis,
just curious how you transfer data from the MMC card to memory? Have you implemented an SPI connection between the PIC and the MMC card for fast data transfer? What sort of speed/time does it take to copy a floppy image from the MMC card to the system memory please?
I'm very interested in this as I'm currently investigating adding memory card (or even USB) support to a floppy disk drive emulator ( http://jeanfrancoisdelnero.free.fr/floppy_drive_emulator/index.html ) and currently it's either implement USB or MMC support in the FPGA (outside my current skills) or use an external uC to move the floppy image data to the shared memory (but an 8Mhz uC would be slow.. ie: about 30 seconds). SPI seems to be the way to go.
EDIT: the 1541-III floppy drive replacement project might have come to my rescue - it's a similar replacement flopy drive with MMC/SD support and it's open source. The author has schematics and source code (with FAT16 support) and is using an pic18lf2620 (http://www.microchip.com/stellent/idcplgidcplg?IdcService=SS_GET_PAGE&nodeId=1335&dDocName=en010284) (tonnes of internal RAM and huge program space) with SPI connection to the flash cards. very nice!
regards,
Gizmomelb
-
Dessimat0r wrote:
That isn't a wiki on the Minimig, it is just a Wikipedia article :)
Going by the Wikipedia explanation of what a Wiki is, this IS a wiki.
'A wiki (IPA: [ˈwɪ.kiː] or [ˈwiː.kiː] [1]) is a type of Web site that allows the visitors themselves to easily add, remove, and otherwise edit and change some available content, sometimes without the need for registration. This ease of interaction and operation makes a wiki an effective tool for collaborative authoring. The term wiki also can refer to the collaborative software itself (wiki engine) that facilitates the operation of such a Web site, or to certain specific wiki sites, including the computer science site (an original wiki), WikiWikiWeb, and on-line encyclopedias such as Wikipedia.'
But no matter what it is, it can be seen as a very short subtract of what has been going on in this thread ;-)
-
@Dennis
Great stuff - Thanks for the update at the end of November!
You know, it would be cool to have the minimig reside on a USB stick and display on the PC (or Mac, MOS) screen in a window and use the PC mouse, keyboard and HDD.
I guess this would be possible without modifying the verilog and just using a couple of extra Pics, right? :-D
-
@Colin_Camper,
If even possible that would be a completly different approach than the minimig is now. In that way, the minimig will never be a independent device.
-
you may as well just put UAE on a usb drive, if that's how you want to use it
-
Just to add, something that I think might be of use.
If its possible, an IDE port on the Minimig. Unless someone would wish to add one ?
I just think the ability to plug in a HDD would be beneficial, Maybe a DMA IDE controller ? :)
-
whiteb wrote:
Just to add, something that I think might be of use.
If its possible, an IDE port on the Minimig. Unless someone would wish to add one ?
I just think the ability to plug in a HDD would be beneficial, Maybe a DMA IDE controller ? :)
Would that be faster than, say, add a high capacity memory card as a harddisk replacement? Sure, they don't come with as much capacity as a modern harddrive but what's the use of having 320 GB of storage on an Amiga? It could be used as a C D-drive controller... but then... I'd say: Finnish the current design and get it to market, leave some space for improvement and call that Minimig2.
-
Yup, get that sucker out the door. A one gig sdcard is awful lot of storage for such a machine. We will see many hacks when this comes out for sure.
The best thing we can do is buy this product when it comes out. When his device gets out he will have done more for the community than Amiga INC has done in the past 6 years and I hope his bank account will reflect that after the sales.
-
Then there is clone-a... I wonder which:
- Will come out first
- Be better
- Be cheaper
-
koaftder wrote:
Yup, get that sucker out the door. A one gig sdcard is awful lot of storage for such a machine. We will see many hacks when this comes out for sure.
I agree. Considering that Tiger Direct are selling a 2GB SD Card for $12.99 (after rebate) or 1GB for $14.99 (no rebate) then that should be aple to hold Workbench plus your favourite ADF game files.
My original A2000 got by with a 128MB HD and my A1200 did the same, along with an Iomega Zip drive and a selection of 100MB discs.
In fact, I just looked at my WinUAE ADF collection and it comes to 220MB.
-
Everblue wrote:
Then there is clone-a... I wonder which:
- Will come out first
- Be better
- Be cheaper
As I see it, they're designed for 2 entirely different markets. The cloan-A looks like a complete Amiga computer re-implementation with individual chips recreated seperately in suce a way that the re-engineered parts can even be subsituted into a "real" Amiga.
The Minimig is a compact "cloan" of an A500, designed to be as basic (parts-wise) as possible.
As there are far less pieces to a Minimig, we can assume it would be a hell of a lot cheaper to produce in bulk and ideal for marketing as an Amiga games console.
The Cloan-A project has more possibilites and could be used as a complete Amiga computer replacement, the basis for an updated Amiga, replacement parts for existing Amigas, a "legacy Amiga Chipset on a PCI card" board for NG Amigas or even a full Amiga-on-a-PCI-card option for inserting in a PC.
Both solutions appeal to me in differnt ways, but I'd really like a Minimig in a tiny case attached to my plasma TV for games and I know my kids would love one too.
-
Darrin wrote:
koaftder wrote:
Yup, get that sucker out the door. A one gig sdcard is awful lot of storage for such a machine. We will see many hacks when this comes out for sure.
I agree. Considering that Tiger Direct are selling a 2GB SD Card for $12.99 (after rebate) or 1GB for $14.99 (no rebate) then that should be aple to hold Workbench plus your favourite ADF game files.
My original A2000 got by with a 128MB HD and my A1200 did the same, along with an Iomega Zip drive and a selection of 100MB discs.
In fact, I just looked at my WinUAE ADF collection and it comes to 220MB.
My collection comes out at ~50MB. Thats a few dozen games some apps and utilities and lattice c.
-
koaftder wrote:My collection comes out at ~50MB. Thats a few dozen games some apps and utilities and lattice c.
I have a lot of junk ones that I honestly am not likely to use, but they're backups for my decaying floppy disc collection.
-
Well, will Clone-A be a 100% amiga 500 remake - or will floppy drives be replaced by the possibility to mount .adf files in a compact flash to be mounted. Also, what about an inbuilt IDE support? :)
Is Minimig a reverse engineered a500 or is it based on UAE?
Darrin wrote:
Everblue wrote:
Then there is clone-a... I wonder which:
- Will come out first
- Be better
- Be cheaper
As I see it, they're designed for 2 entirely different markets. The cloan-A looks like a complete Amiga computer re-implementation with individual chips recreated seperately in suce a way that the re-engineered parts can even be subsituted into a "real" Amiga.
The Minimig is a compact "cloan" of an A500, designed to be as basic (parts-wise) as possible.
As there are far less pieces to a Minimig, we can assume it would be a hell of a lot cheaper to produce in bulk and ideal for marketing as an Amiga games console.
The Cloan-A project has more possibilites and could be used as a complete Amiga computer replacement, the basis for an updated Amiga, replacement parts for existing Amigas, a "legacy Amiga Chipset on a PCI card" board for NG Amigas or even a full Amiga-on-a-PCI-card option for inserting in a PC.
Both solutions appeal to me in differnt ways, but I'd really like a Minimig in a tiny case attached to my plasma TV for games and I know my kids would love one too.
-
I'm not sure quite how Dennis would put it so someone can correct me if im wrong.
He stated (on page two of this thread) that MiniMig was made by working out how everything worked from the technical specifications and documentation about the original Amigas and he only referred to WinUAE and WinFellow to help him out when he got stuck or needed to understand something.
Or at least thats what i've understood of it so far.
Clone-A and MiniMig appear to be different approachs that could achieve the same goals.
Personally i just can't wait to get hold of the MiniMig! :-D
Gimme Gimme Gimme Gimme Gimme Gimme Gimme Gimme!!! :banana:
Andy
-
whiteb wrote:
If its possible, an IDE port on the Minimig. Unless someone would wish to add one ?
There might be a shortage of pins left on the FPGA for an IDE interface. Maybe SATA, if there's enough space inside for it.
-
Everblue wrote:
Well, will Clone-A be a 100% amiga 500 remake - or will floppy drives be replaced by the possibility to mount .adf files in a compact flash to be mounted. Also, what about an inbuilt IDE support? :)
Is Minimig a reverse engineered a500 or is it based on UAE?
Don't quote me on this, but from what I understand, the Cloan-A will be effectively a 100% Amiga remake (but you can't badge it as a A500, A600, etc as it can be configured in different ways and use reengineered versions of any chip revision). I'd be surprised if they included a floppy in this day and age, but would probably include a way of combining a Catweasle for those who want one.
The Minimig is not based on UAE. Take a look at the early parts of this thread and Dennis' descriptions and you'll see that it's far more than just Amiga emulation running on a card. :-)
-
Darrin wrote:
Both solutions appeal to me in differnt ways, but I'd really like a Minimig in a tiny case attached to my plasma TV for games and I know my kids would love one too.
Here here.
Both these projects could open doors for the classic enthusiasts we never thought possible. My hat goes off to Dennis (how he's doing this on his own is beyond me) and to Jens.
With all the old hardware slowly degrading, knowing these two are in the works should keep my spirit going that bit longer. :-)
-
I'm using a CF card PCMCIA card in my A1200 by the way, with the FAT95 filesystem driver and CF adaptor driver :) I get 32mb of FAT16 action, which I can play games straight off, and transfer them from the PC, etc -- It rocks :)
I also have a 2GB hard disk in there, but no cable.
-
Dessimat0r wrote:
I'm using a CF card PCMCIA card in my A1200 by the way, with the FAT95 filesystem driver and CF adaptor driver :) I get 32mb of FAT16 action, which I can play games straight off, and transfer them from the PC, etc -- It rocks :).
Eh? How does that work? Does it appear as another drive on Workbench and allow you to treat ADFs as if they were floppies? Can it boot an ADF of a non-standard boot disk or is it only good for "workbench friendly" disks?
-
Oh, it just appears as a normal 32mb disk in Workbench, formatted in FAT16, using long filenames :)
-
JimS wrote:
whiteb wrote:
If its possible, an IDE port on the Minimig. Unless someone would wish to add one ?
There might be a shortage of pins left on the FPGA for an IDE interface. Maybe SATA, if there's enough space inside for it.
I just had a quick look at the A1200 schematics, you only need 6 signals from the chipset for IDE control, the rest comes from the CPU's data and address buses.
-
Doobrey wrote:
I just had a quick look at the A1200 schematics, you only need 6 signals from the chipset for IDE control, the rest comes from the CPU's data and address buses.
[color=ff0000]slaps palm to forehead[/color]
You're right, I don't know what I was thinking... I'd forgot the CPU was external to the FPGA.
-Jim
-
There's 10 spare I/O pins on the current Minimig board so I guess if someone really want's IDE he could go and hack it on?
-
I've just re-re-read the entire 48 pages of this thread (I am now feeling slightly dizzy!) looked at the photos (for the millionth time) and just thought i'd resurrect the thread in time for the New Year :-D
:bump:
Hope everyone is having a nice time day dreaming of all of the new Amiga hardware!
Andy
-
Yay ! Happy New Year everyone :-)
And I hope the weather is better in Nottingham than it is in Mansfield !!!
-
Let's hope 2007 is the year we all get our hands on a minimig!
:-)
-
Dennis...
Something new on the horizon?! How is you're Minimig project progressing? :-?
-
Fantastic idea, software could come on mem cards like sonys Magicgate, or connect to other devices via a usb port.
Very neat handy and portable, could even be made into a great pda. love to see more
Mark South Wales :-)
-
I'd personally like to see this get off the ground but do you think Amiga would let it??
My ideal situation of a Mini Amiga would be one that had:
USB port to connect USB hard drives, USB sticks & USB joysticks
Component out - for people with high def TV's
USB implementation for keyboard and mouse
Faster speeds - default 7.14 megahertz with an option to go faster (for games and software that would support it eg Vista, Frontier Elite) etc.
DIMM socket to expand the onboard Fast Memory to max of 256 megabytes of ram.
I don't know how much of this is possible but would certainly buy it for $200 as the bare pack (no hard drive, no keyboard, no mouse and no memory).
All of the above would be ideal..along with AGA support would make it even better.
-
@tasmanian
I dont believe that there is anything they could do to stop it even if they wanted too.
Besides which Bill McEwen stated that they had spoken to Dennis about commercialising the MiniMig and he turned them down so it could be released in another fashion (Open Source i hope/think) and he seemed fine with that so i believe MiniMig will eventually come to fruition.
As for all the other features you've just listed; i think we'll just have to wait and see what form the intial release takes. From the pictures that Dennis has so far posted it looks like a solid machine that is a literal re-make of the A500 which is no mean feat and good enough for me. Where it goes from there is upto Dennis and if he releases it as open source the rest of the community that forms around it.
Andy
-
AJCopland wrote:
I dont believe that there is anything they could do to stop it even if they wanted too.
The Amiga hardware architecture is copyrighted and you need a license to produce Amiga classic compatible chipsets. The only person I know of who was ever given a license was Mick Tinker, BoXer motherboard creator. I would hazard a guess that under his license (even if it has not expired) he couldnt sub-license.
As to whether Amiga Inc. is the current owner of said copyrights (due to the amount if IP it has sold off in the past) I dont know.
But someone, somewhere, owns it and they would never let anyone produce a commercial product without substantial cut.
Dennis would be better off producing "general purpose" FPGA boards and stealth releasing the MiniMig FPGA files out the back door. Even then he'd have to be careful it might be less obvious if he teamed up with the Atari ST and X68000 FPGA guys.
Money wise it would be better for him to collaborate with whoever is the current copyright owner towards an Amiga DTV. With a good lawyer and the right deal it could get him the deposit for his first house.
-
Well i think you're sort of right except that it doesn't matter if it's copyrighted or not in this case. The MiniMig is a reimplementation not a copy and copyright only covers a specific implementation. This is largely why you can get those C64-in-a-joystick and other old machines. People still own the copyrights on the originals but they've been re-implemented and whilst compatible are also exempt.
It'd be different if it was patented but i dont know of any issues regarding the old OCS architecture that would have been.
Also Dennis has stated that he's not interested in commercialising the MiniMig so there is no "cut" for someone to receive from.
Finally the only problem would be the ROMs however Dennis loads the from the SD/MMC (can't remember which it is right now) so you'd need to get your own from the Cloanto Amiga Forever thingey anyway. Thus avoiding that licencing issue.
I think you're worrying over nothing. Unless you think someones gonna go after UAE, WinFellow etc etc etc as well :-D
Andy
-
AJCopland wrote:
Well i think you're sort of right except that it doesn't matter if it's copyrighted or not in this case. The MiniMig is a reimplementation not a copy and copyright only covers a specific implementation.
Not true. Not true at all. You make a register compatible A500 out of valves and you'll have to get a license to sell it.
This is largely why you can get those C64-in-a-joystick and other old machines. People still own the copyrights on the originals but they've been re-implemented and whilst compatible are also exempt.
Again this is totally untrue. They (Ironstone Partners) licensed the C64 chipset from current copyright owners and the brand name from Tulip. As they would have to by law.
I think you're worrying over nothing. Unless you think someones gonna go after UAE, WinFellow etc etc etc as well :-D
If it was ever commercialised they would.
-
alexh wrote:
AJCopland wrote:
Well i think you're sort of right except that it doesn't matter if it's copyrighted or not in this case. The MiniMig is a reimplementation not a copy and copyright only covers a specific implementation.
Not true.
Surely it depends on the circumstances. The IBM PC was legally cloned because the BIOS was reverse-engineered, and there was nothing IBM could do about it. Wouldn't the same sort of thing apply here?
-
alexh wrote:
AJCopland wrote:
I think you're worrying over nothing. Unless you think someones gonna go after UAE, WinFellow etc etc etc as well :-D
If it was ever commercialised they would.
Ah good then we're safe since it's not being commercialised :-D
maffoo wrote:
Surely it depends on the circumstances. The IBM PC was legally cloned because the BIOS was reverse-engineered, and there was nothing IBM could do about it. Wouldn't the same sort of thing apply here?
That's what i was thinking, its only in the states where reverse engineering has become a tricky problem due to the DMCA isn't it?
Andy
EDIT: messed up the quoting madness :-?
-
The IBM PC was legally cloned because the BIOS was reverse-engineered, and there was nothing IBM could do about it.
AFAIK IBM used off the shelf parts in their IBM PC. Had they added a custom IBM chip that was the central HUB of the design then, as stated in several historical TV documentarys, things would have been very different.
-
That's what i was thinking, its only in the states where reverse engineering has become a tricky problem due to the DMCA isn't it?
Andy
Would the DMCA even be an issue with MiniMig? I was under the impression that DMCA is about circumventing copy protection, whereas Dennis is just reproducing the behaviour of an A500 chipset (as opposed to copying the physical design of the chips.)
-
I dunno, im a software engineer! :-D
Don't we have any lawyers that read this lot? :-?
I don't think that it'll be a problem anyway, Dennis has already been in contact with AmigaInc and everything seems very amicable, besides which the MiniMig hardware doesn't reproduce ANY of the amiga hardware. It is a 68k CPU, some DRAM and an FPGA. No doubt people above and beyond the Amiga community can find numerous uses for such an amazing device. It is however the FPGA that allows it to emulate the OCS (and in future possibly ECS etc).
Until they outlaw emulation I still think the MiniMig is safe, hell the MiniMig could emulate an Atari, or the Sega Megadrive with a different FPGA image loaded.
So unless someone actually has a lawyer that says, "uh oh better stop or you'll be in trouble." we're just speculating without a clue.
Andy
-
I think you're worrying over nothing. Unless you think someones gonna go after UAE, WinFellow etc etc etc as well
If it was ever commercialised they would.
Come on! You guys should know better than this. Emulators are not illegal. And, as has been pointed out in this thread already, hardware may be patented but COPYRIGHT has nothing to do with it.
-
@DamageX
Exaclty, those were both the points I was trying to make (I just gave up eventually as it was easier :-D).
-
This was my understanding too, as the FPGA is basically emulating the inputs and outputs of say an A500 OCS machine, the way the FPGA is internally programmed is totally different to the logic gates in the actual chips I would imagine, luckily for Dennis :-)
-
Why do you think Dennis did it that way, so he wouldne be buried by beurocratic red tape from whoever's turn it is to own Amiga this week.
-
maffoo wrote:
Surely it depends on the circumstances. The IBM PC was legally cloned because the BIOS was reverse-engineered, and there was nothing IBM could do about it. Wouldn't the same sort of thing apply here?
You can trace signals all day long, but reverse-engineering a technology and re-implementing it does not magically make you the patent owner. You still need to get manufacturing rights. What's important to understand here is the pantent's claims. There are companies who do reverse-engineering for the sole purpose of finding patent infringements. It's a very lucrative game.
I really want a Minimig, in whatever form it comes. But I wonder what hurdles - other than implementation - will be necessary before this product is in our hands...
Fester
-
Come on! You guys should know better than this. Emulators are not illegal. And, as has been pointed out in this thread already, hardware may be patented but COPYRIGHT has nothing to do with it.
True, but you can't use WinUAE without ROMs. You won't get far with the emulator packaged in Zeta either without ROM files.
-
you think Amiga would let it??
I think any entity that would use lawyers or any other intimidating means to interfere with Dennis development of Minimig would be publicly lynched and have their PR with Amiga community (which is their target group?) to go into oblivion. Ie bad idea even if the law says ok.
My ideal situation of a Mini Amiga would be one that had:
The consensus seems to be to get it out of the door. Instead of implementing feature X. I don't think Dennis is eager to do any major reworks now that the board is already produced. There's proberbly other issues just to get it working alright that need to be dealt with first.
USB port to connect USB hard drives, USB sticks & USB joysticks
Could be possible to implement on FPGA directly (including PHY). However it's not likely to be done now.
Component out - for people with high def TV's
If there's going to be highdef, I would prefer DVI.
But I think the VGA output could certainly be reprogrammed into component. (why don't ppl use RGB ??)
USB implementation for keyboard and mouse
Complications, see above.
Faster speeds - default 7.14 megahertz with an option to go faster (for games and software that would support it eg Vista, Frontier Elite) etc.
Could possible be implemented with a tweak to the verilog source provided that a 2x clock is avail on the pcb.
DIMM socket to expand the onboard Fast Memory to max of 256 megabytes of ram.
Nice idea, maybe Dennis care to say anything on this?
Otoh, would proberbly require some pcb reworking.
All of the above would be ideal..along with AGA support would make it even better.
The focus is to get it out of the door now asfaik.
Btw, Dennis do you have any updates..? :-)
-
True, but you can't use WinUAE without ROMs.
Actually I believe that nowadays you can. Of course it won't be "100% compatible with everything", but I guess that some games & demos work anyway.
Also it would be possible to do 3rd party clone of classic amiga rom, don't know if anyone is working on something like that.
-
Jupp3 wrote:
True, but you can't use WinUAE without ROMs.
Actually I believe that nowadays you can. Of course it won't be "100% compatible with everything", but I guess that some games & demos work anyway.
Also it would be possible to do 3rd party clone of classic amiga rom, don't know if anyone is working on something like that.
Or one could run AROS :-)
-
Fester wrote:
You can trace signals all day long, but reverse-engineering a technology and re-implementing it does not magically make you the patent owner. You still need to get manufacturing rights. What's important to understand here is the pantent's claims.
Copyright and patent are two different things. Copyright can't prevent someone implementing something compatible with your software/hardware. Patents can prevent you from making an apparatus implementing a certain feature, but AFAIK all patents concerning the OCS chip set have expired.
greets,
Staf.
-
Minimig... hope it's here sooooon :-)
I had some dull tasks to perform at work today (making technical drawings of 3D models is not the most inspiring thing to do) so I had some time to let my mind slip away a bit (kept on working ;-) ). Where a few years ago I'd go daydreaming about fast cars and goodlooking girls, I've stopped doing that as I'm married, I now had some thinking going on about the possibilities of the Minimig hardware.
Most of us (including me) would be happy just to be able to play the good old Amiga 500 games. But... would it also be possible to implement FPGA-code that would be able to play JAVA, ShockwaveFlash and Macromedia games standalone without the need of a posh PC running?
I had this idea because I see the kids playing the PC, mostly those on-line games looking alot like the good old Amiga games.
-
Fats wrote:
AFAIK all patents concerning the OCS chip set have expired.
Well, it's an interesting subject this Minimig. I hope it happens.
There's an article in the current Total Amiga issue that says it might be released as a kit? That would be fun. I'd enjoy putting it together.
Fester
-
True, but you can't use WinUAE without ROMs.
Actually I believe that nowadays you can. Of course it won't be "100% compatible with everything", but I guess that some games & demos work anyway.
Hmm nowadays?
How can it be done?
Is there a "open source firmware" ROM ?
As far I know the ROM is critical to the Amiga in every aspects ...
-
ptek wrote:
True, but you can't use WinUAE without ROMs.
Actually I believe that nowadays you can. Of course it won't be "100% compatible with everything", but I guess that some games & demos work anyway.
Hmm nowadays?
How can it be done?
Is there a "open source firmware" ROM ?
As far I know the ROM is critical to the Amiga in every aspects ...
For demos and games that take over the system and hit the hardware directly. All you really need is code to read the boot sector and start executing the code.
Hans
-
No ROM, no Amiga! Even emulated or reimplemented.
Even games that booted would make use of at least some OS calls from ROM.
The only exceptions are OS4, AROS, MOS etc which reimplemented the ROM internally.
This is why AROS for 68K Amigas needs a kick start replacement if it is to be binary compatable with OS3.x
bty - How is the great ship Minimig going!
-
This is why AROS for 68K Amigas needs a kick start replacement if it is to be binary compatable with OS3.x
Both bounties are still open by the way:
Kickstart replacement phase 1 (http://thenostromo.com/teamaros2/index.php?number=23)
Kickstart replacement phase 2 (http://thenostromo.com/teamaros2/index.php?number=24)
-
Unfortunately it seems a big task, that nobody want's to do at the time.
I'll first wait for the "dos packets" to be completed, as far as i know it is essential for an "Amiga OS" compatible AROS.
Bye
-
Hmm, you mentioned the 250mb Ram socket.
A guy for a while was building a Coldfire upgrade for the A4000, but didn't really have the time and money needed to get it working, and so I believe it's pretty much vaporware these days. Oliver, correct me if I'm wrong!
The system did however, have some good concepts: a 220mhz Coldfire V4, a vast amount (512MB!?) of SDRAM and integral controllers for USB, Firewire, Bluetooth, FIR. . .
If the design for this system is open sourced I can't see a huge difficulty in applying those upgrades to the minimig motherboard, to give a Coldfire-based Amiga OCS mobo that could be used, perhaps not as a Xeon-Beating supercomputer, but at least make a standardised "Classic" Amiga release? Heck, the boards look small enough that with an LVDS connection, a battery controller, a CF socket and a couple of other odds and ends (Say. . . Ethernet, a couple of sockets for a kickstart and USB) this motherboard could make a pretty neat reference design for a "Classic" Amiga portable. (Say. . . the 3200 mAh 3.7v Battery from a HTC Universal, a 10" Wacom serial penscreen: Drivers are available for Amiga after all, or yer bog standard trackpad, an SD/MMC socket to act AS a floppy disk drive, instead of a image loader for one, I'm sure it's possible with a little creative patching. . ., and perhaps a few external serial ports, it's an Amiga after all.) Hmm. . .
That's a thought. . I wonder if you could work with the MiniMig hardware and a few extensions (Second graphics chip, think Picasso [Or more likely something a little newer, although it doesn't have to be able to support more than 1600x1200 at 24bit] and a few serial ports, maybe even an Amiga floppy disc drive *Catweasel* and of course the required CF/Microdrive) and you could have great fun with an Amiga-based dedicated Graphics computer. I happen to miss Deluxepaint, rather that than modern code sometimes, and oooh. . . I'm rambling, but this looks like fun:
Custom Minimig with OCS VGA, Modern VGA/DVI, Serial Ports, Fast CPU and Memory, 3.1 Boot Rom, Amiga-compatible Floppy Disk Drive, Microdrive Hard Disk socket, USB and Ethernet:
Happily connectible to a serial Wacom, (I have two UD1212's on order ATM) a nice big TFT, maybe even a pair, one good one for the editing chipset, and a cheap one for the OCS output, depending on how you do it, say an 8GB Microdrive or 80GB (not a typo) iPod drive: Just what you can make work; USB for connecting Flash drives, Cameras, etc.
Continue that spec with your Ethernet, I have a copy of iBrowse, don't you? :p a couple of drivers for the Graphics Chipset and Ethernet. TCP/IP and USB stacks.
I could be dreaming here, but any thoughts appreciated, It's always fun to get flamed at for going crazy and drooly!
-
Lets try to keep this in the realm of reality.
Sure, anything is possible given the time but I dont think Minimig is going to suddenly turn into Supermig. I always go the impression that Minimig means simple, the basics, not just a definition of the size of the PCB but the whole concept of the project being to keep it to the basic possible specs for running most Amiga games/software. So talking about FastRAM, VGA graphics chips, super large storage drives, it just seems to be drifting pretty far off the goal of Minimig. If it becomes open source then perhaps, but I do not believe it will be Dennis doing those things. At this point just adding the AGA chipset (even ECS?) sounds like it would be a pain and THAT is a bigger priority in my opinion. Percentage wise little Amiga software needs more an A500 to run hence why that is the goal for Minimig.
At this point it would be cool just to have the basics and I for one would just like being able to plug this tiny little box into a TV and play my classic Amiga games.
As for an Amiga PDA. It wont be long before PDAs are fast enough to run WinUAE (I suspect they already COULD run Fellow had it been ported, it used to run many games properly on Pentium 200MMX PCs) properly so the idea of an Amiga PDA would seem more plausible in software than hardware.
Quite frankly, my HTC Wizard (PDA Phone) can already browse the web superior to any Amiga web browser I have seen so I do not see the point in worrying about running iBrowse on Minimig. Most if not all people have something far superior already.
The Minimig to me is gold dust because my A1200 takes a HUGE amount of space and I cant fit it near the TV with PCs, Xbox 360s and other modern hardware that I use far more often taking up that space.
-
alexatkin wrote:
As for an Amiga PDA. It wont be long before PDAs are fast enough to run WinUAE (I suspect they already COULD run Fellow had it been ported, it used to run many games properly on Pentium 200MMX PCs) properly so the idea of an Amiga PDA would seem more plausible in software than hardware.
I agree that the Minimig should be just finished as it is right now. Again, if Dennis turns her into OpenSource the comunity can do anything (well, about...) with / to her they like. It could go quite extreme I think. Afterall, you're not limited to load the FPGAs with instructions to impersonate and Amiga OCS. But time will tell what's all possible.
Now, a little off-topic, the UAE discussion. I've read many reports of people having a 'fast AMIGA' on Windows. I've tried UAE on machines up-to AMD Duron 1400 (the fasted PC I owned until I got me a Apple) with 768 MB RAM and Creative LIVE soundblaster but felt it always lacked big time in the sound-division. Games like 'The Lost Vikings' just ain't fun with bad sound. So how can UAE run nicely on a Pentium 200?
And a bit more on topic: The trouble the AMD had to emulate an Amiga always make me wonder how on earth an Amiga can feel like an Amiga when heavy emulation has to take place.
Now... for the Creator: Dennis, any progress?
-
So how can UAE run nicely on a Pentium 200?
Probably it can't. Fellow, however, is more lightweight and can run the games (it works with) much better, just like the original poster pointed out.
Haven't tried PSPUAE too much, but I heard it runs some games nicely, too bad I happened to test just the slow ones :-)
-
Now, a little off-topic, the UAE discussion. I've read many reports of people having a 'fast AMIGA' on Windows. I've tried UAE on machines up-to AMD Duron 1400 (the fasted PC I owned until I got me a Apple) with 768 MB RAM and Creative LIVE soundblaster but felt it always lacked big time in the sound-division. Games like 'The Lost Vikings' just ain't fun with bad sound
Never owned a K7 Duron, but my old K8 Turion MT @800Mhz** (with 1MB L2 cache) runs WinUAE 1.3.3 A500 profile reasonably fine e.g. BattleSquadron, Hybris, Elf Mania.
With Amikit 1.2.x/WinUAE-JIT-020, the above setup can easily AmigaAMP MP3 @max decode settings.
**Forced "under clocking" via RightMark CPU clock tool, 1024MB DDR400 memory, AMD RD480 Xpress chipset, AMD Mobility Radeon X700.
-
A guy for a while was building a Coldfire upgrade for the A4000, but didn't really have the time and money needed to get it working, and so I believe it's pretty much vaporware these days. Oliver, correct me if I'm wrong! The system did however, have some good concepts: a 220mhz Coldfire V4, a vast amount (512MB!?) of SDRAM and integral controllers for USB, Firewire, Bluetooth, FIR. . .
Yup, t'was me, and I am still working on it (the last prototype was shown at the BB4, late last year) however time is a real problem...
If Dennis does upload the code to aminet for the great minimig I will have a play with it.
But yeah until I finish the CF card consider it vapor.
The firewire was a TI chipset thats availble now, up to 400mbps if the bus clock is high enough.
the USB was the Cypress 811HS, which is an 8bit USB host controller, there is already an Amiga Zorro card using it.
The IRDA was just off the UART, the coldfire supports that as standard. The card didnt have bluetooth though.
And the SD-Ram was part of the coldfire core, although Xilinx give away IP cores for SD-Ram and DDR ram interfaces for FPGA's.
-
Hey good to hear from you Ollie, it's been a while, hi!
Of course my rant was all on the condition of the design being open sources. The minimig project as it stands looks to be a simpler system, equivalent to an A500, which is fine, because most people interested would just want one to play lemmings and captain planet.
On the condition of it being open soresed, then even I might have a play with it! I was never much of an Amiga gamer, I was more the hacker kinda guy, with cables going everywhere. And my tomato project of course. ;)
In terms of the Minimig, I'd need something more powerful, and after all. . When I have my Coldfire based 512mb dual VGA Amiga, with it's 60gb hard disk (More likely 20gb I hae one of those lying around) then hell, I can run UAE ON IT!!
What's always important to me is a powerful system, still within the Amiga realms. After all, I have a Wacom (with any luck) to use in Dpaint V and a Midi interface arriving in a few days. . . Shame to waste it!
I'd still love the coldfire expansion boards, but they're getting further and further behind and the 68K platform is rapidly being replaced by ARM. Motorola have even sold their silicon company. . .
The most powerful Coldfire to date is after all, only 266/133-DDR, with twin Ethernet.
Compared to the Dual-Core Intel 4GHZ stuff,
The 9 Core IBM 3.2 GHZ stuff,
16 Core 1.2 GHZ SPARC,
16 core PAS Unknown PPC,
and even Freescale's own iMX31, which is an INCREDIBLE processor, with 520mhz, integrated 3D graphics accelerator, FPU, video-in bus, DDR266, USB2 host, Twin-SD, etc etc etc. . .
M68K just can't compete anymore.
I think the Amiga is dead. This is kind of a last ditch project to show that it can go out with a grin. . :(
-
The most powerful Coldfire to date is after all, only 266/133-DDR, with twin Ethernet. Compared to the Dual-Core Intel 4GHZ stuff, The 9 Core IBM 3.2 GHZ stuff, 16 Core 1.2 GHZ SPARC, 16 core PAS Unknown PPC, and even Freescale's own iMX31, which is an INCREDIBLE processor, with 520mhz, integrated 3D graphics accelerator, FPU, video-in bus, DDR266, USB2 host, Twin-SD, etc etc etc. . . M68K just can't compete anymore.
Except the V4e running at 266Mhz is only £15 each at x100 quantities.. :)
It hasnt won against a PC in Mhz terms since the 486/P75.
The key Amiga thing is what you do with what you got. :idea:
I think the Amiga is dead. This is kind of a last ditch project to show that it can go out with a grin. . :(
The Amiga is a retro computer now, its not dead, its reached the end of its development.
I think its safe to say that most proper computer people today will remember their first computer, and occationally want to relive that, and not with emulators, roms and such, they want something they can hold.
That is why the minimig is so great, its an A5 sized Amiga that plugs into your TV and lets you play those old games you played in the 90's.
Sticking DDR ram, terrabyte drives and the like in is just hacking it, Im not into pimped up cars, computers or anything, if it has enough ram or harddrive space, wether its 1meg or 1024 meg, thats fine, if Im only using 1% of the memory thats a waste.
I would love to see the Minimig available as a direct to TV thing, with a single USB (could be connected PS2 ish inside) for keyboard and mouse and a TV out. A SD Card slot would be cool too but thats it.
Stick all your Amiga games on an SD card, sit back, and play cannon fodder until your thumbs fall off. (Actually my best Amiga moment was probably looking through the fredfish CD that came with my CDTV)
Err anyway Im at work so I had better stop here and get something done. :)
-
Hi!
I only got as far as the first 2 pages of this thread but Ive seen enough. Where do I get a minimig? :-)
-
I only got as far as the first 2 pages of this thread but Ive seen enough. Where do I get a minimig?
He wont be selling a complete board, although wether he finds someone else to do it I dont know, there is talk about it going open source, so you can make your own computer from a bag of bits. :-)
-
@Oli_hd
I think you just summed up what people have been saying throughout this thread. We just want the MiniMig, as is, right now please :-D
Andy
-
AJCopland wrote:
@Oli_hd
I think you just summed up what people have been saying throughout this thread. We just want the MiniMig, as is, right now please :-D
Andy
Let's not rush Dennis. If he needs until next week then it's quick enough ;-) .
@mel_zoom:
You haven't read the best pages of this thread, iirc a lot of doubts were fired back then. There's a nice little summary on Wikipedia about the Minimig. It can be found on Wikipedia here (http://en.wikipedia.org/wiki/Minimig).
-
Dennis, any progress?
Yes!
There are now 4 working Minimig boards. The first one is here on my desk and the other three have gone to colleauges and friends. One of them is currently integrating his Minimig inside a small case complete with 5.4" TFT and stereo speakers. It looks almost like a portable (although it has no power and needs external mouse/keyboard/joystick)!
On the core front much has been done too, the OSD display is finally working. It allows the user to select/eject a floppy in DF0: and to reset the Amiga. The OSD works much like a nokia display from the PIC microcontroller point of view. The resolution is 128*64 pixels. (8 lines of 21 characters).
Also, keyboard support has been fully debugged and optimized. The whole PS2 keyboard interface now fits into 49 slices and one blockram. This includes LED support, keyboard reset and OSD control.
This pretty much means that I've completed my initial goal, recreate a console-like A500 to play games on my TV. But, as with all projects of this complexity it is not finished. There are still some features missing and some bugs remaining. To completely fix this will probably take me years. So, all things considering I think the core is ready for an open-source release. I only need to figure out the proper method for releasing the core. Opencores.org? Aminet? Own website? GPL? Any thoughts?
Anyway, the current state of the core is this:
Synthesis report:
Selected Device : 3s400pq208-4
Number of Slices: 2682 out of 3584 74%
Number of Slice Flip Flops: 3122 out of 7168 43%
Number of 4 input LUTs: 4337 out of 7168 60%
Number of bonded IOBs: 134 out of 141 95%
Number of BRAMs: 12 out of 16 75%
Number of MULT18X18s: 2 out of 16 12%
Number of GCLKs: 3 out of 8 37%
Number of DCMs: 1 out of 4 25%
This includes the following:
Agnus (1038 slices with a LOT of room for optimisation)
Denise
Paula (including floppy replacement)
Gary-equivalent
ciaA/ciaB equivalent
Amber-equivalent (scandoubler only)
PS2 mouse emulation
PS2 keyboard emulation
Automatic 2nd joystick/mouse switch.
Overlayed OSD generator
Missing from the core is the following:
parallel port
floppy write support (should be simple)
harddisk/ide support
Still standing bugs:
some silly stuff with ddfstart/ddfstop
floppy incompatiblity with some trackloaders
some silly sprite timing issuse
That is it really. I will be attending the next meeting of the Commodore user group in Maarssen to demonstrate the new board and hope to meet some of you there!
Dennis
-
Dennis wrote:
So, all things considering I think the core is ready for an open-source release. I only need to figure out the proper method for releasing the core. Opencores.org? Aminet? Own website? GPL? Any thoughts?
*snip*
Dennis
Opencores.org, Aminet or your own website... hmm, all of the above :-D plus Amiga.org of course.
If you decide to use your own site then remember to have a PayPal donate button so that the bandwidth costs don't drive you out of house and home when we all come stomping along to download it!
Awesome news and amazing work Dennis.
THANKYOU!!!
Andy
edit: Forgot to say that if you put it on Opencores.org then would it make sense for the whole thing to go under the Prototype Boards section like this FPGA board (http://www.opencores.org/projects.cgi/web/eus100lx/overview)? Including board layout as well as the core.
-
Wow, that's some great news indeed. Now I *really* should come to the next meeting in Maarssen (only problem is that my car is dead, so maybe I'll take the train).
Any chance to do another production run and actually sell a couple of Minimigs at the meeting? :lol:
-
@Dennis
I was really hoping to dedicate my 10,000th post to your work, in this very thread but this news has forced my hand early.
You sir, are a fecking genius!
*takes off hat and bows*
-edit-
Talk about synchronicity though, this is the 1,000th reply in this thread :lol:
-
@ Karlos:
But you did post the 1000th reply to the Minimig thread!
EDIT:
OOps, should have read better, you noticed :crazy:
-
I was planning to write something along the lines of
"Imagine designing a new amiga, hardware-register compatible with the original machine.
Some have said it could not be done. Others have said it perhaps could be done, if the original design information were available, but what would be the point? Some even tried (remember the boXer?). Somebody just went ahead and did it. And that somebody was you."
Oh well :-D
-
When I see a rush of new posts and the Xilinx avatar, I know something good awaits in this thread :)
Well done, Dennis.
-
"Imagine designing a new amiga, hardware-register compatible with the original machine.
Some have said it could not be done. Others have said it perhaps could be done, if the original design information were available, but what would be the point? Some even tried (remember the boXer?). Somebody just went ahead and did it. And that somebody was you."
Where do I buy one?
coldfire
-
It's not that it's compatible with the original Amiga, as I said in a previous post -- various aspects can be really pimped up, and still remain compatible with original games (number of sprites, amount of colours, resolution, memory, etc) :)
We may even see new Minimig-only games.
-
@Dennis
I think you deserve a new avatar :-D
(http://extropia.co.uk/img/dennis.jpg)
-
I have put a new wad of offerings on my Dennis altar. Enjoy =).
-
@Dennis
The only problem with your announcements is the inevitable waiting afterwards :crazy: I'm really looking forward to actually being able to get all the pieces and build a MiniMig
@Karlos
How on earth do you get so much detail out of the tiny size limit we have for avatars?
Andy
-
@AJCopeland
Draw it large and shrink it down, then touch up a few pixels by hand.
Unless it's an animated one which requires working at the correct scale all the way through for the best results.
-
@Dennis,
Outstanding!!!!
I can't wait to build one!
Thank you Dennis - You are the man! :bow:
Regards,
AmigaEd
-
@ dennis
More pictures please!
-
@Dennis: well done! Fantastic that you've been able to do this!
I hope an end product will surface that some of us can buy (with proceeds going to your good self of course).
- Ali
-
I think we should all club together and buy Dennis this t-shirt (http://www.thinkgeek.com/tshirts/generic/8f52/) in recognition of all his efforts :-D
Andy
-
when its released hey ?? :)
-
@Dennis
This is simply Fantastic! Congratz!
It would be nice to get some more pictures of the board, OSD and your mate's case (if he has nothing against it). This would make the waiting easier.
Cheers,
yak
-
I only need to figure out the proper method for releasing the core. Opencores.org? Aminet? Own website? GPL? Any thoughts?
Yeah go for all of them. How big is the archive?
Thank you so much !!!
-
@Dennis
This is great news.
We all look forward to playing with this!
Opencores.org? Aminet? Own website? GPL? Any thoughts?
I imagine Aminet makes most sense initially for the files. Hell, all of them make sense! :-D
What about the PIC code - I suppose you would need to supply that minus the ROM binary somehow. Would it be possible to add a ROM socket to the Minimig board so that the PIC could read in the ROM code from an old A500 or a ROM from Amigakit. I would be extremely interested in producing a kit - but I couldn't see an easy way of doing it (legally) without a ROM socket! (Or talking to Amiga Inc for a license :crazy: )
Good luck!
-
I want one (or more), so when Dennis finally releases Minimig in whatever form or license, how do we get the PCB produced like the four boards he had made, so the rest of us can gather all the parts together and get his code to complete a Minimig of our own?
Do we need to create a group to work together to pay for exact replicas of Dennis' board in a run of 100, 200, ??? to lower the cost of producing them?
Will Dennis suggest small changes to the board design he made for the next production run?
I am excited that Dennis has (for the most part) completed this project and is now ready to release it to the rest of us so we can take it further forward in any way we decide is best for our needs.
Very exciting times for Amiga enthusiasts, don't you agree? :-o
-
amigadave wrote:
Do we need to create a group to work together to pay for exact replicas of Dennis' board in a run of 100, 200, ??? to lower the cost of producing them?
Hopefully we can get someone to manufacture these for us. Otherwise there's batchpcb.com. They seem pretty cheap but the limit is 2-layers.
Hans
-
the PCB overlay Dennis released for v1.0 is single layer, double sided. I seem to recall Dennis saying previously he's gone out of his way to design a simple and cheap to manufacture board for enthusiasts.
The board also isn't very big, so I wouldn't expect small production runs to be that expensive from your local PCB manufacturer. I'd also expect there will be various enterprising techos offering parts or pre-assembled kits for sale at minimal prices, for those who don't know which end of a soldering iron to pick up.
-
@Dennis:
Great to read that the project has been advanced this far already. To quote from a Dutch commercial: 'Goed gedaan jochie!'
You say that there are the occasional oddities in the behaviour of Minimig. Are those oddities software related or hardware? If it's the first then a 'release' of the hardware should be a possibility at the state it is right now, wouldn't it? If more people get their hands on the Minimig I guess the oddities could be corrected easier as well as more people can dive into it. A 'simple' update of the code to program the FPGA should then be sufficient.
@The interested party:
It seems like a good idea to team together and try to arrange a batch production (once Dennis has released the hardware). Not everyone has the skills to hand-solder SMD components so it makes sense to get a pro to do it.
To coordinate such an effort, a central meeting point on the net would be needed I guess. This not only prevents the 'project' to explode in countless directions, it also serves as a good point for support. Also, it should give Dennis a nice idea of how many Minimigs would've been built and used and it could also have a nice little PayPal button to donate a (d*mn... what's the english word for 'vrijwillig'?) fee to show how grateful the people are for this little wonder of technology (this could mean that Dennis has to found the Dennis van Weeren Foundation to escape Dutch taxes). Volunteer... that's it, the word.. ;-). It wouldn't need its own URL or something, could be (should be?) part of Amiga.org.
-
You say that there are the occasional oddities in the behaviour of Minimig. Are those oddities software related or hardware? If it's the first then a 'release' of the hardware should be a possibility at the state it is right now, wouldn't it? If more people get their hands on the Minimig I guess the oddities could be corrected easier as well as more people can dive into it. A 'simple' update of the code to program the FPGA should then be sufficient.
Exactly, the oddities are really not that big. The problem is that I lack the time to correct them. Also, especially people who used to code the Amiga on the hardware level (demo-coderz, crackerz) will probably immediately spot what's wrong.
the PCB overlay Dennis released for v1.0 is single layer, double sided. I seem to recall Dennis saying previously he's gone out of his way to design a simple and cheap to manufacture board for enthusiasts.
It is a double sided (2-layer if you like) board. I've ordered the boards here (http://www.thepcbshop.com/). I have payed about 20,- euro per board. The only thing that prevents Minimig from being mass-producing friendly is that I have mounted some components at the bottom side of the board. For DIY soldering that is not a problem however. Another good move would be to port the Minimig to an universal FPGA board. The C-One comes to mind but the FPGA's of the C-One are probably too small. The TREX C1 (http://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&CategoryNo=39&No=14) seems a good alternative. If one would design a simple daughterboard for the expansion slot containing the 68000 and some joystick ports, a Minimig can be built.
Dennis
-
I would be happy to produce standalone minimigs as well as TREX C1 daughterboards for hardcore minimig fiddlers!
I still think the standalone minimigs would be hard to do, legally, without ROM sockets. I would be happy to modify the PCB and PIC to facilitate this. Is the PIC code assembler or c?
-
Getting the PCB isn't all that a big problem, however the mounting of the components would be I think. I at least would be stumped on how to get the SMD components on there.
Perhaps it's a better idea to (if at all anybody wants to do a small production run) let the surface mounted components be 'pre-assembled'?
@Dennis:
How viable is a version of the board with non SMD components? Is the FPGA at all available in such a form? Perhaps as a socketed version?
-
Colin_Camper wrote:
I still think the standalone minimigs would be hard to do, legally, without ROM sockets. I would be happy to modify the PCB and PIC to facilitate this. Is the PIC code assembler or c?
Dennis has already said that the minimig loads the ROM image from the MMC card. That way he does not get into any legal trouble as he is not supplying the ROM and the minimig board could be used as a basis for any number of other projects. There is NO need for a ROM socket for the Amiga KS at all.
Please just let Dennis release all of the details/source first, if you really want to add a ROM socket then you can modify the project yourself for your own use.
-
I hope this is not too off topic.
I wonder if Amiga OS4 could have a verilog or VHDL datatype?
Could it be possible to program a FPGA on the fly?
I guess Minimig will find itself programmed on the FPGA chip without external chips someday.
Rick.
-
@gizmomelb
Dennis has already said that the minimig loads the ROM image from the MMC card.
Hah! I forgot that! Man, it's hard to find these snippets in this gigantic thread. Thanks for pointing that out! :-D
-
@Thread
Now I feel the minimig is getting closer and closer why don't we have a Minimig modders wishlist thread!
My top three would be; IDE, IDE and IDE!
What about everyone else! AGA?, Accelerator? :-)
-
A wishlist?
Mine would be, parts arrive -> i assemble parts -> feel warm glow of accomplishment -> take over TV and play :-D
As for actual mods. Probably IDE and then a larger FGPA for a vector floating point unit to be added.
You just can't believe the speed difference going from integer maths to a full vector floating point unit with matrix operations. So much more power that it's unholy :crazy: and if there's room in the FPGA then it'd be a nice add-on for the MiniMig. Totally useless for old stuff but great for coding new stuff :-P
Um, yeah. After that I guess incremental stuff, like going from OCS to ECS and onto AGA. Maybe add better chunky mode support. Then another leap upto a higher order 68k cpu like a 68030, 68060 or Coldfire (not strictly compatible but hey we're dreaming here right?) would be... delicious
For now... can I just have the MiniMig as is? Pwwweeeeeease?
Andy
-
Colin_Camper wrote:
My top three would be; IDE, IDE and IDE!
SATA? At least then there is a chance of an Amiga somewhere using relatively modern technology :-)
-
@Everybody
On the occasion of my 10,000th post, I'd just like to congratulate Dennis on what I think would only be fair to describe the most impressive user project I've ever seen!
I had planned to write a few words, but ended up blurting them out earlier. Still, I'll just reiterate them for posterity.
@Dennis
Ever since the demise of Commodore and their manufacturing base, there have been endless discussuions about creating new amiga hardware that would be compatible with the classic machines.
Most of these have pivoted around the wish to have a new chipset, despite the obvious advances that have been made in graphics cards etc. Even so, that's never deterred people from wanting it. Their arguments usually begin something like "Imagine designing a new amiga, hardware compatible with the original machine. Only it could be made better if we did this, that and the other as well!"
Well, some have said it could not be done. Others have said it perhaps could be done, if the original design information were available, but what would be the point? (Big Dave himself has often commented from this angle). Some have even tried - does anybody remember Mick Tinker's boXer project?
In spite of all this debate, somebody just went ahead and did it. And that somebody was you.
Not satisfied with merely doing it, now you are talking about making it open too, thereby ensuring that people will be able to take the idea forward, even if you yourself for any reason do not.
I take my hat off to you sir!
-
Karlos wrote:
I take my hat off to you sir!
Well said Karlos. Agreed.
Excellent work Dennis.
-
Vector floating point ???????
Good luck coding the software for it :).
You do realise you can drop a 68882 onto the bus of a 68000 dont you ? I recall seeing the circuit diagram in Motorola's data sheets for it. Sure its not a VECTOR floating point, but its Amiga :)
-
@whiteb
Coding the software? I already do :-D my day job is a games programmer and I spent about 6 months of last year writing MIPS32 and ALLEGREX assembler for the PSP and it's custom vfpu. Now I'm doing it again for another project that's on Wii, PS2 and PSP. All three with different types of vector units :crazy: the.goddamn.joy
Anyway yes sticking a 68882 onto the board is a real possibility as well. I've never coded for it so I wouldn't know but I doubt that it supports any kind of vector or matrix transformation operations does it?
Taking code that runs using the PSPs fpu and porting it over to it's vfpu can produce an astonishing speedup if done carefully. Which is just the point of custom hardware.
It's something that I would like to see added somewhere down the line (by 2009 maybe?) so that gives the MiniMig an expanded function set.
Andy
-
whiteb wrote:
You do realise you can drop a 68882 onto the bus of a 68000 dont you ?
IIRC, it only hooks upto the 68000 as a memory mapped peripheral so you wouldn't be able to use any '882 instructions directly, but it could be possible to roll some replacement maths libraries to take advantage of it.
Anyway, this is drifting away from Dennis worshipping.
:bow: :bow: :bow:
-
@doobrey.
http://www.freescale.com/files/32bit/doc/fact_sheet/BR509.pdf
Page 15
@dennis.
:bow: :bow: :bow:
-
Different advantages/cons:
Opencores.org? - HDL Community
Aminet? - Amiga community, lack hdl knowhow?
Own website? - Administration chores/bandwidth
Sourceforge - Versioning system
I think any distribution site you use should:
* Not make claims to the source.
* Have versioning system!
* Bandwidth & no {bleep} hoops & loops.
* Be multiplatform friendly (Linux, FreeBSD, MacOSX etc..)
* CVS/Subversion interface?
If you want to allow commercial deratives then the license ought to be BSD. To enforce contribution back will make a GPL a good case (GPLv2?).
It might be beneficial to allow people to build boards and distribute the hdl sources as a package. But I think GPL will work with this too.
The only case where I can see GPL being a hinderce is if someone wants to make additions and sell it as a sw-package.
I checked with a Swedish assembly company and they charge as follows:
440 EUR In startup costs.
8 EUR Per board.
One needs to provide the paste screen aswell.
And it would save lot's of money to have a single component side. Optionally just mount the SMD components and let buyers do the connectors, power regulators etc..
As for fixing bugs. I think it's better to concentrate on hardware bugs as they are harder to fix post-production. While software can be fixed by many-eyes-watching(tm).
If it works good enough now, then the community can work out the quirks.
As for the board itself. It might be useful to have at minimum solder pads for an ZorroII/III bus to allow expansion with ordinary mortal periphials.
"odin"; There are no fpga in non-smd format worthwhile at this moment. The current "non-smd" option is BGA which is a real royal pain to deal with :crazy:
Now we are just waiting for that source .tar :)
-
"voluntary"
:-)
Matt
amiga_3k wrote:
also have a nice little PayPal button to donate a (d*mn... what's the english word for 'vrijwillig'?)
-
whiteb wrote:
@doobrey.
http://www.freescale.com/files/32bit/doc/fact_sheet/BR509.pdf
Page 15
Ta for the link, been a while since I read it..honest guv !
Page 5 ;-)
-
Hi guys,
Please excuse if i bum this up. This is my first posting. I have been following this thread since the first few posts. As an Australia past Amiga owner and current lover, this project will hopefully offer many of us a way in which to rekindle and get in touch with our past. Anyways enough ranting.
Someone previously mentioned the BoXer and it brought up some fond memories of my efforts in offer the Unnof BoXer Site.
The question now is, why hasnt one been started for the MiniMIG. This is a great project that has produced an awfull lot of pics, info and this huge thread.
Why no support site? I mean this project has certainly developed to the point where perhaps it should require one.
What say you all....and Dennis.
Darron
-
I expect Dennis is too busy to do a support site.
I'll be happy to make a start on one - I can (just about!) get xoops running and can produce static pages in Dreamweaver. I would be happy to have help with this.
If it's O.K. with Dennis I could arrange wamp hosting and make a start!
Bty - can anyone suggest a reason why Harald Frank would want this domain; ???
Domain: minimig.de
Domain-Ace: minimig.de
Nserver: ns.serverkompetenz.de
Nserver: ns2.serverkompetenz.de
Status: connect
Changed: 2007-01-08T05:26:12+01:00
[Holder]
Type: ORG
Name: VMC Harald Frank
Oh! and Welcome to AW.Net Grunter!! :cheers:
-
Bty - can anyone suggest a reason why Harald Frank would want this domain; ???
Domain: minimig.de
Domain-Ace: minimig.de
Nserver: ns.serverkompetenz.de
Nserver: ns2.serverkompetenz.de
Status: connect
Changed: 2007-01-08T05:26:12+01:00
[Holder]
Type: ORG
Name: VMC Harald Frank
Interesting find!
Huh, Harald planning to commercialize the Minimig?
-
It might be interesting for Cloanto to present a deal to Dennis...
-
Colin_Camper,
Bty - can anyone suggest a reason why Harald Frank would want this domain; ???
Domain: minimig.de
Domain-Ace: minimig.de
Nserver: ns.serverkompetenz.de
Nserver: ns2.serverkompetenz.de
Status: connect
Changed: 2007-01-08T05:26:12+01:00
[Holder]
Type: ORG
Name: VMC Harald Frank
More info here?
http://www.vmc.de/
Specifically:
http://www.vmc.de/ger_vmc_web_projekte.html
-
humppa wrote:
Bty - can anyone suggest a reason why Harald Frank would want this domain; ???
Domain: minimig.de
Domain-Ace: minimig.de
Nserver: ns.serverkompetenz.de
Nserver: ns2.serverkompetenz.de
Status: connect
Changed: 2007-01-08T05:26:12+01:00
[Holder]
Type: ORG
Name: VMC Harald Frank
Interesting find!
Huh, Harald planning to commercialize the Minimig?
No idea, but http://www.minimig.de/ does hold photos of the Minimig. Just click on the img folder. I almost hope that he is planning to make Minimig boards. I don't have the time to make one myself. I'd rather get straight into messing with the Verilog code.
Hans
-
jorkany wrote:
More info here?
http://www.vmc.de/
Specifically:
http://www.vmc.de/ger_vmc_web_projekte.html
I wonder if Frank is aware that he linked to this site: (http://www.amigax86.de/amigax86_nonsense.html)
"While we were approached early on by Harald Frank in the spirit of cooperation to advance AmiThlon, Harald never delivered anything. Extreme ended up in the middle of the early controversy between the "partners" of AmiThlon. We
managed to avoid the mudslinging and legal issues. Bernie turned out to be a delight to work with, and delivered some unexpected enhancements in the product rather quickly. We worked with Amiga Inc. for the latter half of 2002 to try and resolve all the issues, and finish the product."
:lol:
-
What is that I see on the horizon? Could it very well be the official minimig release? How long still? When will the hurting stop?!?
-
Was anyone at the Commodore Show in Maarssen on the 17.02.07?
Does anyone know when it will be released?
-
Here are some pictures of the Minimig in action. There were two Minimigs at the meeting. One was built inside an LCD monitor. We played lotus 2 with the two Minimigs linked through the serial port.
link:
HCC february meeting (http://commodore-gg.hobby.nl/feb2007col1.htm)
Dennis
-
:bow: :bow: :bow: :bow:
-
Serial link works too! Awesome, I dunno why but I hadn't even considered it. ArmourGeddon co-op used to be so much fun :-D
Andy
-
@Dennis:
Just viewed the pictures on the HCC pages. Especially the 'LCD' version shows how small she really is! Amazing! From those pictures it seems rather likely that it could be used to entertain the back-seat occupants on a long holiday trip. Now, if only it could be available before may this year ;-)
-
Just viewed the pictures on the HCC pages.
Darn! The perspex minimig looks so simple and cool.
I can't wait to try it out!
Dennis! When are we likely to see a release version - even just files! This would be great to play with. :-)
-
::continues to sit patiently on a 1 million gate FPGA waiting for Minimig to get released::
I wanna build an Amiga Laptop, I have an LCD (too small but it'll do), my old Ami HD (the Ami died), a 2GB CF card, and a butterfly keyboard that all really want to talk with the FPGA.
Of course I'll have to modify the Keyboard translator, the butterfly keyboard speaks some weird 3.3v RS232 protocol instead of PS/2. It was made for the Sony Clié T-Series.
As for the LCD it is intended for Composite or S-Video, but there is an RGB mod for it. (Would be less risky to change the video output rather than the LCD interface, but the RGB will provide a better picture). At some later date i'll get a actual portable LCD monitor to use instead.
-
king awsome,i wonder if one would fit into a old sega handheld,I have 3 laptop shells waiting for this board an Amiga laptop has always been a dream of mine.
well done Dennis
-
ivier wrote:
I wanna build an Amiga Laptop
Same :-D I think that'd be awesome!
Looking forward to this coming out!
Thanks Dennis!
- Paul
-
Maybe Dennis might write history....
CNN Newsflash:
The past few weeks has seen something never seen before.
Prices of ancient laptop systems all of the sudden go sky
high. eBay, main source of ancient laptops, says it probably
solved the mistery. A spokesman:
"Some dutch dude by the name of Dennis has re-created the
good old Amiga computer. The new form factor of this system
seems perfectly for building an Amiga laptop, something
never done by, the now defunct, Commodore. Seems many Amiga
users still want a portable Amiga. Back to Washington."
Would be interesting to see if the release of Minimig will cause price increases.
-
Hi Dennis, when you you plan to release it? :)
I`m waiting 24h/day. Anyone else?
-
gimme, gimme, gimme, I want, I want, I want :-D
-
Wow, you got my attention man but you haven't posted a reply in a good while... too bad.
If you need ANY resource I can provide such as custom chips from the OCS to put them under the eyes of a logic analyser then I will help you.
If you need a beta tester for NTSC... just ask and I will provide detailed reports.
-
Where's the best place to get a MC68000?
Is it best (cheapest) to salvage one out of an old Amiga?
Just thinking ahead for the parts needed for the Minimig
- Paul
-
http://cgi.ebay.com/MC68000-Microprocessor-BRAND-NEW-Motorola-MC68000P8-CPU_W0QQitemZ7621752701QQihZ017QQcategoryZ50916QQcmdZViewItem
Like this ?.
Now thats assuming that Dennis is using the DIP package for the CPU. Or is he using PLCC ?
[edit] okay a recheck of the Rev1.0 PCB its PLCC surface mounted style (PLCC with the pins bent out to surface mount position) so that Ebay auction is useless. :(
-
Ah thanks whiteb, im guessing the ones in the old amigas are dips too. I guess i'll have to find one of the plcc ones, alternativly I could try re-arranging the pcb.
-
Isn't it a 68SEC000 in a 0.8mm pitch quad flat pack?
-
mrmkl wrote:
Isn't it a 68SEC000 in a 0.8mm pitch quad flat pack?
Ah yes, thanks mrmkl! :-D
-
Link: MiniMig News on the iAMIGA Search Engine (http://www.google.com/custom?hl=en&safe=off&client=pub-0987631822357778&channel=1773060945&cof=FORID%3A1%3BAH%3Aleft%3BS%3Ahttp%3A%2F%2Fwww.iamiga.info%3BCX%3AiAMIGA%3BL%3Ahttp%3A%2F%2Fsearchisland.googlepages.com%2FiAMIGAmodified.jpg%3BLH%3A52%3BLP%3A1%3BLC%3A%230033FF%3BVLC%3A%23FF0033%3BGALT%3A%23666666%3BGFNT%3A%23AAAAAA%3BGIMP%3A%23AAAAAA%3BDIV%3A%23FF0033%3B&q=%22Minimig%22+more%3A2006_-_2007&btnG=Search&cx=002218398616714700785%3Awfjswecfzy0)
-
Any news?
-
Where's the best place to get a MC68000?
Out of an old Mac! :lol:
-
Digi-Key appears to have 64pin QFP 68SEC000 in stock.
http://www.digikey.com
( MC68SEC000AA16 MPU 32BIT 16MHZ 64-QFP Freescale Semiconductor 64-QFP $12.18000)
:-)
-
@Dennis
Could you tell us something about your relationship with VMC (Harald Frank) who set up this website (http://www.minimig.de/) including a Minimig webforum? (http://www.minimig.de/forum/)
Are there any plans by VMC for production runs of Minimig after the design is finished?
-
I'd say Harald Frank is a business man, and probably looking into producing Minimigs and selling them off to people who can't make them themselves.
-
I hope this gets made and stays open source. At firt I though a company like Jakks Pacific should do an Amiga TV game like the pac man and Namco versions Jakks has made...
BUT...I don't think a toy company would take this and develop especially if Amiga Inc are so hard to deal with. If a toy company did get a hold of it, I don't think they'd be too keen on Amigan's purchasing them to hack their product into computers after paying Amiga Inc's licensing fees.
I am surprised Chinese toy companies have not reverse engineered the Amiga yet... They did it with Nintendo and Atari 2600. A pal worked at toymax (original developers of the TV game concept) and they said they got idea for TV Games after seeing the readily available bootleg or "fanban" (Chinese slang for bootleg) versions sold on the streets and shops of Hong Kong... They had a hard time getting it to work initially but the factory engineers just kept going at it.
I wonder if a factory in China would develop such a beast provided they can take the opensource design and apply it to actual product use.
Would this work like the Linix opensource licensing?
Hope this stays opensource though. Could be a great platform to teach game development or hobbyist platform.
Imagine an AGA version with 040 or 060 speed powering a media player or retro game system.
Good Luck, hope this ships sometime soon
-
@Dennis:
I was there at the demo in Maarssen... VERY impressive!
-
DrWhoop wrote:
@Dennis:
I was there at the demo in Maarssen... VERY impressive!
And did Dennis say something about availability or VMC?
-
Hi Dennis,
have you already got a release date because you wrote that it should be soon.
Plz answer! :-)
-
And did Dennis say something about availability or VMC?
I believe he did mention plans of making it "open source", but I didn't get all the details, as I wasn't standing very close to where it all happened.
-
Dennis, why don`t you answer the questions? Do you still plan to release it?
-
He's probably just busy. Don't rush the master :-)
-
Sorry if this has been answered, but there are 54 pages now. Does this project have an official homepage anywhere? I keep checking every few months to see if this project is finished. I'd love to get my hands on one. I'd even be happy to assemble it myself if the PCB and FPGA code were available. I found the article on wikipedia, but does not point to a homepage.
-
On the one hand;
I find the idea that Repo01 can be impatient with Dennis shocking and disconcerting. :-?
If I were, in any way, disrespectful towards Dennis then I imagine that I would spontaneously combust! :madashell:
On the other hand....
I'm glad he asked the questions! - I find the waiting is getting harder and harder! :boohoo:
-
Dennis asked for input on how to release it on the same post that he said it is ready to release, i don't see many offering him feedback on the best release method. And on the best open source licensing model. So perhaps if more responded to this, it would get released sooner.
Opencores - The premiere open HDL site
Sourceforge - The premier open source anything site
Aminet - The Amiga Community
Wiki - a documenting system, but could also be used for distribution
Other - most likely a custom site
While opencores would be more appropriate, it's still a immature site, and I am uncertain of how long it will be around (another site often referenced is freecores, which no longer exists).
Sourceforge has the longevity we need for a distribution point, but they can be intolerant of non-mainstream open source licenses (such as custom written ones).
Aminet has the Amiga fandom to do Minimig justice. As well as major backup capabilities. However very few Amiga users are likely to have the ability to configure the FPGA themselves. (though the capability could be built into Minimig itself theoretically)
While a wiki is not strictly speaking a distribution system, it could be linked to a CVS/SVN hosted on either Sourceforge or Opencores, providing full documentation on Minimig, and allowing people an easy forum to post their own mods to the core.
Custom sites provide the most capabilities and options but to get those you also have to go through the most headaches. In my experience coders and hardware engineers don't make the best looking websites.
As for Licenses
There are many possibilities.
GPLv2 - GNU Public License (most common but somewhat viral)
LGPL - Lesser GNU Public License (non viral)
BSD - Berkley (Allows relicensing provided license file remains in tact in all distributions)
MPL - Mozilla Public License (i don't know the specifics of)
VBPL - Vorpal Bunny Public License (made to be fair and just for BOTH sides)
Custom - Usually the best choice, but takes a lot of thought on exactly what it should cover and how it should be worded. (courts are very picky about wording which is why the GPL is so fricking long)
My personal choice on this is the Vorpal Bunny Public License, it's not well known and has not been tested in court, but it's clearly written, and gives room to allow using tiny pieces freely, while still encouraging updates to the code from outside sources
-
@ Ivier
Whichever method is chosen, I hope it is one that doesn't prevent someone from producing a commercial "for profit" Minimig product. I think that most of us don't want some sort of DIY kit, but rather a ready to go product or at the very least a fully assembled motherboard. For that we require a third party to produce a product and to encourage that we need to allow them to make a healthy return on their investment.
This will probably upset the "free meal" brigade, but it will be the best solution for the Minimig when it comes to distribution.
-
ivier wrote:
My personal choice on this is the Vorpal Bunny Public License... [/quote]
Would that be because you wrote it?
I managed to locate a copy of the VBPL here:
http://camerahacks.10.forumer.com/viewtopic.php?t=537
I'm assuming by the email address listed there you are one and the same person.
Not a bad effort for a license. Only negative point about it in relation to Minimig is its single author bias in alot of the conditions (ie. changes must be emailed to author for inclusion in next release). That could pose problems with a multi-author project which I think Minimig could possibly become. It could also possibly pose problems with future forks of Minimig, if that was to happen.
I think a license like the LGPL could be a good choice for Minimig as it allows both free and non-free components to 'link' against it while still compelling additions/bugfixes to be contributed back to the project.
-
This will probably upset the "free meal" brigade, but it will be the best solution for the Minimig when it comes to distribution.
I agree. The way it should be done is released as an open-source project so people can built it themselves and add improvements as time goes by. A perfect example of why to do this is to look at the Commodore DTV and what has happened there. it was a great product and now it is dead because of legal battles between the parties involved in making it. Had it been open-source, this wouldn't be the case.
However, it should ALSO be available as a commercial product, made with a custom board and case that people can buy without having to solder anything. In fact, if some kind of arrangement could be made with Amiga (the corporation) so that it could be distributed with licensed Amiga workbench and stuff would be great. In fact, if there were enough demand, one might go for converting the code to an ASIC design so that the whole thing might be produced for like $30. Then TONS of people would buy it.
-
Dennis, why don`t you answer the questions? Do you still plan to release it?
Sorry for the slow replies, I didn't follow the thread so closely lately. I will try to answer all questions:
Do you still plan to release it?
Yes, Like I said before. Minimig will be released as open-source, including the schematics and layout of the rev1.0 (reference)board. The planned release date is June 2007 at latest.
Where's the best place to get a MC68000?
I got mine from Digikey. I am actually using the MC68SEC000 in a QFP-64 package. The MC68SEC000 is a 3.3V compliant, embedded variant of the original 68000.
Any news?
Not much...
Could you tell us something about your relationship with VMC (Harald Frank) who set up this website including a Minimig webforum?
Good question. Harald Frank contacted me in a very early stage to ask my permission to create a Minimig fan-page. He also offered to set up a small production run of Minimigs. Although I declined his second offer, I did allow him to create a fan-page. But, I am not completely happy with the way this turned out. The www.minimig.de domain name suggests it is the real, official homepage of minimig (.de .nl and so on sound like company domain names too me), which it isn't. I am also not happy with a dual-langauge forum. English is *the* langauge of the internet and an international support/development forum should be completely in english so that the information is accessible to everyone. So, in short, there is no relation-ship between Harald or www.minimig.de and me.
I was there at the demo in Maarssen... VERY impressive!
Thank you. The new board with on-screen display looks much better doesn't it?
have you already got a release date because you wrote that it should be soon.
See above.
Does this project have an official homepage anywhere?
Not yet, but I am now realizing that I really need a homepage :roll: .
I am trying to work out the best option for me. Like I said in a previous post, I need some place the dump all the files. I checked out opencores.org, but frankly, opencores.org is badly organized. (Anyone ever tried to download the actual vhdl/verilog files of a project there?) Aminet should be OK, but an official homepage is probably best. Stay tuned!
@ivier
I share your opinion on opencores.org but what do you mean with GPL being somewhat viral?
Dennis
-
The GPL license is potentially quite nasty - from memory, it basically stipulates that any project that uses even a single line of GPL code in any way shape or form must also be covered by the GPL.
This was why the LGPL (Limited/Library GPL) was created, so that you can write libraries that are GPL but can be used by projects which do not fall under the GPL banner. For instance, a shareware MP3 encoder using a GPL mp3 encoding link library would have to be under the GPL also, whereas one using an LGPL library would not.
I think that's the difference anyway, it's something along those lines....
-
Hi Dennis!
Thank you for your answers! Good to hear that everything is ok. It was the right decision not to allow the production run.
I agree with your homepage idea. It`s the best place to host the files. If you need any help (logo, homepage, instructions, donations ... ) I´ll help and I´m sure others will, too.
-
To save on bandwidth costs it could be a good idea to host the files at one of the other sites, but still have the homepage with links to the files, news & forum, etc.
-
Repo01 wrote:
Hi Dennis!
Thank you for your answers! Good to hear that everything is ok. It was the right decision not to allow the production run.
I agree with your homepage idea. It`s the best place to host the files. If you need any help (logo, homepage, instructions, donations ... ) I´ll help and I´m sure others will, too.
I'll second the thank you for the answers and specially the news that it will be released very soon (by June 2007).
I don't entirely see why a small production run (or for that matter a medium or large run) would be a bad idea though, unless it is just that the person that volunteered is not the right choice. I am still hoping that a kit will be provided that is easy for those of us that have limited skills can also complete a Minimig that matches or exceeds the version 1.0 board that Dennis produced.
-
I would love to see, and believe this project could single handedly create a standard board format. Maybe a MMB (MiniMig Board) format. Basically, dictate, the maximum dimensions of the MiniMig board, and a standard where the mounting holes and connections would be. This way, we could start working on our custom MMB compliant cases. Of course this would not need to be done by Dennis. The work could be done by someone else, as long as Dennis endorsed the spec, and was willing to give input on the design before it became final.
-
I actually regestered here just to post my thoughts on this project.
I can't wait!
The best thing about this has to be that it's going to be open source giving the general public a chance to effectively "make" their own Amiga.
Seriously, I don't think I've been so excited about a project in years!
Thanks Dennis!!
-
I'm Loven' it !!!
:-D
-
It was the right decision not to allow the production run.
What makes you say that?
Anyway - I am looking forward to the release! :-)
-
Colin_Camper wrote:
It was the right decision not to allow the production run.
What makes you say that?
Anyway - I am looking forward to the release! :-)
Hi!
When you give a licence to a company to produce something that you invented :idea: then you`ve got to take care about different things like contracts, lawyers, warranty of the product... and all that costs a lot of money.
When you sign a contract with bad conditions for you then you could lose the ownership of your product or even worse you`re personally liable. :rtfm:
-
@Repo01
Yes, I take your point.
I don't suppose we will see any 'kits' or 'runs' of minimig until Dennis releases it under public license.
I hope this happens asap - I think the minimig community will be a lot of fun - lots of mods and hacking - great stuff! :lol:
-
Dennis, I much appreciate all your hard work on Minimig and I'm looking forward to it's release! I know I'd like 1 or 2 or 3 or, well you get the picture. :lol:
-
The Clone-A project is far advanced. It will be demonstrated on the 03/31/2007 in Aachen, Germany. Quote from Jens Schönfeld from Amigaworld.net http://amigaworld.net/modules/news/article.php?storyid=3641
All Commodore chips of the demonstrated machine will be clones: CIAs, Paula, Gary, Agnus, Denise. The CPU will be an original Motorola. The final chip will also include AGA, but that's the step after next. The current plan includes a major demonstration every quarter with prize money/goods for any incompatible software that someone discovers. We might be able to sell the first FPGA-based units late summer 2007. If you're good at beta-besting and finding/writing software that discovers incompatibilities, the prize money could pay your Clone-A board!
The parallel port of Clone-A works, so 4-player games can be played.
-
Would that be because you wrote it?
HMmmmmmm Could be :D We are partial to our own creations i think. But your right, it is single author centric. I suppose i should make the VBPL2 more thoroughly support group efforts.
-
@ivier
I share your opinion on opencores.org but what do you mean with GPL being somewhat viral?
Dennis
The GPL requires that if you use a small part of a GPL project in your own project, your project must be GPL as well, this is called Viral Licensing. Open source projects typically share fragments of code, accelerating development. Lets say someone who is making a c64 on a FPGA decides to borrow code from your CIA implementation as the 6526 and the 8520 are nearly identical in operation. Under virtually any other license, only the shared bit of code would be covered under Minimig's license. But, if Minimig were under the GPL, the C64 on a FPGA would have to be as well. Thus the GPL infects projects like a virus.
Now i have heard ALOT of people say the GPL isn't viral lately, and maybe the GPLv2 isn't. I haven't sat down with a few hours to read legalese. GPL licenses are notoriously dense! Which is why i hold little value in people telling me it's not viral, when i doubt any of them have read it. When i read the GPL back in the early 90's it most definitely was viral, and at the time i thought that was cool, but hey i was 17.
Which i why when i released my first useful open source project i wrote my own license, to express how i want the code to be shared. A layer friend of mine even said he was quite impressed with it (i plugged a lot of loopholes in very few words).
Now VBPL is still technically an immature license, granted. so if we look at seasoned licenses. I would have to say LGPL as it's the traditionally non viral form of the GPL. And the GNU people pay layers to write these things in legalese, so they hold as much weight in court as possible.
Oh, on a final note. Contrary to many what GPL advocates say, the GPL has NOT held up in every court case it's be through, it has lost in a great many cases in many countries, BECAUSE of it's viral nature. Though this may be the reason behind GPLv2. Like i said i haven't read that one yet.
-
"viral" is a rather strong term to use, it implies disease, illness, badness, etc.
I've never heard about the GPL losing in court, simply because it hasn't been tested many times. It is only a license - one which clearly states that if you use any of the code in your own projects you have to make your code open, etc, as well. This is the right of the software author, and if somebody else wants to use that source code they have to abide by the license, or negotiate their own license with the author.
More fool the closed-source company that uses GPL code in their software in breach of the license. It isn't difficult to understand the GPL at a basic level.
The real issue with GPL software is people using it without really thinking if it is how they want to distribute their software. For some it is - if you've written a tool for yourself, then you can benefit from other people's improvements if you GPL it (and other people actually improve it, hehe).
-
Calling the GPL "Viral" is simply MS 'Viral' FUD. They spread it, and you caught it. The terms of the GPL are not something that 'infects' your code. It is a feature that you choose to add. You don't just wake up one day, look at your code, and realize that your code is sick.
-
Just past the 1100 reply mark... :-o
-
How many of those posts were on topic???
-
Also, has anyone read the entire thread, lately?
-
It is only a license - one which clearly states that if you use any of the code in your own projects you have to make your code open, etc, as well.
Could you point me to the point where it states that? I know you can't, becouse it doesn't state that.
You probably refer to the (often misunderstood) part where it says, that you need to offer sources for everyone who has the binary.
Sounds close to what you said, doesn't it? But on the other hand, it, for example, allows companies internally use GPL powered software without worries of having to "leak" some company secrets outside - as long as no-one leaks the binary.
Any news on the topic btw? :-)
-
Been looking around the web for other FPGA retro computer projects...
In the link below..They mention the Amiga... would be nice to have an Amiga the size of a couple of decks of cards like this one MSX clone...
http://www.bazix.nl/onechipmsx.html
This fellow has a lot of work to do yet..
http://home.freeuk.com/fpgaarcade/atari_amiga.htm
Someone should just start petitioning Jakks pacific to do an Amiga version of their TV GAMEs unit. I'm sure Amigans cand find out how to hack it. They usually want an order of 100,000 units or more to justify tooling for such a beast. That means interest from US retailers like Walmart or Target. Specialty orders of 5,000 units won't cut it.
What games would you like to see on such a device?
I'd Like:
SpeedBall2
Xenon2
Super Stardust (AGA)
Sensible Soccer
Lemmings (not possible as sony owns it)
Worms
Body Blows
Turrican (1,2,3)
R-Type
Banshee (AGA)
Battle Squadron
Syndicate
BC Kid
-
What games would you like to see on such a device?
I'd rather have a Minimig and not be restricted to certain games :popcorn:
-
@Dennis
The planned release date is June 2007 at latest.
Only a couple of months at the most. I can't wait! :-D
-
Colin_Camper wrote:
@Dennis
The planned release date is June 2007 at latest.
Only a couple of months at the most. I can't wait! :-D
@Colin:
And for some freaking reason I believe it's a realistic target. Hope some of the other developers of amazing new hardware take notice! :-D
-
i heard the dutch people can buy it first :P
-
amazing wrote:
i heard the dutch people can buy it first :P
Buy what first? I thought Dennis was only releasing the code and was not going to sell hardware of any kind. Open source was the word I had heard.
-
@Amazing:
amazing wrote:
i heard the dutch people can buy it first :P
It's a pitty that I lack two things at the moment:
1) Time
2) Talent
If I had both, I could see a 'remake' of the beer-commercial of Heineken (no, I'm not advertising Heineken here, I don't like their beer, I like good, dark Belgian beer)... you know the one: Setting - Some space control center. The event - The 'landing' of a planet explorer on The Red Planet. The lander comes to a halt. A expedition leader says: "If there is life, the Dutch will find it" Then a Dutch guy (probably it actually IS Dennis ;-)) presses a button, the explorer transforms into a typical Heineken pub... The guy takes a relaxed position and says: 'Now we wait."
So that's the original...
Now this could be tweaked a bit... The planet should be easily replaced by our beloved Boing-Ball. The expedition leader should say something like: "If there is an Amiga user out there, the Dutch will find him." Then, at the point where the Dutch guy presses the button (he should be replaced by Dennis) the explorere transforms into The Minimig. And then Dennis should say something in the lines of: "Now we get creative"
-
This is my first post here after reading this ENTIRE thread over three evenings, and all I can say is OMFG!!!!
I never believed that this kind of thing could be achieved by a hobbyist, Dennis, you are talented beyond most people's wildest dreams, I truly am not worthy...:bow: :bow:
To all the impatient people out there, re ember that good things come to those that wait.... patience.... you know it will be worth the wait.
@ Dennis, remember Field of Dreams? If you build it, they will come, never a truer word spoken...
If Dennis' creation has got so much publicity just by threads like this and word of mouth, wait and see what happens when he releases it, I can see a few websites seeing DOS attack like hits because of this.... who ever hosts it had better have SERIOUS webhost power....
Dennis, you are a GOD, I can't see how you can top this one...
WotTheFook
-
@Dennis
Now that ACube is releasing announcements, I am crossing my fingers to see an ACube-Dennis announcement one day. ;-)
-
Dennis, you are a GOD, I can't see how you can top this one...
AGA :lol:
-
Are we still on track for a 'before' June release of Minimig!
I sure hope so! Less than 2 months! :-)
-
@Dennis:
I, as a dutchman, know whe've been blessed with shiny, almost summerlike, weather the past few weeks. So, it wouldn't be too much of a surprise if you've been more busy with the bbq instead of Minimig. However, it doesn't take away my curiousity about how development is progressing, what needs to be done? You're still in the race to produce (or atleast deliver the stuff needed to produce) the first 'new' Classic Amiga in over a decade!
Hope I can soon post a small comment here stating nothing more than:
Dat he-j mooi daon!
--Bert
-
I was just logging in to ask the same thing... though without the Dutch :-D
Hope things are going well. Still really looking forward to the MiniMig becoming a reality (for us not for you for whom it obviously already is :banana: ).
Andy
-
Dat he-j mooi daon!
:lol:
I can't wait either, I'm dying to stick this little baby in a nice case next to my TV, and make my Playstation completely useless :-)
Maybe it would be cool to fit it in an old NES or Sega Master System case :-D
-
I was hoping/planning on an A1200 case :-D
Plenty of room for whatever people come up with then and would just feel good to have it in an appropriate housing!
Andy
-
I am wondering if the MiniMig 1.0 board was a success?
If it was fully functional, will the schematics, FPGA pinlist files etc. be open-sourced as well?
Even if the HDL files never get released I'd like to try and make a similar board for home-projects including the open source Atari ST.
I was going to make my own, then MiniMig 1.0 PCB was displayed and I thought I'd hold on for that. Then Mike of FPGA Arcade said he was going to make one and I said I would hold off for that.... Basically I am a lazy git :roll:
-
@alexh
I agree, for several reasons it'd be interesting to see how everything went from the v1.0 made of discrete boards to the final finished design.
I'd guess that the v1.0 system might be better for people who would like to tinker with things as well.
Lazy? No surely not? :-D Lets face it if we weren't all so lazy then someone would have done this. We should all praise Dennis for not being lazy! :lol:
@Dennis
Any chance of an update, despite the talk of new Amiga's in the other threads I'm still more interested in the MiniMig!
Andy
-
AJCopland wrote:
@Dennis
Any chance of an update, despite the talk of new Amiga's in the other threads I'm still more interested in the MiniMig!
I think the Clone-A project is better for the following reasons:
100% compatibility with the original chipset.
More likely to be actually available to buy.
No dicking about with Kickstart ROM files.
Warranty and support.
That is not to say Dennis' work is not amazing though, it is.
-
@alexh:
From what I understood is that the first batch works almost 100%, only small bugs to squash which didn't had to do with the PCB, but with software.
-
@CannonFodder:
I'd prefer Minimig. Reasons:
Small
Flexible
High potential for new implementaions
and most of all
Dutch
-
@CannonFodder
No dicking about with Kickstart ROM files.
This is still the question. AFAIK Amiga Inc. suddenly broke up negotiations with Jens Schoenfeld. Jens was trying to license the kickstart rom from them. Now he is looking into some workarounds, i.e. a replacement kickstart. Using these could result in a much lower compatibility.
-
humppa wrote:
@CannonFodder
No dicking about with Kickstart ROM files.
This is still the question. AFAIK Amiga Inc. suddenly broke up negotiations with Jens Schoenfeld. Jens was trying to license the kickstart rom from them. Now he is looking into some workarounds, i.e. a replacement kickstart. Using these could result in a much lower compatibility.
My understanding is that the kickstart replacements he's looking at ar good enough to boot up hardware-banging games/demos that disable the OS. It's not going to be able to boot up workbench.
Hans
-
Is it only me or is the minimig not going to do me any good at all .. The way i understand it "am i wrong inform me" is that i have to build the damn thing my self.. I just want too buy one no interest in working my arse off for lesser a500 compability mobo,,, The way i see it better as a amiga in a joystick prebought than a project only a hand full of us can put together .... sorry but just the way i see it ...
-
Well one option would be for us to organise getting a small volume of them put together. Either by a group of us from these and Amigaworld's boards or professionally.
I'm hoping to be proficient enough by then (I'm terrible right now but getting better!) with a soldering iron and some friends help to build one myself and I'd bet that we could put together a how-to guide for building one.
Maybe Clone-A will be better suited to purchase but I don't see this being a home-built solution as an insurmountable problem for those of us who don't (yet!) have the ability to build it at home.
As for the other issues:
I don't see compatibility as a problem, even if it is at first with a few titles, that can be fixed later.
---
"No dicking about with Kickstart ROM files." isn't an issue for me as I have A500, A500+, A600, A1200 all working with ROM's or if it becomes possible i'll use an alternative dummy ROM like it looks as though the Clone-A will have too.
---
"Warranty and support." Yes that could be attractive, but if I've built it myself what I have trouble fixing?
Plus the MiniMig(s) would be our machine, designed and built by a non-professional member of the Amiga fan community. I don't know why but after so many years of failed promises that really matters to me. I had all but given up on the Amiga until this single effort inspired me again.
Andy
-
adonay wrote:
The way i understand it "am i wrong inform me" is that i have to build the damn thing my self..
As far as I read in the previous issue of Total Amiga, the intention is to release the Minimig as some sort of hobby kit.
As a tinkerer, I find that quite agreeable.
Perhaps Clone-A would suit you better?
Fester
-
humppa wrote:
Now he is looking into some workarounds, i.e. a replacement kickstart. Using these could result in a much lower compatibility.
Kickstart ROM Replacement (Phase I) (http://thenostromo.com/teamaros2/?number=23)
Kickstart ROM Replacement (Phase II) (http://thenostromo.com/teamaros2/?number=24)
We don't need Amiga, Inc. We, as a community, need to move beyond the need for them.
-
@Dennis:
Any news? Silence just before the storm?
-
I put my hands on a Altera Nios II board: it has 8MB of 32-bit SDRAM and 1MB of 32-bit SRAM plus ethernet and CF card and a nice Stratix 1S40 device (equivalent to a Spartan-3 1600). The card is on sale for $396 - US only! (it is not ROHS compliant).
It just lack a VGA and audio output but you can add these for $175.
Now, I am dreaming about an 8MB Chip RAM Amiga clocked at 114.5 MHz :-D
-
humppa wrote:
This is still the question. AFAIK Amiga Inc. suddenly broke up negotiations with Jens Schoenfeld. Jens was trying to license the kickstart rom from them. Now he is looking into some workarounds, i.e. a replacement kickstart. Using these could result in a much lower compatibility.
GODDAMN :flame: :flame: AINC ! :destroy: :destroy:
Why can't he put a 32 pin socket into the design so the users can put their own kickrom ? Or a mmc/flash card which will contain a legal rom file user will get from amiga forever package ? That would be perfectly legal and a minor hassle for the users.
-
@countzero: Fantastic new avatar! :lol:
-
Hmmm as for me i will most liketly NOT buy clone a from jens ... I will not support that product at any cost but the minimig is a different story.. Seems i have to make it my self then .... although it will take alooooong time with my current life situation and spare time ... :lol:
-
Why can't he put a 32 pin socket into the design so the users can put their own kickrom ?
That wouldn't be too practical. Changing rom version would involve changing the chip, and nowadays, you can fit the very same amount of data (and more) in a LOT smaller space.
Or a mmc/flash card which will contain a legal rom file user will get from amiga forever package ?
afaik, that's the plan. Of course you should also be able to use a rom image extracted from your own classic amiga.
This approach also makes it easier to use totally custom rom images (that's what the non-official rom is, right?)
In any case, I find especially second solution much better than the original plan to licence ROMs from Amiga inc. - this way Amiga inc. won't be getting ANY income at all of Clone-A sales, which will likely exceed "Anything Amiga Anywhere" clearly :-)
-
@countzero
Why can't he put a 32 pin socket into the design so the users can put their own kickrom ?
If Jens target is still to achieve maximum compatibility and to target Joe Average and not only Amiga geeks, there is currently no alternative to either selling it together with AmigaForever or licensing the original ROM.
-
Ah, seems that today, Queensday in the Netherlands, the day we celebrate the birthday of our late queens' mother, brings even more hope that the Minimig will be the first (and probably only?) new Amiga-compatible hardware in a long long time. What a sillyness that law-suit between AInc and Hyperion.
-
sorry it was a 40 pin socket after all not 32 :) It's great that jens is going the softkick way.
@madsjm
thanks !
-
The weekly bump :bump:
Dennis.... any progress with the Minimig? I guess we (atleast I) would appreciate some positive news in the Amiga-scene. Or atleast some news we can all trust without serious doubt.
-- Bert
-
Dennis, what was the reasons to not allow "thay guy" to set up a small production run of Minimigs..?
What's the status of the hardware..?, I guess the "software" piece is good enough to make most users happy. Or allow them to fix things anyway.
I also notices the Xilinx 500E chips in non-bga has become available.
Belial6, There's already chassi standards to build from. Like miniATX, or nanoATX (17x17cm?). Or 3,5", 5,25" etc..
With the added benefit of an already existing market.
CannonFodder, As for the Clone-A. If I'm going to be mean, It might turn out like this :-)
100% Lawyer lapdog.
Likely only available in the shape the producer wants it.
Kickstarts are available in your own A500.. or your friends.
Warranty involves messy shipping.
And support is made up of clueless callcenters.
But most commercial projects try to prevent people from tinkering with the machine. And will dump it at will. Leaving no support.
If an commercial project would been good. Then there should been one available since long. The free model is the way to do to avoid getting stuck with legalese BS.
-
amiga_3k wrote:
The weekly bump :bump:
Dennis.... any progress with the Minimig? I guess we (atleast I) would appreciate some positive news in the Amiga-scene. Or atleast some news we can all trust without serious doubt.
-- Bert
Dennis is just building up demand! The more I wait... the more I want one :-D
-
@bloodline
Who here wouldn't want one?
-
humppa wrote:
Why can't he put a 32 pin socket into the design so the users can put their own kickrom ? Or a mmc/flash card which will contain a legal rom file user will get from amiga forever package ? That would be perfectly legal and a minor hassle for the users.
Would there be any legal issues if we developed a complete Kickstart replacement ourselves? No reverse engineering, just basing it on the existing ROM documentations. It will also allow us to bypass any possible quirks/hw bugs.
Hell, I'll help. I know parts of it both in and out.
(Jens, Dennis, you guys reading this? ;))
-
Karlos wrote:
@bloodline
Who here wouldn't want one?
Well mostly I think think of it as a really cool thing... but I have UAE... then every time this topic pops to the top, I get all excited... and start wanting some new hardware!
-
I see multiple uses for it, mainly because of its small design. For instance, it would be cool to have the machine as back-seat entertainment-system for the kids on long holidy journeys. Or as a nice add-on to the tv-set or as a nice quick-enough and silent alternative to running WinUAE on a noisy PC.
-
amiga_3k wrote:
I see multiple uses for it, mainly because of its small design. For instance, it would be cool to have the machine as back-seat entertainment-system for the kids on long holidy journeys.
That's a cool idea... though most PDA/Smart phones with a CPU greater than 400Mhz can run PocketUAE... Not tomention using a laptop... which would also allow modern games and more serious software.
Or as a nice add-on to the tv-set...
That's a nice idea... and probably how I would have one set up!
...or as a nice quick-enough and silent alternative to running WinUAE on a noisy PC.
Or a nice quiet MacBook or MacMINI ;-D
-
@skurk
Heh, you misquoted me, I never wrote that.
Hell, I'll help. I know parts of it both in and out.
I say go for it. A kickstart replacement with high compatibility would be very useful, not only for the Minimig, but also for AROS and A-Clone (see my earlier post). :-)
-
bloodline wrote:
That's a cool idea... though most PDA/Smart phones with a CPU greater than 400Mhz can run PocketUAE... Not tomention using a laptop... which would also allow modern games and more serious software.
Now, I've got three kids in the back-seat, chances are slim (well 0) that they all can have a smartphone or a laptop. On top of that, the power-consumption of the minimig is low which can come in handy with the aircon turned on. Sure, modern games apeal to some kids more than the classic Amiga ones but... I was surprised to see how much fun the kids had playing old-school 'Sonic' games on the Gamecube. Another benefit of the classic games is that you won't have to tell them how to use it every 10 minutes (try doing that while doing 140 km/h).
Or a nice quiet MacBook or MacMINI ;-D
I surely hope the minimig won't be in the same price range as the MacBook or MacMini ;-)
-
amiga_3k wrote:
bloodline wrote:
Or a nice quiet MacBook or MacMINI ;-D
I surely hope the minimig won't be in the same price range as the MacBook or MacMini ;-)
:lol: Though I fear that it wouldn't be far off a lowspec MacMINI!!!
-
Hi all,
Just to let you know, I'm aiming for a late May / early June release of the Minimig sources. I have been busy building a tube (or valve :-) ) amplifier for my brother. It is almost finished so I will finally have some time to figure out how Sourceforge exactly works. I have joined Sourceforge but the readme on sourceforge is quite big... :-o
Hang on just a little longer :-)
Dennis
-
Im glad to hear that there are still people out there (Such as yourself) who can work with FPGA's, single-handley design and build their own amiga, and still have the knowledge and intrest to work with valves...
:-D
As far as I have got with valves is to make a HT rectifier into a magnetism sensor... :idea:
Hodgkinson.
-
Dennis,
I am looking foward to the release of MiniMig !
In addition to the FPGA verilog source files, are you also going to place the schematic and layout source files on Sourceforge as well ? Keep up the good work.
:-)
-
@Dennis:
This is just the kind of announcement I was hoping since all the hectic of last week ;-).
And tube-amplifier... now that would be cool to incorporate in some Minimig-for-audio-freaks setup :-D.
-
Dennis! Very good news. Keep up the good work!
-
@Dennis
Great!
Is it a guitar amp - Purple Haze man!
I hope I don't blow a valve waiting! :-)
-
Would there be any legal issues if we developed a complete Kickstart replacement ourselves? No reverse engineering, just basing it on the existing ROM documentations. It will also allow us to bypass any possible quirks/hw bugs.
I dont think there are legal issues. There is even a AROS Bounty open for this to happen. If you could write a Kickstart Replacement maybe have a look at it, or ask Jens as he only seems to have a "workaround" that can boot up games - starting Workbench not possible.
-
Donar wrote:
Would there be any legal issues if we developed a complete Kickstart replacement ourselves? No reverse engineering, just basing it on the existing ROM documentations. It will also allow us to bypass any possible quirks/hw bugs.
I dont think there are legal issues. There is even a AROS Bounty open for this to happen. If you could write a Kickstart Replacement maybe have a look at it, or ask Jens as he only seems to have a "workaround" that can boot up games - starting Workbench not possible.
Not only is there a bounty for a Kickstart replacement, but some interesting things are in progress...
-
Is it a guitar amp - Purple Haze man!
It is an integrated stereo hifi amp.
It has 2 6550 valves per side in ultralinear push-pull arrangement doing 50 watts per channel :-D .
This amp will surely impress the neighbours :lol:
Dennis
-
amiga_3k wrote:
and tube-amplifier... now that would be cool to incorporate in some Minimig-for-audio-freaks setup :-D.
Talk about your reto-computing.
-
AWESOME! I go offline for 3 days and miss new hardware that I really care about... no not that super-amiga from ACK but the MiniMig! :-D
Good luck and thanks for keeping us informed.
Andy
-
Are there some site about MiniMig? I don't want to read the entire insane forum post.
http://en.wikipedia.org/wiki/Minimig
-
Someone need to put up a production line of the Minimig, not all of us have the setup needed to make our own "copy"
Btw, would it be possible to add maybe a 030 + fpu instead of the regular 68000?
-
Many games would not work anymore with a 68030.
-
A better question would be, can we use an FPGA implementation of the 68k? The 3Volt 68K Dennis used can't be that easy to get hold off... the 68k as a design is on borrowed time as it is... coupled with the fact that FPGAs are only going to get bigger, we might be able to reduce the chip count, by having the 68k in with the chipset!
Don't some FPGAs have an embedded CPU? Can we run a 68k emulator on one of those?
-
@bloodline
The 68k is hardly on borrowed time. You don't think that they keep manufacturing them just for laughs do you? They're used in millions of embedded devices. It's on of the reasons they're still actively developing the Coldfire series and now refer to them as "68k/Coldfire" as they've got closer to 68k compatibility with each version.
Besides which the 68k is a pretty complicated CPU and the FPGA to simulated/replicate it would eb much more expensive than simply using the plentiful and cheap embedded 3.3v version of the 68k itself.
Andy
-
How about using a standard 5V 68030 with some bus level shifters to interface it to the 3.3v logic in the rest of the Minimig?
-
AJCopland wrote:
Besides which the 68k is a pretty complicated CPU and the FPGA to simulated/replicate it would eb much more expensive than simply using the plentiful and cheap embedded 3.3v version of the 68k itself.
Ed Hepler, chip designer/engineer of CBM ("father" of both AAA and Hombre) made a Verilog version of a 68000 running at 80Mhz. According to David Haynie at least... ;)
Now he's running his own company (http://vlsi-concepts.com/)
-
Wishmaster wrote:
Many games would not work anymore with a 68030.
Ever heard of something called "Whdload" ? :)
-
And how exactly is whdload going to help loading adf images from a memory card on the minimig ??????
-
Yeah even the Pentium Pro(...and III...) made a come-back last year!
(I say this with some truth, if you've followed the whole Intel Core Duo stuff...)
-
JJ wrote:
And how exactly is whdload going to help loading adf images from a memory card on the minimig ??????
Well..............
1) its not released yet.
2) Never underestimate the resourcefulness of coders for the Amiga to put options in.
You are arguing about something that is not in the Minimig yet, but remember there is also no 68030 either, so if minimig was to receive an 030 upgrade, whats stopping the core code from accepting larger Memory cards, ADF, IPF, IDE etc.
WHDLOAD itself, does allow many games, not designed to run on a system with more than a 68000, to run on 030, 040, 060
-
I know what whdload does.
-
SKAN wrote:
Ed Hepler, chip designer/engineer of CBM ("father" of both AAA and Hombre) made a Verilog version of a 68000 running at 80Mhz. According to David Haynie at least... ;)
Now he's running his own company (http://vlsi-concepts.com/)
How about a 266Mhz one? (http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MCF5474&fsrch=1)
:-D
-
@keedon
Heh, that would be lovely but as people will be quick to point out compatibility might suffer. Depending on who replies first that'll either be a drastic or a minor problem :-D
Maybe for a MK2 or MK3 MiniMig eh.
Andy
-
Maybe for a MK2 or MK3 MiniMig eh.
Oh, i'll take one...
-
If 68030 or any other cpu will execute original 68000 instructions cycle accurate. And be electricaly compatible with the fpga it should work fine. A switch could make it run at true 7.14 Mhz speed.
The catch could be the (ab)use of undefined opcodes.
Maybe the undefined opcodes of 68000 could hint how the processor works?
Re kickstart. Once there's a working piece the barrier to fix replacements for parts of the system is way less.
-
Donar wrote:
Maybe for a MK2 or MK3 MiniMig eh.
Oh, i'll take one...
Rather like the dream device in your signature too :-D I think many of us would. MiniMig v1.0 is the important one to get our hands on first though.
Andy
-
@Dennis
Just to let you know, I'm aiming for a late May / early June release of the Minimig sources.
Great, Dennis!!!
I am looking forward to Minimig.:banana:
Keep up the excellent job!
-
Any idea on cost?
-
@keedon
We believe, and Dennis seems to have confirmed, that this will be a source only release. I.e; the plans for the pcbs, the schematics and the source code for to program the FPGA will all be made available for free.
There's been some discussion about getting a batch of them made up for those of us (like me:-D) who aren't good enough with a soldering iron to build our own. Until the actual designs have been released however guessing at the cost is a bit tricky.
Andy
-
Considering the way that the sources are being released, does the license allow a company to use them to design and build a complete Minimig that can be sold for profit? If so, I expect someone will produce a professional product based on Dennis' work, otherwise I feel we'll only see a few half-arsed DIY projects.
-
Darrin wrote:
..otherwise I feel we'll only see a few half-arsed DIY projects.
Don't forget also a few full-arsed DIY projects.
-
paul_nz wrote:
Don't forget also a few full-arsed DIY projects.
LOL. Minimigs installed in cardboard cigar boxes...
I'm sure there's a decent market for professionally manufactured Minimig consoles for those of us who don't have the time, ability or parts to assemble one from the sources, however nobody is going to make a retail-quality product for sale unless they can make a worthwhile return on the project.
-
With Dennis having done all the hard work of development my guess is that just about any electronics company could build a batch and sell it with just the right profit. Just think of the Commodore64 joystick thing.
Maybe the name Amiga should not be on the box to prevent law-suits. Hey, they can even disguise it as a CD player. All that's needed is some code to make the FPGA think it is an CD player.
-
You wouldn't need to hide it as a CD player. You would just put one freeware game on it with a freeware kickstart replacement. My suggestion would be to find, or have someone write an air combat game with tiny little Russian planes. Hence the name MiniMig.
-
Belial6 wrote:
You wouldn't need to hide it as a CD player. You would just put one freeware game on it with a freeware kickstart replacement. My suggestion would be to find, or have someone write an air combat game with tiny little Russian planes. Hence the name MiniMig.
LOL. Now there's an idea.
I'm surprised that considering the retro computer market that Amiga Inc aren't all over this thing like a rash. I mean, produced in bulk to reduce costs and sold with the Amiga logo, it should have a goood worldwide market similar to the C64 joystick and just in time for Christmas 2008.
-
actually, at this present moment in time a little combat game with funny little russian migs sounds great
does such a game exist anywhere? :-D
and I'm really looking forward to the minimig being completed, may be that final thing that pushes me into learning vhdl
-
Darrin wrote:
I'm surprised that considering the retro computer market that Amiga Inc aren't all over this thing like a rash. I mean, produced in bulk to reduce costs and sold with the Amiga logo, it should have a goood worldwide market similar to the C64 joystick and just in time for Christmas 2008.
Nah, Amiga Inc. are too busy picking their arses inbetween hissy fitting it out in a court with hyperion to actually pay attention to what goes on with the Amiga.
-
Don't forget also a few full-arsed DIY projects.
Yeah.. Just look at my website at some of the little contraptions people in the community have come up with on the DTV. I have a gallery of some of the best:
http://galaxy22.dyndns.org/dtv/common/hacks/index.html
As you can see, some are half-arsed, some are not. Several of these are my own work, and I wouldn't mind making up a few batches of Minimigs myself.
-
Darrin wrote:
I'm surprised that considering the retro computer market that Amiga Inc aren't all over this thing like a rash. I mean, produced in bulk to reduce costs and sold with the Amiga logo, it should have a goood worldwide market similar to the C64 joystick and just in time for Christmas 2008.
All they need to do is bide their time, wait for someone else to do the hard work, sweep in with a "you owe us lots of money" law suite and own whoever made it :)
-
There's a couple of really nice designs there. I love the first one and there are a couple where the DTV board is mounted inside a PS2 keybaord that look really great.
-
No dicking about with Kickstart ROM files.
The ROM:s are always available for purchase through Amiga Forever (http://www.amigaforever.com/system/).
-
Dennis, is it okay to use the picture:
http://www.amiga.org/gallery/images/3556/1_2890.jpg
As public domain ..?
-
I saw mention of the Minimig in the most recent issue of Make magazine. (#10) In the "Retrocomputing" column, there's about a paragraph about Minimig. Strangely, it describes the original version using the Spartan 3 dev board, but uses this very picture of the 1.0 board as an illustration.
-
sounds intriguing! keep us updated.
-
Hail to the King! :bow:
How goes with tube amp? :) It's nice to hear taht digitall genius like You likes to work with tubes :) Maybe an output filter for Paula on tubes? :)
Can I ask what source matherials have You used about Amiga hardware? I'm starting project to bulid my own Turbo card for Amiga 1200. Maybe even I will try to put Motorola core into FPGA but I'm not sure for now. My question is where to search the descriptions of Amiga hardware?
And my most serius question. What was the style of writing it in Verilog? Is it something like this?
module Amiga (Joy_port,Video,sound.....)
begin
your magic
end
module Paula (pin1,pin2,pin3.....)
begin
Your magic
end
and ebery custom chip in separated module? It would be awsome, bedause then there is possible bo replace damaged chip with FPGA in real Amiga.
About this bug: I think tahat it doesn't work because Your custom chip is without a bug (which bug is necessary for some particular programs to run :)
Best wishes and keep going! You are creating new history of Amiga!
-
mosfet wrote:
What was the style of writing it in Verilog? Every custom chip in separated module? It would be awsome, bedause then there is possible bo replace damaged chip with FPGA in real Amiga.
Clone-A being created by http://www.ami.ga definately will have that heirarchy as they have already done exactly that, re-created the Amiga chips one by one.
Denis said that MiniMig was a "recreation" rather than a "replication" of the Amiga, not all the pins and buses which existed in the REAL Amiga have been recreated exactly as they were (because of several reasons including that they are all inside one FPGA and there is no need to).
We'll just have to wait and see exactly how he's structured it.
I'd like to see his PCB board get manufactured... even just a short run of a few thousand :-) Without a platform the code will be fun, but for the majority of people useless.
-
I agree with Alex that working with the Mini-mig Verilog code will be interesting. I am wondering if the Verilog code is using an on chip bus such as Wishbone. If there is a defined on chip bus, adding to Mini-mig would be easier. I guess that we will have to wait and see.
:-)
-
I've always dream about portable *.mod player. Yes Yes Yes.. I know I can make mod->mp3 and listen them on every mp3 player.... but that's not he point. Amiga in FPGA would be just perfect mod player :).... I hope that Denis's decision wold be to GPL the Verilog code. Then I'll design my own board and make something in kind of Amiga Palmtop.
-
I'm wondering what the legal status would be of selling a run of the Minimig Rev1.0 boards bundled with Cloanto's Amiga Forever and OS1.3 and the ROM images removed from AF and pre-installed and configured on a SD Card.
Any thoughts?
-
@Darrin:
As long as there's no mention on the Amiga Forever CD that you are only allowed and licensed to use it on Windows PC / Apple Mac machined then I guess everybody should be able to pick-up a copy of Amiga Forever and use it just the way they desire.
-
I am wondering if the Verilog code is using an on chip bus such as Wishbone. If there is a defined on chip bus, adding to Mini-mig would be easier. I guess that we will have to wait and see.
There will be no "on-chip-bus" per-say, certainly not a wishbone or anything like that. If anything I would suspect Denis has removed some of the bus structure of the original Amiga Chipset as it was no longer needed when all chips are inside a single FPGA. With the 68000 being external, adding things to the 68000 bus should be relatively easy... you'll just have to know your Zorro ;-)
I'm wondering what the legal status would be of selling a run of the Minimig Rev1.0 boards
Selling the boards unprogrammed should be fine, that's what I'd want :) Sadly I feel that the most that will happen is the schematics and layout become available.
The only thing I worry about legal wise is "Commodore-Amiga Inc. Amiga Hardware Reference Manual : Third Edition"
Published in 1991 it is a book which discusses large aspects of the OCS/ECS Amiga chipset, how they work etc. and as such it's contents are covered by 70 years of copyright.
Changing language (e.g. English to Verilog) may not be enough to prevent the copyright holder having a claim.
-
I think Jens Schoenfeld mentioned two different kickstart replacements which were enough to boot most games but not workbench.
-
amiga_3k wrote:
As long as there's no mention on the Amiga Forever CD that you are only allowed and licensed to use it on Windows PC / Apple Mac machined then I guess everybody should be able to pick-up a copy of Amiga Forever and use it just the way they desire.
The reason I'm thinking about this is because to reduce costs of a production run you'll need maximum volume. While all of us here know how to play with WinUAE or even extract our own ROM images from our real Amigas, Joe Public doesn't. So, to market a Minimig "games machine" to him, it would need to be already set up with the ROM images in the right place and a few "free" games installed (along with a OS1.3 install if the Minimig allows a virual hard drive setup from a directory on the SD card). While selling Amiga Forever shouldn't be a problem, would providing some of the data from the package on aditional media (The SD card) break the rules?
I have a feeling that it wouldn't be allowed as it would be similar to selling a VHS tape along with a DVD-R recorded off the same tape.
-
Not to spin off in a whole new legal discussion, (if you decide to though, do post a link in this forum!) but I'm going to say that would be ok. There are plenty of legit places offering VHS-2-DVD conversion. While it was more wildly popular when DVD's first came out, and Joe Public's favorite movie wasn't available on DVD, it still is out there.. So I don't see anything illegal about that. Somone better warn the AF fellas.. And heck - They might might be a little excited because people would be flocking to their CD now. I mean come on - Most of us (that I know of) don't touch AF because, well, we already have Ami's at the house, have converted anything/everything we have, we don't need the product. But as it was pointed out - If Dennis' product heads to the masses, it's going to need to be legal, and what better, easier way, right? Cuz I just don't see whoever owns the IP behind Amiga blessing off free use of the code..
I should also point out that this obviously still won't affect us, the core "who are ready to own/build/buy" one of Dennis' creations, because many of us sit here with the software in hand - well on HDD now, anxiously awaiting the day :)
-
@ Rooster.
I hope you're right. My ex-wife is a computer illiterate and gets confused using Windows Update. However, she was a demon on Lemmings (solved all 3), beat me on the Alien Breed games and was hooked on Rodland. Anyway, she's a prime customer for one of these gizmos, but if you told her that she had to install her own ROM image then you'd get a blank look. Even if you gave her a step by step guide then she'd fail in the attempt. So, for her and the thousands like her, a pre-installed package that's ready to go once you hit the power button is the only real option.
-
I think that this project is mainly adressed to manniacs with red-hot soldering iron. :) For 'normal' computer users buliding Minimig would be very tough. Manufacturing is legal up to point of spartan soldered into a board....even configuration file with Amiga inside :) could be a cause of legality problems..but if ovnership of Amiga is so foggy, who's gonna sue us? :)
-
AF is installable. That means that you have the right to make a copy for active use. You can be sure that there is no caviate that says the install must be done to 'magentic media'. What would be cool is if after the minimig get released, if someon would make a hack that would load the ROM images from an SD card into memory from the default install locations of AF. This would allow even the most novice of users to legally use the roms without having to know an more than they would to install any Windows program. Well, they would have to know that they need to install to a drive other than c: E.g. L:, R:, or whatever their SD card show up in Windows as. Of course this would be awsome for the CLONE-A project as well. Jens should build it so that the Kickstart rom files are always run from ram. Then the replacement kickstart can load by default, but if there is a kickstart.rom file in the SD slot, it would load that instead. This would allow easy upgrades of the replacement kickstart. It would allow those who already have legal kickstarts from other sources to run them. And, it would allow Amiga to make deals with the owners of Amiga titles to sell their games on SD cards, since they could license the distribution of those roms to anyone that wanted to sell 'carts' to MiniMig/Clone-A owners.
Win.Win.Win.
-
Sounds like a job for the AROS Kickstart Replacement Bounties and AROS 68K. We would then have a totally open source Amiga.
-
AF ROMs are encrypted, you can't just copy them over to SD/MMC...
-
Argo wrote:
Sounds like a job for the AROS Kickstart Replacement Bounties and AROS 68K. We would then have a totally open source Amiga.
I'm really interested in this. First and foremost to get an open-sourced ROM replacement, but also to make ROM improvements easier for the future. SetPatch is for whimps ;)
-
Start donating! There are one or two devs interested in the UAE integration bounty. That is goint to need AROS 68k and replacement ROMs from what I understand.
-
Argo wrote:
Start donating! There are one or two devs interested in the UAE integration bounty. That is goint to need AROS 68k and replacement ROMs from what I understand.
I was thinking of getting my hands dirty doing the job myself. No promises or anything, this is just a pet project for now.
I started a couple of days ago. My goal is to rewrite all essential parts based on official Amiga ROM documentations for Kickstart 1.3 and to pull this off without any reverse engineering. Time will show how it turns out...
-
Argo wrote:
Start donating! There are one or two devs interested in the UAE integration bounty. That is goint to need AROS 68k and replacement ROMs from what I understand.
At least UAE integration as it is worded in the Bounty should lead to an 68k binary compatible AROS or AOS Box inside AROS, should help a lot with a binary compatible 68k AROS...
-
That's actually what I don't care for with that bounty. The benefits I see to having a Kickstart replacemnt is so that UAE can be integrated into Aros for backward compatibility, and even better for minimig/cloneA. As much as running Aros on a real Amiga (I would count mimimig and cloneA as real Amigas) would be neat, I would be far more interested in running older software on Aros, or either Amiga replacement hardware projects.
-
Im think Minimig is a great idea and if it ever came to production i'd buy one
-
@skurk
I was thinking of getting my hands dirty doing the job myself.
Good Luck! :-)
BTY
Are we still on for a release by this Friday? :-D
-
And my most serius question. What was the style of writing it in Verilog? Is it something like this?
Yes (real code has tabs offcourse :-D )
// this is paula
//
// written by Dennis van Weeren
//
//
// 06-03-2005 -started coding
// 19-03-2005 -added interupt controller and uart
// 04-09-2005 -added blitter finished interrupt
// 19-10-2005 -removed cck (color clock enable) input
// -removed intb signal
// -added sof signal
// 23-10-2005 -added dmal signal
// -added paula part of DMACON
// 21-11-2005 -added floppy controller
// -added ADKCON/ADCONR registers
// -added local horbeam counter
// 27-11-2005 -den is now active low (_den)
// -some typo's fixed
// 11-12-2005 -disable syncword interrupt
// 13-12-2005 -enable syncword interrupt
// 27-12-2005 -cleaned up code
// 28-12-2005 -added audio module
// 03-01-2006 -added dmas to avoid interference with copper cycles
// 07-01-2006 -added dmas for disk controller
// 06-02-2006 -added user disk control input
module Paula( clk,reset,regaddress,datain,dataout,txd,rxd,sol,sof,int2,int3,int6,_ipl,dmal,dmas,user,
_step,direc,_sel,side,_motor,_track0,_change,_ready,
_den,din,dout,dclk,
left,right);
//bus interface
input clk; //bus clock
input reset; //reset
input [8:1]regaddress; //register address inputs
input [15:0]datain; //bus data in
output [15:0]dataout; //bus data out
//serial (uart)
output txd; //serial port transmitted data
input rxd; //serial port received data
//interrupts and dma
input sol; //start of video line
input sof; //start of video frame (triggers vertical blank interrupt)
input int2; //level 2 interrupt
input int3; //level 3 interrupt
input int6; //level 6 interrupt
output [2:0]_ipl; //m68k interrupt request
output dmal; //dma request (to Agnus)
output dmas; //dma special (to Agnus)
//disk control signals from cia and user
input [2:0]user; //user disk control
input _step; //step heads of disk
input direc; //step heads direction
input _sel; //disk select
input side; //upper/lower disk head
input _motor; //disk motor control
output _track0; //track zero detect
output _change; //disk has been removed from drive
output _ready; //disk is ready
//flash drive host controller interface (SPI)
input _den; //async. serial data enable
input din; //async. serial data input
output dout; //async. serial data output
input dclk; //async. serial data clock
//audio outputs
output left; //audio bitstream left
output right; //audio bitstream right
//--------------------------------------------------------------------------------------
//register names and addresses
parameter DMACON=9'h096;
parameter ADKCON=9'h09e;
parameter ADKCONR=9'h010;
//local signals
reg [4:0]dmacon; //dmacon paula bits
reg dmaen; //master dma enable
reg [14:0]adkcon; //audio and disk control register
reg [8:0]horbeam; //horizontal beamcounter
wire [15:0]uartdataout; //UART data out
wire [15:0]intdataout; //interrupt controller data out
wire [15:0]diskdataout; //disk controller data out
wire [15:0]adkconr; //ADKCONR register data out
wire diskdmal; //disk dma request
wire audiodmal; //audio dma request
wire diskdmas; //disk dma sepcial
wire audiodmas; //audio dma special
wire rbfmirror; //rbf mirror (from uart to interrupt controller)
wire rxint; //uart rx interrupt request
wire txint; //uart tx interrupt request
wire blckint; //disk block finished interrupt
wire syncint; //disk syncword match interrupt
wire [3:0]audint; //audio channels 0,1,2,3 interrupt request
wire [3:0]audpen; //audio channels 0,1,2,3 interrupt pending
wire [3:0]auden; //audio channels 0,1,2,3 dma enable
wire dsken; //disk dma enable
//--------------------------------------------------------------------------------------
//dataout multiplexer
assign dataout=uartdataout|intdataout|diskdataout|adkconr;
//dma request multiplexer
assign dmal=diskdmal|audiodmal;
//dma special multiplexer
assign dmas=diskdmas|audiodmas;
//--------------------------------------------------------------------------------------
//DMACON register write
//NOTE: this register is also present in the Agnus module,
//there DMACONR (read) is implemented
always @(posedge clk)
if(reset)
dmacon<=0;
else if(regaddress[8:1]==DMACON[8:1])
begin
if(datain[15])
{dmaen,dmacon[4:0]}<={dmaen,dmacon[4:0]}|{datain[9],datain[4:0]};
else
{dmaen,dmacon[4:0]}<={dmaen,dmacon[4:0]}&(~{datain[9],datain[4:0]});
end
//assign disk and audio dma enable bits
assign dsken=dmacon[4]&dmaen;
assign auden[3]=dmacon[3]&dmaen;
assign auden[2]=dmacon[2]&dmaen;
assign auden[1]=dmacon[1]&dmaen;
assign auden[0]=dmacon[0]&dmaen;
//--------------------------------------------------------------------------------------
//ADKCON register write
always @(posedge clk)
if(reset)
adkcon<=0;
else if(regaddress[8:1]==ADKCON[8:1])
begin
if(datain[15])
adkcon[14:0]<=adkcon[14:0]|datain[14:0];
else
adkcon[14:0]<=adkcon[14:0]&(~datain[14:0]);
end
//ADKCONR register
assign adkconr[15:0]=(regaddress[8:1]==ADKCONR[8:1])?{1'b0,adkcon[14:0]}:16'h0000;
//--------------------------------------------------------------------------------------
//Paula local beamcounter
always @(posedge clk)
if(sol)
horbeam<=0;
else
horbeam<=horbeam+1;
//--------------------------------------------------------------------------------------
//instantiate uart
uart pu1 ( .clk(clk),
.reset(reset),
.regaddress(regaddress),
.datain(datain[14:0]),
.dataout(uartdataout),
.rbfmirror(rbfmirror),
.rxint(rxint),
.txint(txint),
.rxd(rxd),
.txd(txd) );
//instantiate interrupt controller
intcontroller pi1 ( .clk(clk),
.reset(reset),
.regaddress(regaddress),
.datain(datain),
.dataout(intdataout),
.rxint(rxint),
.txint(txint),
.sof(sof),
.int2(int2),
.int3(int3),
.int6(int6),
.blckint(blckint),
.syncint(syncint),
.audint(audint),
.audpen(audpen),
.rbfmirror(rbfmirror),
._ipl(_ipl) );
//instantiate disk controller / flashdrive host interface
disk pd1 ( .clk(clk),
.reset(reset),
.enable(dsken),
.horbeam(horbeam),
.regaddress(regaddress),
.datain(datain),
.dataout(diskdataout),
.dmal(diskdmal),
.dmas(diskdmas),
.user(user),
._step(_step),
.direc(direc),
._sel(_sel),
.side(side),
._motor(_motor),
._track0(_track0),
._change(_change),
._ready(_ready),
.blckint(blckint),
.syncint(syncint),
.wordsync(adkcon[10]),
._den(_den),
.din(din),
.dout(dout),
.dclk(dclk) );
//instantiate audio controller
audio ad1 ( .clk(clk),
.reset(reset),
.horbeam(horbeam),
.regaddress(regaddress),
.datain(datain),
.dmacon(auden[3:0]),
.audint(audint[3:0]),
.audpen(audpen),
.dmal(audiodmal),
.dmas(audiodmas),
.left(left),
.right(right) );
//--------------------------------------------------------------------------------------
endmodule :-D
-
Dennis, is it okay to use the picture: http://www.amiga.org/gallery/images/3556/1_2890.jpg As public domain ..?
Yes, no problem.
I saw mention of the Minimig in the most recent issue of Make magazine. (#10) In the "Retrocomputing" column, there's about a paragraph about Minimig. Strangely, it describes the original version using the Spartan 3 dev board, but uses this very picture of the 1.0 board as an illustration.
Cool!, they asked permission to use the photo but I never saw the article itself. Can you provide a link or something?
I am wondering if the Verilog code is using an on chip bus such as Wishbone. If there is a defined on chip bus, adding to Mini-mig would be easier.
Minimig is using a simple synchronous bus. Wishboning the Minimig however is definately on my wishlist but it is not so easy due to the nature of the Amiga chipset. It should be possible though.
anything I would suspect Dennis has removed some of the bus structure of the original Amiga Chipset as it was no longer needed when all chips are inside a single FPGA.
Yes, but unfortunately my bus doesn't allow for wait states, everything must happen within one clockcycle, without pipeline delays. This makes it difficult to interface to even an SRAM. Wishbone should be much better but would also mean a redesign.
Are we still on for a release by this Friday?
Uhmm, I'm going on a 2 week vacation next monday. I have finished the tube amp I was working on and am now busy cleaning up the code, layout and schematics for a release. I hope to have it finished next weekend, but it's still a lot of work. (otherwise I will release it after my vacation)
-
Thanks for the update Dennis. I hope someone is manufacturing these in time for Christmas. :-D
-
Uhmm, I'm going on a 2 week vacation next monday. I have finished the tube amp I was working on and am now busy cleaning up the code, layout and schematics for a release. I hope to have it finished next weekend, but it's still a lot of work. (otherwise I will release it after my vacation)
Cool. I can't wait! I hope you manage to get it out.
Have a good holiday! :-)
-
Quick question before you jet, Dennis: how hard would it be for someone to put a different CPU on there? Like say an 020 or better?
And what about a different memory configuration, so it could use 72 pin SIMMS?
-
Thanks a lot for repply. I wish You nice holiday.
Why not whith 72-pin SIMS?
I think there's serious pin-numer issue. Using this kind of memory requires 'bigger' FPGA CHIP. Bigger FPGA's are shipped inside BGA cases, which require VERY expensive boards. The next thing is refreshing. When we use SRAMS we don't care for refreshing. SIMS with DRAM on them would be very tricky.
I'm not as 0.0000001 good as Denis, but I think this is the main reasons.
-
Quote:
----------------
I saw mention of the Minimig in the most recent issue of Make magazine. (#10) In the "Retrocomputing" column, there's about a paragraph about Minimig. Strangely, it describes the original version using the Spartan 3 dev board, but uses this very picture of the 1.0 board as an illustration.
----------------------
Cool!, they asked permission to use the photo but I never saw the article itself. Can you provide a link or something?
----------------------
I checked the web site, but only subscribers can view the article online. I'll scan the hardcopy and send you a link in PM.
-
Dennis, thanks for the snapshot of a small portion of the Verilog code in your post. Your explanation got right to the point about your internal bus used in the Minimig FPGA and explains a lot since everything must occur without any wait states. This explains the use of SRAM instead of SDRAM or DDR SDRAM. I agree that converting the FPGA to Wishbone should be possible and take a fair amount of work. However, if this is done at some point, we can use a bunch of the peripherals from www.opencores.com (and maybe some from Lattice Semiconductor, who also uses Wishbone) to expand Minimig with more peripherals.
I look foward to looking over the Verilog code soon.
Keep up the good work !
:-)
-
mosfet wrote:
Why not whith 72-pin SIMS?
I think there's serious pin-numer issue. Using this kind of memory requires 'bigger' FPGA CHIP.
Since the minimig has it's own cpu, I don't see why you couldn't attach extra ram directly to the CPU, without involving the FPGA at all. Just like the fast RAM on a 'real' Amiga.
-
JimS wrote:
Since the minimig has it's own cpu, I don't see why you couldn't attach extra ram directly to the CPU, without involving the FPGA at all. Just like the fast RAM on a 'real' Amiga.
I dont know much about this, but the Ram itself isnt directly hooked to the CPU directly. You need buffer circuitry.
The only part of the RAM that will be hooked to the CPU will be the Address and Data pins, but there is a hell of alot more to it than JUST that. (Alot of Logic gates).
-
B00tDisk wrote:
Quick question before you jet, Dennis: how hard would it be for someone to put a different CPU on there? Like say an 020 or better?
In theory it could be done, There are Technical specs from Motorola on how to Hook the 68882 onto the 68000/010, I am sure that you could hook 68030 etc into 68000, but what you have to deal with is would the timing of the FPGA be affected at all.
-
whiteb wrote:
Ram itself isnt directly hooked to the CPU directly. You need buffer circuitry.
I wouldnt call it buffer circuitry but yeah, you have to go from 68000 bus to DRAM bus.
The only part of the RAM that will be hooked to the CPU will be the Address and Data pins, but there is a hell of alot more to it than JUST that. (Alot of Logic gates).
I wouldnt say "a lot". Look at the existing RAM cards, they have like 2-3 PAL/GAL logic chips e.g. my A1200 RAM card (which will possibly be more complex than an A500 RAM card) uses a GAL16V8B-25 (http://www.datasheetarchive.com/datasheet.php?article=1531089) and a GAL22V10B-25QP (http://www.datasheetarchive.com/datasheet.php?article=1532839)
That is a total of 18 Macrocells!
Or you could just buy a DP84322 (http://eshop.engineering.uiowa.edu/NI/pdfs/00/50/DS005003.pdf)
Making a DRAM interface for a 68000 was the first hardware project I did at Uni way back in 1995. 3 of us making a wire-wrapped VME board with 74LS logic. I remember hunting round everywhere for an Open Drain part.
-
whiteb wrote:
I dont know much about this, but the Ram itself isnt directly hooked to the CPU directly. You need buffer circuitry.
The only part of the RAM that will be hooked to the CPU will be the Address and Data pins, but there is a hell of alot more to it than JUST that. (Alot of Logic gates).
I was speaking of the DRAM as a functional block, not as individual components. The DRAM controller would interface directly to the 68000, without the need to go thru the FPGA. Personally, if I wanted more ram, I'd just use bigger chips on the minimig.
-
I agree with Alex that a 68020 or a 68000 DRAM controller is a fairly straight foward to design. In the past, I designed one with GALs and TTL parts; and I designed another using an off the shelf DRAM controller. Using older technology parts, a 68020 DRAM controller could be done with a GAL22V10 (state machine/timing), a GAL16L8 (address decoder), and a handfull of 74F257 muxes.
However, in Minimig, the SRAM must be shared between the MC68HC000 and the FPGA which emulates the orignal Amiga chipset. The SRAMs would have to look as a pseudo multi-port memory via the FPGA to allow for 68K access and for chipset (i.e. frame buffer or other) access, etc.
:-)
-
---edit----
ffs Oli_hd read up before you post :)
-
Any people here that have the capability to make pcb and especially get the chips assembled with automated machines?
One option is to have just the pcb made, and the big chips (Xilinx + M68k) pre-assembled. And let people do the rest themselfes?
-
@Dennis
A lot of people want more I/O on the Minimig and the big problem is the number of available I/O available on the spartan FPGA.
You should use the "I/O expander" technique : add a small CPLD in your design with let's say 40 I/Os and deport all the "slow" I/Os (keyboard, serial, joystick, parallel port, floppy) to it. The I/O updates to the CPLD will be done through a high speed syncronous serial port. Your CPLD is like a big shift register. Moreover, some CPLDs are 5V tolerant...
This should free up enough pins for an IDE port, a bigger CPU or an expansion port.
Can you please release your code before your vacation ? I will be on vacation too and without the Minimig source to read, that won't be funny :-(
I just found a quite interresting stuff : finally someone has re-created a 68000 in VHDL !
This is the amazing work of Wolfgang Foerster who is re-creating an Atari ST. Check this out :
http://download.experiment-s.de/Suska/Configware/2K6B/rtl/vhdl/WF_68K00_IP/
I have compiled the design under Quartus II for an Altera Stratix 1S40. It takes more than 9,000 LE ! That barely fits into a Spartan 3S400 :-(.
The Fmax is a little bit disappointing : 30 MHz.
At least, we can have a Supra Turbo 28...
Quite a lot of stuff can be optimized in his design : the use of built-in multipliers for MULS/MULU, barrel shifter for ROR/ROL/LSR/LSL, memory for the registers...
In comparison, the company CAST has a 68000 that fits in half the footprint and runs at 70MHz.
Regards,
Frederic
-
One could upgrade the FPGA to XC3S1000 or XC3S1500 which can be programmed with the free version of ISE.
It should swallow the minimig + 68k core. And force a minimum redesign. An alternative is XC3S500E which have non-bga variant.
One could also put two FPGAs on the pcb. And do the I/O stuff mentioned before as a bonus.
Edit: "2x XC3S500E"
-
An XC3S1000 costs about 10 dollars more than an XC3S500 and a MC68SEC000FU20 together. It is also only available in BGA, requiring a redesign of the board, a much more complex and expensive board (more layers) and BGA soldering. Combined with potential problems with the 68k HDL core, this doesn't sound like a very good idea to me.
-
I thought about bga package issues that maybe with an oven reflow + surface tension and some luck.. it could be done. Or as a pre-assembled chip. Anyway I have to agree in the end it just complicate things but it would be neat to have it all as one big chip doing everything.
Maybe a 2x XC3S400 (or XC3S500E?) could be the thing?
Would allow one to put slow I/O + cpu on a second fpga.
At least the free m68k vhdl sources will make this more future safe. Should the m68k cpu become unavailable. FPGAs will likely have fallen in price at that time to allow for a more powerfull fpga.
Is there any small modifications that can be made to allow this "platform" to replicate more obsolete computers?
-
The slow I/O trick is with a CPLD not an FPGA : XC9500 for Xilinx or MAX for Altera. The configurations of these circuits are non-volatile. Moreover, they have less gates than the FPGAs.
Splitting the design into two FPGA does not help : you automatically have a bus between them that eats up a lot of I/Os.
The best you can find today is a QFP package with 208 pins => 140-150 I/Os at best. Altera has a Cyclone III EP3C25 in this package, it is bigger than a Spartan 3S1000.
The use of SDRAM helps also reducing de pin count, especially with DDR chip : you have 32-bit data per clock cycle with only 16 data lines. And it is really cheap!
Some spoke about using a full 32-bit 680x0. Well, just the address and data take 64 I/Os... Plus, the control lines, you imagine.
-
Cyclone III has actually the PQFP 240-pin package.
Maximum user I/O:s are 160, 148 or 128, and the number decreases with fpga gate count (Altera say they need more grounds with larger amount of logic.)
Some synchronous DRAM are available in 4-bit wide configuration. To get 16 bits you would utilise the pipelined burst mode of SDRAM (in 4 4-bit data cycles.)
-
FrenchShark wrote:
The slow I/O trick is with a CPLD not an FPGA : XC9500 for Xilinx or MAX for Altera. The configurations of these circuits are non-volatile. Moreover, they have less gates than the FPGAs.
An added advantage with CPLD is they have 5V mode asfair.
And their non-volatile operation may allow to them to replace the PIC mcu processor chip.
Splitting the design into two FPGA does not help : you automatically have a bus between them that eats up a lot of I/Os.
The best you can find today is a QFP package with 208 pins => 140-150 I/Os at best. Altera has a Cyclone III EP3C25 in this package, it is bigger than a Spartan 3S1000.
The bus between two FPGAs will not eat lots of I/Os because you put them into highspeed operation.
I found alteras products are nice, but they better up their free linux offerings.
The use of SDRAM helps also reducing de pin count, especially with DDR chip : you have 32-bit data per clock cycle with only 16 data lines. And it is really cheap!
Dram operation will complicate things so things will likely stay sram for some time.
Some spoke about using a full 32-bit 680x0. Well, just the address and data take 64 I/Os... Plus, the control lines, you imagine.
You can always serialise the lines and let some other chip take the I/O pin count.
-
Dram operation will complicate things so things will likely stay sram for some time.
Did you look at the SRAM price ? If you want to be above 2MB it costs a fortune! With current SDRAM technologies, you can get 16-32 MB of memory and have a Chip RAM 64 times faster than the original Amiga.
You can always serialise the lines and let some other chip take the I/O pin count.
Not really, you will add latency on the CPU bus and have at least one wait state on your access. I think the synthetized 680x0 is the way to go : with some optimization, maybe some instructions reductions (like ABCD/SBCD, MOVEP, TAS, ...), we should beat at least a 030.
We can also take a 32-bit RISC CPU and add some custom instructions to ease up the 680x0 emulation.
-
FrenchShark wrote:
Dram operation will complicate things so things will likely stay sram for some time.
Did you look at the SRAM price ? If you want to be above 2MB it costs a fortune! With current SDRAM technologies, you can get 16-32 MB of memory and have a Chip RAM 64 times faster than the original Amiga.
Well that sounds reasonable. But the A500 was limited to 16MB (24 address bits on the cpu).
I'm sure thoe that some ramdisc could be useful :-)
You can always serialise the lines and let some other chip take the I/O pin count.
Not really, you will add latency on the CPU bus and have at least one wait state on your access. I think the synthetized 680x0 is the way to go : with some optimization, maybe some instructions reductions (like ABCD/SBCD, MOVEP, TAS, ...), we should beat at least a 030.
We can also take a 32-bit RISC CPU and add some custom instructions to ease up the 680x0 emulation.
I think the speed difference will hide the reality of latency from the cpu. As theoreticly 24bit adr+ 16bit data at 7.14Mhz would result in 286 Mbps wich an fpga could handle. Besides joystick, parallel port etc.. is much slower than this.
-
Dennis! I hope you had a great vacation! :-)
Any minimig news? :-D
-
FrenchShark wrote:
Did you look at the SRAM price ? If you want to be above 2MB it costs a fortune! With current SDRAM technologies, you can get 16-32 MB of memory and have a Chip RAM 64 times faster than the original Amiga.
PSRAM is available in large densities ( 64MB, 128MB, etc) extremely cheap and is a drop in replacement for sram.
-
Hi Dennis,
ive sent you a PM about purchasing the project. I didnt have your email sorry.
please reply me when you can.
thanks
-
ive sent you a PM about purchasing the project.
Heh, heh. Yeah right.
-
retroasia wrote:
Hi Dennis,
ive sent you a PM about purchasing the project. I didnt have your email sorry.
please reply me when you can.
thanks
The Minimig is not "For Sale" as such.
When Dennis is ready (Previously mentioned June 2007) he will be releasing everything to open Source. (Board schematics, FPGA core etc).
-
I've watched this thread with interest for ever now it seems so I thought I should finally put some input into it.
Regarding cvbs/s-video output, opencores did have a few cores to do this in the past but for some reason they updated their webpages around early 2004 and it was discarded. Using the excellent archive.org you can find the simplest one here (http://web.archive.org/web/*/http://www.opencores.org/projects/lite_videocoder/). I've tried it using webpack and it takes about 50k gates as mentioned in the pdf but someone else will have to enlighten me on the external logic required to use it. I guess it would be a resistor ladder to get the levels like the minimig does for rgb or a DAC IC.
As for input i've been playing around myself with using a playstation joypad using verilog from
here (http://translate.google.com/translate?hl=en&sl=ja&u=http://office-dsan.hp.infoseek.co.jp/pspad/pspad.htm&sa=X&oi=translate&resnum=3&ct=result&prev=/search%3Fq%3Dpspad%2Bsite:.jp%26hl%3Den%26client%3Dfirefox-a%26rls%3Dorg.mozilla:en-US:official%26hs%3DLiy%26sa%3DG). I'm having some success with this one but cannot get it 100% working.
I hope everyone else tinkering with FPGAs is having some success out there!
Cheers
-
I haven't been reading this thread, but I see it pop up a lot and it's been going on forever. So, I clicked in to see the latest and found some people questioning whether or not this minimig thing is real.
Have there been any public demonstrations or anything proof-wise on this thing? Was this Dennis character active on any Amiga boards (or retro or electronics, even) anywhere before popping in here to post about minimig and, I guess, nothing else?
-
weirdami wrote:
I haven't been reading this thread, but I see it pop up a lot and it's been going on forever. So, I clicked in to see the latest and found some people questioning whether or not this minimig thing is real.
That was 18 months ago.
Have there been any public demonstrations or anything proof-wise on this thing?
Yes. Aside from videos etc. it's been demonstrated at a couple of Amiga shows. To answer your underlying question, this is all legit and real. The best part is it's all going to be opened up (some code has already been posted here) so it will be availabe to anyone who wishes to produce it.
-
As for input i've been playing around myself with using a playstation joypad using verilog from
here (http://translate.google.com/translate?hl=en&sl=ja&u=http://office-dsan.hp.infoseek.co.jp/pspad/pspad.htm&sa=X&oi=translate&resnum=3&ct=result&prev=/search%3Fq%3Dpspad%2Bsite:.jp%26hl%3Den%26client%3Dfirefox-a%26rls%3Dorg.mozilla:en-US:official%26hs%3DLiy%26sa%3DG). I'm having some success with this one but cannot get it 100% working.
Use a PIC instead, much cheaper.
http://aminet.net/package/docs/hard/appe_v32
-
@uncharted
So the questions of it's realness were old news that I shoulda checked the date on? :oops:
-
Given that Dennis has not logged onto this board since 2007/5/28, the "realness" of Minimig is, I think, an open question. If you control the demo right it would be possible to fake what's been shown so far. I hope it's real, but in the Amiga World, reality is often just a photoshopped image...
-
persia wrote:
Given that Dennis has not logged onto this board since 2007/5/28, the "realness" of Minimig is, I think, an open question. If you control the demo right it would be possible to fake what's been shown so far. I hope it's real, but in the Amiga World, reality is often just a photoshopped image...
:roll:
Stop the crap talk already. Minimig has been demoed in the Netherlands and been shown to the public (including Jens Schoenfeld) multiple times. There is no question it is for real.
And what does the fact that Dennis is on vacation and hasn't logged in for 3 weeks to do with the "realness" of Minimig?
-
@All the lazy people out there questioning the Minimig:
Minimig has been demoed a few times. One report can be found at the Commodore GG (GG stands for User Group) site, which you should be able to reach here. (http://commodore-gg.hobby.nl/feb18.htm)
Another hit in Google: http://0xaa.org/2006/ (http://0xaa.org/2006/)
YAHIG:
here (http://eab.abime.net/showthread.php?t=26524)
AYAHIG:
German link (http://www.amiga-news.de/de/news/AN-2006-02-00097-DE.html) translated by Google: Machine english (http://translate.google.com/translate?u=http%3A%2F%2Fwww.amiga-news.de%2Fde%2Fnews%2FAN-2006-02-00097-DE.html&langpair=de%7Cen&hl=en&ie=UTF8)
YAHIG:
Seal (http://www.seal-amiga.co.uk/bigbash4.html)
Well, I guess there have been enough occasions where Dennis would have been proven a fraude if he was. Give the man a brake for a few reasons:
1) The fun part is over. Fine tuning costs time and is no real fun as progress is hard to see.
2) He has other projects to attend, one of them being his house he bought during the Minimig project.
3) Putting all on the net in a way that others can work with the sources is hard work.
4) The man is enjoying his holiday.
It's still june and if I remember correctly, the time window was even a big broader expanding into july 2007.
Guess the Amiga vs Hyperion law-suit doesn't help here either, although Minimig seems to imune to this kind of law-suits.
-
Good luck to him, and, can I have one too!!!!!
-
@persia
I saw minimig for myself at Big Bash last September and can assure you that it is real and works. Dennis doesn't have any plans to sell it himself but like others have said he will open source it when the time comes.
Once he's satisfied that it is ready for release and gets the open source license sorted out I'm sure that people will start producing kits and complete boards to buy.
-
Spoke to a big factory in China visiting New Yorks international Licensing show. Thry specialize in toys and electronics. I explained how MiniMig works based on what I've read here... and they have no question that it could be manufactured. The real issues is a viable market for it. At the end of the day as long as they get paid on time they will manufacutre it. Minimum orders are like 50000 units.
Hire a toy rep to get placement in mass or specialty market stores and that pays for tooling. Negotiate with copyright holders of our favorite amiga games...
Of course we can't put the Amiga name one it....
Maybe Commodore 16 bit games!
If a Amiga games Joystick version was engineered one could use the same tooling and release an amiga compatible clone. Similar to the C-64 thingy that everyone hacks into a commodore 64 machine.
Years ago a Chinese manufacturer actually released an Amiga clone that was fashioned into a Internet appliance for the Chinese Market. So all of this is possible.
In Hong Kong the streets are riddled with shops that sell "fan Ban" (Cantonese slang for Boot-leg) Nintentdo, Sega, and Atari style TV games that are cheaply manufactured.
The open source route will take forever...
-
Crom00 wrote:
Spoke to a big factory in China visiting New Yorks international Licensing show. Thry specialize in toys and electronics. I explained how MiniMig works based on what I've read here... and they have no question that it could be manufactured. The real issues is a viable market for it. At the end of the day as long as they get paid on time they will manufacutre it. Minimum orders are like 50000 units.
The open source route will take forever...
Did you get the factory's business card ?
What was their approximate CPI (Cost Per Item) from what you described ?
Can they package it (As in make a case, to fit the minimig Board) ?
For me, Australia's close proximity to China might be better to ship 50,000 units to Australia than the USA ? (Australian Customs depending).
Another question, is if you have 50,000 units, how will you go about getting 50,000 Spartan FPGA's (if there is that much in stock somewhere).
I suppose Kickstart is in a file on the MMC (if my memory serves me well), you just need to pay someone (Not A.inc i hope) for those 50,000 licences. If you can get each unit made for $50 and sold for $100 thats a good start.
Would people be prepared to pay $100 for a plug in and play unit ? How many games can you supply with the unit for that $50.
Lots of questions.
-
retroasia wrote:
Hi Dennis,
ive sent you a PM about purchasing the project. I didnt have your email sorry.
please reply me when you can.
thanks
Who are you now???
I think offers like this are not fitting here. This project ist supposed to be open source. Seems like you try to make us belive, you have a big asian computer company or sth.
Or are from a company that starts with "A" and ends with "a" or similar??
You bettter don't let you punk with fake offers, Dennis! ;)
-
They had price sheets for everything from action figures to plush with Minimum order quantities and the like.
I asked for a milestone sheet but was refused.
I'm not from Amiga, I work in the Toy industry and am surprised no one has done this already.
I'm not trying to cause trouble just lamenting that such a project has not been done.
I do know folks from the company that did the Jakks Pacific games (toymax before they were bought out)and they totally developed the hardware in china.
It was a lot of work but they did it.
We did not go into specifics about what hardware they would use to execute it but the key here...
It's possible, will it get done?
Probably not..
PM me and I will send the factory contact info.
-
Right... buy the design lol. We can see how well the A1 and C64TV joystick turned out...dead after just a year or two.
OS2, and a host of other OS's which never had the demand to sustain the cost.
Open source is the way to go. He may not make $$ directly from board sales (unless he sell kits himself) but it creates a ton of job opportunities that can carry someone further than a single batch of overpriced computers sold to hobbyists.
-
C64 DTV dead ? Check this out :
http://pouet.net/prod.php?which=30870 A very good demo, only for DTV.
-
Amiga180 wrote:
You bettter don't let you punk with fake offers, Dennis! ;)
First message here on this website and this is what you have to say?
Please translate what you are trying to say and then try introducing yourself and telling us who and where you are.
I don't think there was any malice intended in retroasia's message, just ignorance on his part about this project.
-
I imagine that any proper production run for the mass market would involve turning the Verilog code for the project into a VLSI chip that would be fabbed on an old cheap process (250nm or larger) - presumably with an embedded 68k core as well rather than a separate CPU. FPGAs are an expensive solution for mass market consumer products, although fine for people like us who just want to recreate the Amiga, maybe hack on new features because we can (like that 256 colour mode on the C64-DTV).
The only restriction we would have for hackability on a 'in-a-joypad' design would be if all the pins we'd require were available via solder pads on the final design, like with the C64-DTV.
The entire C64 was redone as a single chip that didn't even get it's own package with pins - it was dumped on the PCB and covered in resin. I bet the chip was under 5mm on a side.
I think the Megadrive has also had this treatment, judging from the number of 6-in-1 megadrive games in a joypad packages you can get.
EDIT: I know no-one will really care, but the C64-DTV die size is 3.785mm by 3.759mm, or 14.22mm^2.
-
I bought two of those C64 in a joystick things from QVC shopping channel. It's pretty neat. I'd love to see an Amiga 500 version done. I know I used to have an 880K disk for a 1MB Amiga 500 I had that had a TCP/IP stack and dialer and an IRC client (grapevine) on it along with a striped down version of Amiga OS 2.1 I believe. Anyway....I made a bunch of money from peecee users betting them I could log onto the Net with a 1 MB 7mhz computer with no hard drive and chat on IRC. A couple of them made me open the case to make sure it was for real. It should be possible now to do all that in one chip.
Amiga makes it possible!
:)
coldfire
-
It is possible to fool people, there was this fellow named Colin Powell who got up in front of the UN and turned pictures of absolutely nothing into weapons of mass destruction...
-
persia wrote:
It is possible to fool people, there was this fellow named Colin Powell who got up in front of the UN and turned pictures of absolutely nothing into weapons of mass destruction...
Difference being that this Powell stood in front of a bunch of, well... let's call them not so bright people. A bunch of people with a great history of meaningless wars, started only to be re-elected during the next elections. These Minimig demonstrations have been held on shows attended by people who know the Amiga, who are capable to ask all kinds of tech questions and are capable to know if the answers are correct.
So, please, do not put the Amiga comunity at the same intellectual level as the Powelss, Bushes and Britney Spearses of this world.
-
Why do suddenly so many people want to make it into a commercial product?
If that happens, all hope of upgrading the hardware (HDD controller, AGA etc.) will probably be lost.
It might not even be possible to use other games with it than the ones that are programmed into it's ROM.
Luckily, it looks like Dennis has already decided open source is the way to go.
Anders M.
-
Please keep the politics in the politics threads.
-
I think the driving force of "commercial boards" is the complicated smd circuits and the small connection paths involved.
But there's another approach to the problem. Use already in production development boards that have capacity for both A500 verilog+ m68k verilog code.
http://www.xess.com/prod035.php3
(XSA-3S1000 Prototyping Board with 1,000,000 gate Spartan-3 FPGA and 32 MBytes of SDRAM, vga, ps/2 200 USD)
Using an FPGA rather than an ASIC allows users to alter the design at will.
With ASIC any bugs will be unfixable, and AGA or other add-ons will be way harder.
-
There is an actual "retro arcade game" developers board I've seen out there. A minimig style board is soooo much cooler simply becuase of the Kisckstart and Amiga software library.
I have been asked by Toy Folk...about getting rights to the Amiga and when I give the history of the years since the demise folks shake their head and disbelief. I've been asked "Are they cursed"
I can tell you that an XBOX running AUEX amiga emulator was a hit at a toy company I worked at. Some folks had not heard of the Amiga yet we would stay after work and play Pinball dreams, Lemmings, and SpeedBall, even sword of sodan. And these guys have lots to do after work...
Goes to show you what we all know, with the propper support this Amiga thing could work on multiple levels.
Mattel, Jakks, Playmates, Hasbro would properly exploit the IP that Amiga supposedly have rights to into real products.
If apple had this IP library you'd be downloading Amiga game system to your ipod and it would be heralded as really hip, groovy retro yet still cutting edge stuff.
But what do we have? Really bad Breakout clones from Amiga Inc.... ooooffff.
Good Luck Dennis you deserve success with this item.
-
If the Minimig is real it would be perfect to build into an ipod like device. How about a Minimig phone?
-
persia wrote:
If the Minimig is real it would be perfect to build into an ipod like device. How about a Minimig phone?
No it wouldn't... it would be better to run UAE on the phone...
-
persia wrote:
If the Minimig is real it would be perfect to build into an ipod like device. How about a Minimig phone?
I kinda liked the portable minimig pictured on one of the show reports. A small lcd panel, minimig board, some Li batteries, good to go. ;-) It's too slow to play mp3s like an ipod. But Futurlec makes an MP3 decoder board. It uses an SPI interface, so you'd need to put in an SPI port in the minimig source. I wonder if the 68000 could keep up with the mp3 decoder?
-
@Crom00
There is no "Amiga software library" as you imagine it. Amiga Inc does not have any special rights to Amiga games. The only way they could have any would be by purchasing the rights.
Amiga inc, does however have the rights to Amiga Kickstart ROM, and any commercial endeavour wanting to use the KS ROM would need to negotiate a license deal with them.
Additionally deals would need to be negotiated with the rights owners of all the included games (Public Domain games being the only exception).
-
The power consumption of a Spartan3 FPGA is pretty high and really not meant for a portable applications; thus, the battery would be pretty big and expensive such as a Lithium polymer laptop battery. Whenever someday the Verilog source code is in the GPL on source forge, lots of options would open up to actually use much lower power FPGAs such as Actel ProASIC3 or Igloo. Lattice FPGA ECP/ECP2 may also be lower power consumption than Spartan3 but higher power consumption than Actel. Power vs performance trade offs can be made for future derivatives made from the original MiniMig once MiniMig 1.0 is released to open source someday ....
:-)
-
Where's Dennis? He hasn't checked in for a month!
-
I have been wondering that too..
I guess his vacation was longer than 2 weeks. Or someone actually succeded to buy the project. Or he tries to get the quality of the hw/sw to some quality goal that took longer than thought. Or he just simple got caught up in other stuff that leaves the minimig on ice.
Should the worst happen one could always take an xilinx development board and use the existing m68k core and work from there. Should lowering the entry barrier.
Then once one have a working core it's possible to put one group of coders for each chip to accomplish parallel development.
-
Dennis, when do you plan to release it? Please answer. Thank you!
-
Hi guys,
With regret, with 1265 replies, this thread is dragging the server down a bit and needs to be closed. Please feel free to create an additional "carry on" thread to continue the discussion.
Wayne