Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Protek on July 12, 2008, 06:20:06 PM
-
Having read about Dennis' MiniMig and Georg's project, as well as the NatAmi, I've really began to hope that someone with skills and resources would redesign the A1200 mainboard with a component layout that would make it possible to create a pci or even agp busboard with single slot that would fit inside regular A1200 desktop case. In practise, the accelerator connector should be in upper left corner, leaving enough space for the backside connector leads. This way, there would hopefully be enough height for custom busboard and a graphics card fitted the same way as slimline CD-ROMs. The IDE connector and hard drive would reside under the trapdoor, providing easier access. Of course, there would be possibility to soup up the hardware but in the beginning, I think that reorganising the components would be enough.
-
This new board should also use the superaga chipset.
-
And if it could end world hunger and all wars that would be good too :-)
George's project was made possible because most (all?) the chips on the A1000 motherboard were socketed. On the A1200 they are all surface mount. Harvesting the chips for a new A1200 motherboard would be prohibitively difficult.
I have a feeling that George was also a PCB designer by trade and was able to "borrow" on a lifetime of experience and access to the latest PCB tools which greatly reduced the NRE of the PCB.
One of MiniMig's great strengths (only strength?) is its form-factor. Being so small compared to any real Amiga was a great selling point. Had it been the same form factor as an A500 I doubt it would have sold in as many numbers.
I don't believe their is a willingness to develop a replacement A1200 motherboard (at the moment). This may well change if an open source AGA implementation becomes available.
Although... I was just thinking... I wonder why Jen's schoenfeld did not consider making his new A1200 scandoubler into an ultra basic RTG capable video card?? The FPGA and the RAM would almost certainly have facilitated one.
Maybe there were not enough CPU signals going to the Lisa chip to which it clips.
-
Does the Amiga 1200 and the 600 have the same layout on the back except for the mouse ports?
I was wondering if the A600 board could fit into an A1200 case.
If one board was designed to fit in both it could be a seller.
:-D
-
There seems to be no shortage of people willing and able to design new boards, we have 4 already, so why not an a1200 replacement?, we certainly would not reuse 15+ year old components, a new design would use new modern components.
-
trekiej wrote:
Does the Amiga 1200 and the 600 have the same layout on the back except for the mouse ports?
Yes (http://amibay.com/viewtopic.php?f=4&t=675&st=0&sk=t&sd=a&sid=085d987733d1208938291685efdb8770)
-
A6000 wrote:
There seems to be no shortage of people willing and able to design new boards, we have 4 already
Four?
A6000 wrote:
so why not an a1200 replacement?
Because of the availablity of the AGA chips. Or rather lack of it.
A6000 wrote:
we certainly would not reuse 15+ year old components, a new design would use new modern components.
You HAVE to use some old components. You cannot get anything to replace the Commodore AGA custom chips.
Unless you design new ones (ala MiniMig / NatAmi). Which isn't going to be easy, quick or cheap (not to mention compatible). Also, MiniMig is not AGA (or finished) and NatAmi is a closed source project :-(
-
alexh wrote:
You HAVE to use some old components. You cannot get anything to replace the Commodore custom chips. /quote]
SUPERAGA.
-
A6000 wrote:
SUPERAGA.
Doesn't exist.
-
Even if ignoring every other unrealistic bit in this thread:
Who would write the drivers for the PCI and/or AGP cards? They don't just magically work once plugged in.
-
Piru wrote:
Even if ignoring every other unrealistic bit in this thread:
Who would write the drivers for the PCI and/or AGP cards? They don't just magically work once plugged in.
LOL, we're hoping for you Piru ;)
I must admit that I also catch myself dreaming for a new A1200-like machine.
But I think that Natami and MiniMig are our two best chances so for of producing such a machine.
MiniMig:
As I understand part of it is based on the UAE code. Now I seem to have read that the FPGA doesn't have enough umph or gates to do AGA. But this could change. Maybe a rev. design could put another FPGA on it, or another clever solution might fall into place. I think that if the MiniMig get's enough time to mature, it'll have both USB, IDE and AGA. But it will take time, lots of time.
Natami:
Now, this kit will give you PCI. And a new AGA dubbed "SUPERAGA". Now alexh says that it doesn't exists. Well, it's perhaps a bit harsh. If we agree that Natami exists, then it follows that the new AGA chipset exists, because Natami doesn't use any of the Amigas original custom chips.
The hard thing for us who are impatient, is that a Natami is probably just as far away as a "Minimig v2". The difference between the two projects is monumental: Minimig started out small and has delivered, Natami has huge ambitions and is still in development. What people hold against Natami is it's ambitions and that there has only been one demo.
But back to the question of dreaming of a new A1200. I think that if you want a new A1200, side with either of these two excellent efforts. Can you code, then look into what you can contribute with. If you're a user, give encouragement and voice your support, and most importantly in our small community: Buy the products, and support the few remaining vendors.
-
arnljot wrote:
MiniMig:
As I understand part of it is based on the UAE code.
Wrong. Denis said that he used the Commodore HRM (Hardware reference manual) to create MiniMig.
arnljot wrote:
Now I seem to have read that the FPGA doesn't have enough umph or gates to do AGA.
Partly right. It's not so much not having enough capacity. More that it doesn't have a 68020 CPU or enough RAM.
arnljot wrote:
if the MiniMig get's enough time to mature, it'll have USB
I hope not. Native USB and Amiga are not a good combination.
arnljot wrote:
"SUPERAGA". Now alexh says that it doesn't exists. Well, it's perhaps a bit harsh.
Why? It doesn't exist. You cannot buy it anywhere. The source code is not available. For all intents and purposes it doesn't exist.
arnljot wrote:
If we agree that Natami exists
But we don't, because it doesn't. (Yet)
arnljot wrote:
then it follows that the new AGA chipset exists, because Natami doesn't use any of the Amiga's original custom chips.
But even if Natami did exist and was for sale it today, it would come with it's own PCB!
If you are asking if NatAmi could be made to fit into an A1200 case, that is practical and possible sometime in the future, if NatAmi is released and the designers are willing. It's certainly not something you can start planning today.
arnljot wrote:
What people hold against Natami is it's ambitions and that there has only been one demo.
And I don't think it was a demo of AGA? It was the "Super" extensions which the author appears to have created for a previous (paid) project.
I wish the creator of NatAmi all the luck in the world. It is obvious he is an Amiga enthusiast who is also a knowledgeable software/hardware engineer who, having done a 3D accelerator project thought... wouldn't it be cool if the Amiga had these features and set out to create an Amiga compatible chipset with extensions.
But at the moment the website contains aspirations which are embarrassingly touching on fantasy. The technical details are too low, they contradict themselves and seem to negate compatibility with classic Amiga. I am almost sure a lot of the website is not the work / words of the developer but the people who are surrounding / helping him with the website.
But if he's a real hobby developer, he wont give a crap what other people think. He's doing this for himself, for fun, screw what anyone else says or does! He's doing the bits that are fun, skipping the bits which are boring. I think it's cool.
-
Natami:
Now, this kit will give you PCI.
But it what it won't give you is the drivers. Having PCI slots you can't plug any cards into is quite pointless.
Regarding Natami, it does sound awful lot like the author has a bad case of "featurutus", that is piling more and more features that'll always further complicate the project and make it less and less likely to ever get finished.
Such stories of creeping featurism (http://www.catb.org/jargon/html/C/creeping-featurism.html) are not unheard of in the amiga world. The story of BoXer comes to mind.
-
@Piru
Yes, scope creep. I can´t believe that it hasn´t dawned on me before. But with the 070 and all. It seems like they are/he is engeneering towards a moving target.
@alexh
I more or less agree with all of what you wrote, even the part of the Natami "not existing", because I can´t buy one. To my knowledge there existed boxer boards, but they were perhaps not working? But even if it ever did, it does not exist. Much like Clone-A...
But,
alexh wrote:
I hope not. Native USB and Amiga are not a good combination.
As a person who is not a hardware engineer, and neither knows the AmigaOS insides or APIs, but as a user I must disagree on this one. Just a mouse and keyboard alone almost justifies this one to me, and then add mass storage and it´s perfectly clear to me that this is a very important feature.
Yes, one can have PS2. And yes it´s probably cheaper to implement. But isn´t the point to always move forwards ;)
There should however be a prioritized list of features, so that it´s constantly evolving. Features don´t need to incubate for years before they appear.
slightly off topic:
It seems that the 68040 are abundant, in speeds 25mhz, 33mhz and 40mhz. Would it be a suitable choice for a "power minimig"? Which one of the 68 series is most common, except for the 68000?
-
Part of the issue with USB are that you need a CPU side USB stack/driver to control the peripherals that you plug into it and that it actually takes a bit of processing power to do run that. There are already USB stacks for the Amiga though so it must be workable.
For building a "power minimig" or "MiniMigV2" etc the 68060 makes more sense from a simplicity point of view. This is because the 060 is a 3.3v part whilst the 020, 030 and 040s are all 5v parts. The rest of the MiniMig (FPGA, RAM et al) is all 3.3v so attaching them will require voltage level shifting and different power planes for the two voltage regions on the board which makes it all more complicated and a lengthier process to build.
Of course the 68060 is harder to get hold of and more expensive...
Andy
-
arnljot wrote:
as a user I must disagree on this one. Just a mouse and keyboard alone almost justifies this one to me
But they dont work!!
Workbench yes... Games No. (Well mainly no)
You see the games hit the hardware registers of the custom chips. They do not go via the OS and so (AFAIK) do not work with USB keyboard and mice.
Most USB mice and keyboards support legacy PS/2 over USB so with an adapter you can use them directly with MiniMig via the PS/2 ports which are wired to the hardware of the custom chips.
arnljot wrote:
and then add mass storage and it´s perfectly clear to me that this is a very important feature.
You get mass storage via IDE, replicate the GAYLE registers and you have a 100% compatible interface AND drivers.
arnljot wrote:
slightly off topic:
It seems that the 68040 are abundant, in speeds 25mhz, 33mhz and 40mhz. Would it be a suitable choice for a "power minimig"? Which one of the 68 series is most common, except for the 68000?
It has to be 3.3v I/O to interface to the FPGA. This limits which 68k you can use. You cannot use any 040 from the Amiga days. Not without extra electronics to convert the I/O levels. You cannot even use a regular 68000 out of an A500! (MiniMig uses a special 3.3v I/O 68000)
As I have said before the 68060 is NOT 3.3v I/O (unless the specs at Freescale are wrong). It is 3.3v core, 5v I/O.
-
@alexh
Yup, I know about PS2 and USB and that most peripherals include chips that operate on both.
And perhaps it's over engineering to map usb->ps2->amiga inside the minimig. And also allow a usb stack to access the usb ports... But:
Yes, it takes CPU to run a USB stack, but it can be done on the Amiga1000, and on OS classic with Poseidon. So it should be doable on the Minimig, but certainly hardly usable with a 7mhz 68000.
And yes, IDE will give us mass storage. But how cool would it not be to be able to boot of a memstick which you also use at work with WinUAE? Hmm...
Any way, if the SD card could be used as a IDE drive, I'd be happy as a clam :)
When it comes to form factor and the A1200 plastics, I'm sure a 17x17cm Minimig will fit :)
-
alexh wrote:
As I have said before the 68060 is NOT 3.3v I/O (unless the specs at Freescale are wrong). It is 3.3v core, 5v I/O.
Ah bugger, I must have missed that, sorry.
Andy
-
I'm unsure... some of the Freescale documentation says 5v I/O in some places and 3.3v I/O in other places.
I'm wondering if it does both?
-
arnljot wrote:
But how cool would it not be to be able to boot of a memstick which you also use at work with WinUAE? Hmm...
I already do it now using A4000 and a CF card (which is IDE).
-
alexh wrote:
I'm unsure... some of the Freescale documentation says 5v I/O in some places and 3.3v I/O in other places.
I'm wondering if it does both?
If the 060 doesn't drive the I/O at 5v, the 68k support devices that it needs to interface with wouldn't work!?
-
Not if it is just digital. You should look up TTL thresholds. 3.3v is still a HIGH.
-
alexh wrote:
Not if it is just digital. You should look up TTL thresholds. 3.3v is still a HIGH.
So the the 060 could happily run at 3.3v and simply tolerate 5v inputs...
-
bloodline wrote:
So the the 060 could happily run at 3.3v and simply tolerate 5v inputs...
Maybe, I think it is more complicated than that. It probably requires voltages on certain pins at certain times...
-
I can't believe that this has hit two pages, and no-one has mentioned ColdFire yet :-D
-
It's like the amiga communities variant of Godwin's Law (http://en.wikipedia.org/wiki/Godwin%27s_law) :-o
Andy
EDIT: damned non-html links blah blah blah .. nazis
-
I suppose an obvious question would be how do the Blizzard card handle it? You can swap out the 040 for an 060 and you change the voltage regulator. Does that regulator change the voltage going just to the core or to the IO as well via level changers?
Andy
-
Maybe Dimlow of EAB (http://eab.abime.net/member.php?u=11736) will donate his Apollo1240/60 regulator design to the Minimig? (http://eab.abime.net/showpost.php?p=428730&postcount=80)
Could it be done with out too much redesign to theMinimig hardware and firmware?
-
Eh?? That is just a remake of a standard voltage regulator board for an Apollo card. An ultra simple board which has been on sale for years, that PG has been selling to people on this board for years.
All the "good bits" (if there are any) are actually in the 1240/1260 board design itself, not this addon.
The MiniMig PCB would require a full re-design to use an 060 CPU. MiniMig is a two layer PCB which would almost certainly not be possible for an 060 design due to the number of power and ground pins there are compared to a 68000
Not to mention that MiniMig is a 16-bit design and the 68000 is a 32-bit processor, there are not enough pins on the FPGA that MiniMig uses for a 32-bit processor so you'd have to change to a different FPGA too.
And 1.5Mbytes of RAM is not enough ;-)
-
@alexh
Just goes to show how little I understand about hardware :D
MiniMig is twolayer, so much I've caught on. But are you saying that BLizzard, Cyberstorm and Apollo 060 cards are singlelayer, or much more (4++)?
What would these "good bits" of such a card be? The basic good bits that is, baring the interfacing with the Amiga side.
-
arnljot wrote:
@alexh
are you saying that BLizzard, Cyberstorm and Apollo 060 cards are much more (4++)?
Yes.
arnljot wrote:
What would these "good bits" of such a card be?
The power & ground routing mainly, handling all the extra pins, perhaps some electronics to support dual voltages.
-
So a fair recap then for an Amiga1200-like MiniMig:
1) It must have more memory.
alexh: Is it true that a challenge with with the current design is that they must reimplement a memory controller for newer more common memory?
2) It must have a better processor than a 68000, the 68060 is best because it's 3.3v just like the MiniMig.
alexh: Are there FPGAs around that can handle a 68020 if Tobiflex's FPGA 68000 gets reimplemented as a 020?
3) Some rerouting needs to be done. Either to make it ITX (industry standard, easy to get a case) or A1200 like so that it goes right into our old boxes :)
4) Mass storage needs to be implemented. IDE is on the wishlist for many people.
This could end up then with an AGA A1200-like machine with either an expensive 060 or an FPGA implemented 020. PS2 ports for mouse and keyboard.
alexh: If this is done, do you think that the board would need two FPGAs? If so, what would that do to the cost of the board? If we would wait for the semiconductor industry, is it likely that we in the next year would get a cheap FPGA that hold AGA+020?
Many questions, but I'm trying to get a picture of this so that my expectations to who ever is working on the MiniMig is reasonable ;)
Both in terms of costs (to end user), features and time scale for advances :)
Of course I know that you're not working on it, but you certainly understands these issues better than me.
-
On the other hand, it might make sense, if Georg's PCB design was adapted to A500. It's true that even those components are hard to come by but the design exists.
Having looked at an A1200 mobo (I know, I could've done that earlier), it doesn't help that much in fitting a PCI bus. As Piru stated, drivers would also be needed, unless existing busboard drivers can be used.
I suppose there's no hope of better WB graphics for us desktop A1200 users.
-
arnljot wrote:
alexh: Is it true that a challenge with with the current design is that they must reimplement a memory controller for newer more common memory?
Yes, but it has already been implemented by Tobiflex for the DE1 / DE2 boards.
arnljot wrote:
alexh: Are there FPGAs around that can handle a 68020 if Tobiflex's FPGA 68000 gets reimplemented as a 020?
Sure, but how much do you want to pay? And I dont see an 020 coming out any time soon. The 68000 isn't perfected yet.
arnljot wrote:
4) Mass storage needs to be implemented. IDE is on the wishlist for many people.
For MiniMig, you wouldnt want true IDE. Just re-route access to the MMC card.
arnljot wrote:
This could end up then with an AGA A1200-like machine with either an expensive 060 or an FPGA implemented 020. PS2 ports for mouse and keyboard.
Where did you get your AGA core from??
arnljot wrote:
alexh: If this is done, do you think that the board would need two FPGAs?
Depends on economics. Probably one big one.
arnljot wrote:
If so, what would that do to the cost of the board?
Too many unknowns to even speculate.
arnljot wrote:
If we would wait for the semiconductor industry, is it likely that we in the next year would get a cheap FPGA that hold AGA+020?
Of course. But what is this "waiting" you talk about? It's not like you have anything to put into an FPGA if you had one in front of you.
arnljot wrote:
Many questions, but I'm trying to get a picture of this so that my expectations to who ever is working on the MiniMig is reasonable ;)
AFAIK there is no-one really working on MiniMig, there are about 3-4 people who play with it in their spare time, as a hobby, nothing more.
-
Get an Efike motherboard and keyrah - make them fit into A600/A1200 case - install MOS2.0 - done. :-D
Would be nice if there was Efika with only pinhead connectors instead of "tall" sockets, like certain variants of the nano-ITX cards, would make it much easier to install into odd cases, and also much more suited for embedded use. Now, how was the story for Efika2 again...?