Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Generale on March 13, 2005, 11:20:26 AM
-
Hey all!
Just finished building and debugging (lots of debugging) the a500ide interface. Works beautifully. Even with the proto board rats nest!
I have an IBM 540mb connected and it works fine. As long as I don't use the IDEMount program. It sits there for several minutes before declaring that it doesn't know.
My main problem now is making a half decent boot disk to get the thing up and running without too much extra crud. It's for WB1.3 by the way.
Can anyone suggest to me nice simple setup to cleanly boot it up? In all honesty I'm not sure about what I can get away with pruning.
Thanks!
-
I can´t help you with your problem but I do want to congratulate you for what you have done so far !!! :-D
-
Thankyou very much :)
I spent hours trying to find out why it was reading garbage. Found a few minor mistakes, but it wasn't the problem. Turned out the problem was a little bit of superglue jamming down one of the contacts for the connector I made. Hours of probing with a multimeter and tracing wires for nothing!
...At least it has a pretty blue LED
I got it to boot from a floppy, but it needs work. Currently dumping an ADF so I can't cause too much damage.
Next projects are the kickflash and the AT Keyboard, then the case mod.
Anyway, the drive is brilliant! So quick. Now to put my favourite apps on it. No more disk swapping!!!!! I'm so happy!!
I'd definitely recommend this project if you want a cheap solution and don't mind using a boot floppy.
-
Why use a boot floppy. I hear your quite a skilled hardware builder. Then it should be no problem for you to make your own ROM that has the drivers gor your IDE drive. I thought the A500 has 256k Rom and mirrors its ROM in the 256k after the rom space. So it should be possible to create somthing that takes advantage of that.
Or an A600 3.1 Rom. I hear this work on the A500 too. You might be able to replace SCSI.device from that one.
Anyway, good work and congratulations.
-
I've been considering putting the drivers on ROM. The only problem is I don't really know how.
If you know of any docs on how to do this I'd really appreciate it.
I was going to take the A500 up to WB 2 or 3, which pretty much cancels out the whole 256k rom advantage. But I noticed there's still a little space on the end of the KS roms. Maybe not enough though. Anyway, there's nothing saying that I can't shove in a ROM elsewhere in memory and just have a jump to it on the KS ROM.
Thanks for the ideas :)
-
Hi Generale,
I was just wondering what set of plans you built your interface from. I might want to give that a try myself.
Thank you,
AmigaEd
-
Sorry I have no docs. But at work we do use Flash devices and Eeprom devices. As for kickstarts; ther is always reverse engineering. Copying a Rom is not difficult if you have a compatible device to put it in. Finding the devices you want to replace is another. A collegue of mine claims he made a custom Rom once. So I shall ask him next week when I see him again.
-
Bleh... dont beleive ya Generale :-P as an Aussie sceptic I need pics of the working board and drive
Well done btw!
-
@AmigaEd
The design I used was at http://wuarchive.wustl.edu/aminet/hard/hack/a500ide.lha (http://wuarchive.wustl.edu/aminet/hard/hack/a500ide.lha)
Alternatively you could go to the site that the maker of this hack has the design and drivers too. Thats at http://www.students.tut.fi/%7Eleinone3/ide/index.html (http://www.students.tut.fi/%7Eleinone3/ide/index.html)
If you do build it, you don't need the 7404. I have no idea why that's on the schematic. And the wiring for the activity LED is only on the PCB, not the schematic, so if you want LED, you'll have to look at the PCB.
@doppie
I was going yo use redskulls design for the kick flasher. it's at http://wuarchive.wustl.edu/aminet/hard/hack/DC-KF500.lha (http://wuarchive.wustl.edu/aminet/hard/hack/DC-KF500.lha)
As an odd coincidence, I have the type of flash ROM the design specifies. Bought them surplus from unisys.
One thing to be wary of. The driver seems to have a massive footprint. I'll have to look into it.
My a500 reports 886816 free with a standard wb1.3 disk and 652840 free with the driver. Kind of irritationg because it hasn't left me with enough for HAM in dp-iv.
-
It sounds to me like your next stop will be here (http://main.aminet.net/hard/hack/8meg.lha) although you won't be able to exceed AmigaOS 3.1 with that 68000 processor... even if you do boost it with this (http://main.aminet.net/hard/hack/14MhzA500.lha).
-
Here's some pics.
If you want to build it you'd be best off ordering parts from Futurlec (http://www.futurlec.com.au) because its impossible to find anything in Melbourne.
Drive Details (http://img.photobucket.com/albums/v291/princess_of_pain/a500-3.jpg)
Drive nearly empty (http://img.photobucket.com/albums/v291/princess_of_pain/a500-2.jpg)
card sticking out of the A500 (http://img.photobucket.com/albums/v291/princess_of_pain/a500-1.jpg)
I couldn't get a pic of it with the drive LED on because it's too quick :-)
-
SamuraiCrow wrote:
It sounds to me like your next stop will be here (http://main.aminet.net/hard/hack/8meg.lha) although you won't be able to exceed AmigaOS 3.1 with that 68000 processor... even if you do boost it with this (http://main.aminet.net/hard/hack/14MhzA500.lha).
Would you believe I was just looking at that hack about a minute before I saw your link to it?
Been considering the 14Mhz one. Except I doubt I'd be able to get the CPU, So I was thinking 'I wonder if there's any reason I can't use the original if I can dissipate the excess energy'
It may throw the 8mb hack into chaos anyway because of the delay method.
Still, I did have an idea which I may try later on. It involves using latched buffers as an interface between a ram board and the puter. And running the ram at a multiple of the clock speed, so in the times that the bus is not ready for sending/recieving data the ram can be refreshing and/or accessing the adresses specified and readying the data. So it'd kind of be creating pseudo 1T ram.
I have heaps of dead ibm clone mobos to take simm slots off anyway (even though i think dontronics sells them) and a box of SIMMs, so yeah. Only prob is I'd have to make another board....and another connector (oh crud), with both things built in. erk!
-
Perhaps it would be a good idea to contact Jens Schoenfeld at nospam[at]jschoenfeld.de...
-
Dandy wrote:
Perhaps it would be a good idea to contact Jens Schoenfeld at nospam[at]jschoenfeld.de...
*eyes glazed*...I know that name. Why do I know that name? I don't remember. help.
-
Didn't he design the Catweasel floppy controller?
-
That could very well be it. It was bothering me not knowing where I knew the name.
If anyone has tried the project and had different RAM use stats, could you please tell me?
In the meantime I may have to doctor a RAM expansion. New thread time.
-
Well done, it a good feeling isn't it when that mess of wires you have just soldered together actually works.
I've got pictures of my second prototype in the images section that plugs into the CPU socket.
BTW, how long is your IDE cable that you are using?
I found only the shortest ones work reliably especially with a CD-ROM as a slave as well.
-
the distance from the controller to the hdd is 10-15 cm. If you want extra distance, an 80 conductor cable may help. After all, they are designed to sink crosstalk.
I'm just using a 40conductor cable though. Standard IDE length. It's trailing end is currently draped over the hard drive and near the PC monitor, so i'm surprised it works.
I was wondering if using the CPU socket would work. Guess so!
by the way.... What's the rough memory eaten by your drive? like I said earlier its eaten about 250k of ram. I was just wondering if its only me thats having that problem.
It is a good feeling seeing something work though. I thought I'd never get it running.
-
Man, Generale, as sharp as you are I'm surprised you haven't started on an A500 CPU upgrade project.
-
Heh. I've been tempted. but the prob is that CPUs cost $. And the reasons I build instead of buy are twofold.
1: I don't pay for anything that I can do myself.
2: It's in my price range.
I did come up with an idea on how to 'upgrade' a 68k today which I actually came up with ages ago to trick up a 6502 (but never tried). If I say what it is I'll probably be burned at the stake as a heretic.
Alternatively.....I can't remember how the interrupt handling works on a 68k. If it's anything like MIPS RISC I can cheat. It's amazing what a person can do with an interrupt handler and some overclocking.
I want to try a CD-ROM on my Ami. But I can't :-( Apart from power drain concerns, is there any support available for ISO9660 in wb1.3 anyway?
-
The Xetec driver on l8r.net supports KS1.3, so that should work for the CD-ROM. I used KS/WB2.x so the choice of CD filesystems was greater though I can't remember which one I actually ended up using. AsimCDFS I think.
Memory useage by the driver was a little under 250k I think, I will plug it in and check it out time permitting.
Another way to overcome the cable issue is to add some buffers to the interface which isn't difficult but with short lengths I didn't deem it necessary.
I made a startup disk that loaded the driver and passed the remainder of the startup-sequence and assigns to the harddrive to speed up booting though I too did consider a modified ROM based driver to autoboot from the HDD but I never got very far with that idea due to lack of time.
I may still look into it though.
I also built the 14mhz hack and it works fine with 68000's from later (rev 6+) A500's. I ended up using a socketed square 68000 from an A600 for my project wired into a standard DIL 68000 CPU socket.
Regards,
Clockmeister.
-
Generale wrote:
Heh. I've been tempted. but the prob is that CPUs cost $. And the reasons I build instead of buy are twofold.
1: I don't pay for anything that I can do myself.
2: It's in my price range.
Well how much does a bare 060 or hell even a low end G3 CPU cost? I'd be willing to donate to that project.
-
@generale
Ralph Babel has a KS1.3 Cd-rom driver here
http://babel.org/amiga/#omni
connect everything up, see if it works :-)
-
Laugh at me if my question is stupid, but would a CDTV KS1.3 ROM fit in an A500 rom socket, and more importantly would it be of any use?
-
The CDTV 1.3 ROM is identical to the A500 one. The CDTV has a pair of extra ROMs that patch themselves into the system at boot time.
-
hm. So many bits and pieces to reply to. I'm guessing most people here are either vampires or in a different timezone.
Okay. First yeah, buffers work wonders. Especially a nice high speed CMOS buffer. Preferably a schmitt triggered. It'll have those signals nice and crisp. Which is needed moreso if its under the cpu. I suggested the 80 connector cable because its cheaper and simpler.as for better, probably not.
Thanks for the URLs for the drivers. I'll definitely have to try them out. I know I can play audio CDs atm, but it hasn't been enough to justify connecting a cd-rom. I do know the first audio cd and song I'll test it with though. Can you guess what they are?
So it wasnt just me the driver was big for. Hm. Well, an ide interface should only be a few hundred bytes or so, so I'm guessing the rest is interrupt handling and OS interfacing.
The hardest part of interfacing a faster CPU to an a500 as near as I can tell would be the core and IO voltages for the CPU. As for the PPC I have had _no_ experience with that apart from using the computers at uni.
My idea for 'upgrading' a 68k is much more evil and still uses the original 68k, roms, a high speed microcontroller and some tristatable buffers. If you want to know more, say so and I'll spawn another thread.
I considered the CDTV rom too for its fs drivers. Never got further than that.
Extra ROMs? Wow, they really did cheat.
Oh yeah. Like I said, rusty on 68k theory of operation. Been ages and that was hust for altering features on apps for my palm that irritated me. So, it does have an interrupt for illegal opcodes, right? Has anyone made a small patch already that takes advantage of that yet...or does the OS fill in the blanks for the missing opcodes on the more lowly CPUs?
-
http://www.students.tut.fi/~leinone3/ide/
did you get the latest driver from this site?
also the lucas/frances accelerator project might be of interest to you
http://ftp.funet.fi/pub/amiga/hardware/accelerators/
-
Those are the drivers/schematic I'm using. Thanks for that though.
I'll have a look at the accelerator project in a min.
Oh, would this be right for the mountlist entry?
CD0: Handler = L:BABELCDROMFS
Startup = ide.device/2
StackSize = 4000
Priority = 10
GlobVec = -1
The 'Startup' line bothers me. Would it be /1 or /2 (slave on controller)..More to the point, is this a standard syntax used to specify the unit number?
-
My idea for 'upgrading' a 68k is much more evil and still uses the original 68k, roms, a high speed microcontroller and some tristatable buffers. If you want to know more, say so and I'll spawn another thread.
Yeah, I'm interested... Spawn please :-)
-
@Generale:
I have the Schematics for the Lucas board in Gif format if you need them.
They are available here: http://www.ianstedman.co.uk/Downloads/68020-68881.gif (http://www.ianstedman.co.uk/Downloads/68020-68881.gif)
http://www.ianstedman.co.uk/Downloads/Pals_&_Glue.gif (http://www.ianstedman.co.uk/Downloads/Pals_&_Glue.gif)
You would need to replace the obsolete 16L8 GAL devices, luckily the PLD source code is included.
-
mdma wrote:
would a CDTV KS1.3 ROM fit in an A500 rom socket, and more importantly would it be of any use?
Not any use, cos the CDROM in a CDTV isn`t IDE, so the driver wouldn`t work.
-
Doobrey wrote:
mdma wrote:
would a CDTV KS1.3 ROM fit in an A500 rom socket, and more importantly would it be of any use?
Not any use, cos the CDROM in a CDTV isn`t IDE, so the driver wouldn`t work.
Would the CD32 ROM be incompatible/useless with a 500 too?
-
Yes, a CD32 ROM would be useless. A CD32 is a 32-bit machine. The A500 is 16-bit.
-
Well, a 32 bit ROM can be used..I think...If it's burnt onto a 16 bit rom or a bit of addressline trickery is done, because all 68ks are 32bit internally anyway. It's just the bus is crippled (like the 8088 was a crippled version of the 8086). The only concern is whether the memorymap is the same between hardware.
Good point about the scsi driver in the CD32. Speaking of which, the scsi/atapi packet tool that came with the ide.driver looks like some fun to play with
-
B00tDisk wrote:
Well how much does a bare 060 or hell even a low end G3 CPU cost? I'd be willing to donate to that project.
New 060s are rather expensive. The budgetary price (when buying in lots of 1000 or more) is $283 for the MC68060RC50 (50MHz 060). If you're willing to settle for an EC part (no FPU and no MMU) you can for MC68EC060ZU50 for only $83, but it's a BGA part which requires some funny equipment to solder. Most PowerPC parts I've seen are also BGA though I seem to remember the G3s and possibly later Macs using a ZIF socket so perhaps there is a PGA version of some sort.
-
Generale wrote:
Well, a 32 bit ROM can be used..I think...If it's burnt onto a 16 bit rom or a bit of addressline trickery is done
For the CD32 ROM, it can be used..but it still won`t work.
1. It`s a 512x16 bit ROM, so it only needs an extra address line hacked in to access it.
2. Some of the code is probably for 020 CPU`s, so it wouldn`t run on a plain A500.
3. Although the CD32 rom has a scsi.device, it`s the cd.device that controls the CD-ROM , and just like the CDTV, it isn`t an ATAPI drive :(
-
I tried Ralph Babel's driver. Either there's something I'm overlooking (most likely), or the babelcdromfs and ide.device don't like each other. I'm not sure. I can't get it to work :-(
-
The Xetec drivers can use the ide.driver it seems. The only problem is that I cant actually use it because it complains that the version isn't official. I dumped across the files.
After that, I tried putting the ADF onto floppy. Well, It looks like 1.3 doesn't know how to cope with the format of the disk. Help!!!
-
Have you tried using IDEfix97 drivers, they were made for IDE drives. I cant ever remember using them under KS1.3 tho.
Something else for you to try :-)
http://main.aminet.net/disk/misc/IDEfix97.lha
Az
-
I'll have a look at the driver soon.
Babel almost works. It keeps tormenting me when I try to mount with a dialog telling me volume is read only. with retry and cancel as the options. So I hit cancel and then it just moans about it not being a dos disk.
Is my pressing cancel saying "Hey, I won't load the driver now?".
Okay then. Next!
How do I mount an FFS formatted floppy in wb1.3? I found an old thread here but it just doesn't work for me. I just want to give the Xetec drivers a try, and the floppy is in FFS format!
-
FastFilesystem in 1.3 only worked on harddrives. The support for floppies requires KickStart 2.04+. I admire your attempts at bringing an A500 up to date with simple hacks but you can only get so far. :-(
-
Heh. I know it'll never be remotely up to date. But theres nothing wrong with updating a little. I enjoy strangling every last bit out of everything I have. I like a challenge.
I tried IDEFix. It didn't want to play with me. Ah well. Thanks for that.
As for FFS on a floppy in 1.3. I thought it was possible, just that it involves a bit of tinkering with the mountlist.
Its actually a little frustrating. I know Xetecs drivers will work for me on 1.3, but I can't use them because of the ffs floppy issue.
I have to admit, right now FFS has a couple of meanings to me. I just keep hitting dead ends!
Thanks for the input though. Much appreciated!
-
Oooh, very excellent!
I was expecting a breadboard with untrimmed leads sticking out all over along with a real rats nest of wire!
Good job!
-
Hehe. Thanks.
It could have been tidied a little more, but given my construction error rate I was happy with the way it was, because it worked. I made sooo many mistakes. Fixed them all though.
I actually made some wires a tad too short. I had a lot of trouble plugging in the ICs as a result.
I think spending a few minutes trying to arrange the components for optimal fit before I started soldering paid off. I'm not sure but it seemed to have benefited the signal/noise ratio. I can use a standard IDE cable error free!