• Please review our updated Terms and Rules here

Reviving an Orchid PCturbo 286e Accelerator Card

th3bar0n said:
New memory controllers arrived - and both work perfectly, so both boards now pass the turbochk memory map test.
This weekend I will do a full install of the drivers and see if either of them initialise completely.
Click to expand...
That's great! Looking forward to the results.
 
th3bar0n said:
New memory controllers arrived - and both work perfectly, so both boards now pass the turbochk memory map test.
This weekend I will do a full install of the drivers and see if either of them initialise completely.

@Digitalman I'm thinking it might be possible to copy across a custom bios to the board that could do some basic diagnostics like a ram check - A sort of Ruud's diagnostics rom but for PCturbos. It could help to narrow down your issue.
Click to expand...
Fingers crossed great progress!
 
If it turns out the memory controller is the point of failure on Digitalman's card too, maybe I should pick up a few for my own card just in case.. I know from experience that this chip gets terribly hot, so it's expected to have a more limited lifespan than the other chips.

Also, about the PAL chips. I see that these L type can be brute-force read since there's no registers in there, so I will try that using an Arduino later this weekend.
 
A productive weekend! Got a proper setup going –
  • IBM 5160 256kb
  • IBM CGA card
  • 16-bit Multi-I/O floppy controller
  • Rancho RT1000B 8-bit SCSI card (no IRQ needed)
Did a fresh install of the Orchid drivers on a DOS 3.3 partition and set each board to Port 300h and IRQ7 (to avoid any issues with the short).

Board 2 (from @Mike1978) – With the nice new memory controller, it totally works!! Ran a few programs, particularly benchmarks and was completely stable, no invalid opcode messages. Obligatory Topbench score.. I think very close to a stock Inboard 386/PC!

topbench.png


Board 1 (my original board) – Still no go. I think the same problem as always. There is an issue with the 286 executing the bios code and it never signals back correctly to the turbo.com driver running on the host.

not_responding.png


It’s a massive step forward though, as I now have an identical working board to compare to. I’ve tidied up an earlier schematic that I did last year (attached) and I think the next step is to patch turbo.com with custom bios code that just executes some basic instructions (maybe just NOPs in a loop?) to see if I can see what is happening on the logic analyser. Although any suggestions welcome! ( @modem7 :) ?)

I think the startup is almost identical to an IBM 5170 (except the !RES signal to the 82284 comes from the TU-50 PAL, not power good), so something like the tests here (but for a 5170/286) could be a starting point for more detailed troubleshooting.
 

Attachments

  • startup.png
    startup.png
    112.2 KB · Views: 6
Awesome, nice work! So glad you got it going. I'm glad I found this thread, you've been an enormous help. I don't think I'd be as far along in my own troubleshooting with out it. I have a memory controller on route, but will be a couple of weeks yet.

For comparison, here is my setup and TopBench score: Close!
  • Commodore PC=10 II (640kb)
  • ATI Graphics CGA Compatible (w/ plantronics color plus)
  • XTIDE CF Adapter - 2 partitions C: and D: ~32mb running MS DOS 3.31
  • Adlib Compatible ISA sound card
IMG_20240121_101456.jpg
 
Digitalman said:
Awesome, nice work! So glad you got it going. I'm glad I found this thread, you've been an enormous help. I don't think I'd be as far along in my own troubleshooting with out it. I have a memory controller on route, but will be a couple of weeks yet.

For comparison, here is my setup and TopBench score: Close!
  • Commodore PC=10 II (640kb)
  • ATI Graphics CGA Compatible (w/ plantronics color plus)
  • XTIDE CF Adapter - 2 partitions C: and D: ~32mb running MS DOS 3.31
  • Adlib Compatible ISA sound card
Click to expand...

Great! I also watched your Youtube video, it certainly shows how picky these boards are.
 
Small update from my end.... I finally got around to installing a new (old) memory controller into my "bad" board. As suspected it has helped to stabilize the board but it still gives me the parity or opcode errors. I swapped RAM boards between my two cards with no success. (both boards test good on the GOOD card). Although this is only testing half the RAM, I would guess the rest of it is ok too. So what are we left with? Some faulty logic chips?

Here's the interesting thing; I can bounce around inside of DOS all day with the "bad" card. It's not until I load something. I can load "SOME" things, but usually very small programs. See the attached photos, they are all from the so called "bad" card. I am able to run TOPBENCH with a -i but not fully. These are the errors I get with everything else. Thoughts?

IMG_20240210_165933.jpgIMG_20240210_170225.jpgIMG_20240210_173353.jpg
 
Another update here.... I decided to play some "chip roulette" and start swapping out some logic ICs. Since I was getting parity errors I figured I would start with those. I replaced all the 74F280ANs, 9-bit parity IC. All I could find were 74F280BN, I assume these are compatible. I also changed the 74F74s near by. None of this has helped, in fact the only thing I've managed to do is ensure I get a parity error almost immediately now. I can still boot to DOS, but if I do a directory listing on my C: drive it throws a parity error about half way through. I can list my D: drive but it's smaller. Interesting? When I launch a program I still get OPcode error. Swapping the original ICs back produces the same result, although both 74F74s had leg snap off during extraction which I found interesting.

Clearly I am in over my head.
 
Digitalman said:
Another update here.... I decided to play some "chip roulette" and start swapping out some logic ICs. Since I was getting parity errors I figured I would start with those. I replaced all the 74F280ANs, 9-bit parity IC. All I could find were 74F280BN, I assume these are compatible. I also changed the 74F74s near by. None of this has helped, in fact the only thing I've managed to do is ensure I get a parity error almost immediately now. I can still boot to DOS, but if I do a directory listing on my C: drive it throws a parity error about half way through. I can list my D: drive but it's smaller. Interesting? When I launch a program I still get OPcode error. Swapping the original ICs back produces the same result, although both 74F74s had leg snap off during extraction which I found interesting.

Clearly I am in over my head.
Click to expand...
Quite the journey! Small steps but better to understand these things.
 
It is worth noting that the 74F series of chips are quite fast devices, and are known source of flakiness if the hardware is not designed to handle the sharp/fast transitions. If your replacement chip is too new, then it might for instance be a little bit quicker and have a little more RF overtones in the transitions.

At the point you are at, I would suggest having a go with a logic analyzer to check the parity level and data lines for every memory access. If your analyzer is wide enough, also capturing the address bus can also greatly help.
 
per said:
At the point you are at, I would suggest having a go with a logic analyzer to check the parity level and data lines for every memory access. If your analyzer is wide enough, also capturing the address bus can also greatly help.
Click to expand...
I figured, unfortunately I don't have that kind of equipment.
 
Digitalman said:
Another update here.... I decided to play some "chip roulette" and start swapping out some logic ICs. Since I was getting parity errors I figured I would start with those. I replaced all the 74F280ANs, 9-bit parity IC. All I could find were 74F280BN, I assume these are compatible. I also changed the 74F74s near by. None of this has helped, in fact the only thing I've managed to do is ensure I get a parity error almost immediately now. I can still boot to DOS, but if I do a directory listing on my C: drive it throws a parity error about half way through. I can list my D: drive but it's smaller. Interesting? When I launch a program I still get OPcode error. Swapping the original ICs back produces the same result, although both 74F74s had leg snap off during extraction which I found interesting.

Clearly I am in over my head.
Click to expand...

RL has severely limited retro time recently but I’m hoping I can get back to this a bit more now.

I assumed the parity errors were coming from a bad RAM chip, something like this. Did you try swapping the memory board with the working Orchid?
 
th3bar0n said:
RL has severely limited retro time recently but I’m hoping I can get back to this a bit more now.

I assumed the parity errors were coming from a bad RAM chip, something like this. Did you try swapping the memory board with the working Orchid?
Click to expand...
I did replace the RAM board and got the same results. But of course this is only half of the ram.
 
You must log in or register to reply here.
Back
Top