exxos's Atari, Amiga & retro forum

ROMsoak and TTDiag run from floppy.

Your TTDIAG and ROMSOAK results look good. TT-RAM passed everything with no faults, and the ROM copy matched the ROM chips for 13 passes with no errors.

IDETEST needs running again, as I think the last run tested the UltraSatan rather than the IDE drive on the ST536. The transfer speed was about 500KB/sec, which is what I would expect from the UltraSatan, not the ST536 IDE port.

To make sure it tests the right drive, please unplug the UltraSatan for this test, so the only hard drive connected is the IDE drive on the ST536. Run IDETEST.PRG from floppy, and when it asks which drive to test, pick a partition on the IDE drive. Please take a photo of the first screen showing the drive letter and free space, then let it run the full 5 passes and post a photo of the end result.

After that, please run BUSTEST.PRG from floppy and leave it running for at least half an hour. This is the only one of the four tests that runs with the 68030 caches on, which is closer to what GEMBench does when it crashes.

Can you also tell me which firmware version is on the board, and whether you powered the machine off before running ROMSOAK, or just pressed reset? ROMSOAK reported that TOS was running from the ROM chips rather than the copy in TT-RAM, so I need to know which it was. If you reset then shadow ROM is lost until next power on.

IDEtest failed when run on the IDE drive. It would never complete or load

Bustest - never completed but did run for a while occasionally.

coonsgm wrote: 10 Nov 2026 18:29 Bustest - never completed but did run for a while occasionally.

Thanks for running those. BUSTEST is the first test that fails on your board but passes on mine, so that's a good lead.

I have updated BUSTEST so it now runs with the caches off first, then with the caches on, and it shows the first errors on screen as soon as they happen. That way we get the details even if the machine hangs afterwards. There is also a second version, BUSTESTS.PRG, which is the same test but runs from ST RAM instead of TT-RAM.

New kit here:

Please do a cold boot (power off for 10 seconds), then run BUSTEST.PRG from floppy. If any ERROR lines appear, please take a photo of them straight away, as the machine may hang shortly after. Then do another cold boot and run BUSTESTS.PRG from floppy, and photo the result the same way.

Which firmware version is on the board at the moment? That makes a big difference to what we're looking at.

Firmware is version H

Running the new version without cache and no errors as yet.

coonsgm wrote: 10 Nov 2026 19:19 Running the new version without cache and no errors as yet.

Interesting.. photo any ERROR lines the moment they appear in the caches-on half.

@coonsgm

Something else that would help narrow this down: if you have an H4 or H5 board, could you try your ST536 in that as well ?

My test machine is an H4, which has pull-up resistors on the ST data bus, and your 520ST and Mega most likely don't. If BUSTEST passes on the H4 or H5 but fails in the 520ST, that points at the machine rather than the ST536. If it fails in both, we look at the ST536 board itself, and run both BUSTEST.PRG and BUSTESTS.PRG on H4 /H5.

I have an H5 I can test it in.

Surprisingly this time it completed the Bustest cache off and on. On pass 2 now.

coonsgm wrote: 10 Nov 2026 20:02 I have an H5 I can test it in.

:thumbup:

The signals are a lot better on those boards are be interesting if that has any effect or not.

Surprisingly this time it completed the Bustest cache off and on. On pass 2 now.

Now that is weird...

I will have to check if something like burst was turned off and not turned on again next.. bear with me...