VOGONS


First post, by flynth

User metadata
Rank Member
Rank
Member

I have this 386sx 25mhz board (a Dat302 rev. B) that in general runs well, however sometimes it has a problem on post. Specifically normal boot process is:
- gpu message appears on screen
- bios message appears on screen, then it counts 4mb of ram
- it displays word Wait and almost immediately I hear floppies do their thing (the seek noise) - i have two floppy drives there.
- the BIOS table shows up with details of the cpu, com ports etc
- it starts dos

However there are two places where it sometimes fails.

First one is after the word Wait. It just hangs there until I press ctrl-alt-del then it restarts and posts normally. It happens maybe once every 10th boot.

The second is, it goes all the way, but instead of starting ms-dos it displays garbage. It actually continuous to boot showing another message like when himem is loaded, but all messages are garbled. It then freezes, but this time I need to use the power button to start again. This second error is much more rare. It only happened twice in a long time.

So, should I be washing this board, deoxing the slots etc? Or is this some known issue perhaps?

Reply 1 of 9, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

Looks like bad options selected in BIOS.
In case you have a CF as a HDD, have you selected the correct options? Is it connected to power? Sometimes those cards start but can turn unreadable if no power is connected (usually a floppy connector).

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 3 of 9, by flynth

User metadata
Rank Member
Rank
Member
douglar wrote on 2022-11-27, 14:27:

Could be those things. Could also be a boot sector virus. Find a copy of fprot for dos and give your system a scan.

Hmm, I have all bios settings on default (one would assume on a MB with a soldered in cpu and no peripherals the defaults are reasonable).

I do have a CF card. I set the settings from the datasheet (incidentally this is the system that runs fine on a 33k cylinder cf card i mentioned in another thread - no 1024 cylinder limit is enforced here).

However, I doubt the CF card or even bootsector have anything to do it as yesterday I was testing a couple of old hdds so I disconnected the CF card and I booted from just the floppy (before I connected the drive) and it still froze once. Perhaps the hdd/fdd controller needs it's contacts cleaning? (it has lots of jumpers).

Reply 5 of 9, by douglar

User metadata
Rank Oldbie
Rank
Oldbie
flynth wrote on 2022-11-28, 09:49:

However, I doubt the CF card or even bootsector have anything to do it as yesterday I was testing a couple of old hdds so I disconnected the CF card and I booted from just the floppy (before I connected the drive) and it still froze once. Perhaps the hdd/fdd controller needs it's contacts cleaning? (it has lots of jumpers).

Based on your description, it sounds like this is something unique to the cold boot process, yes?

If it was dirty contacts, I'd should expect the errors to continue intermittently after booting, yes?

So the items that come to mind at this point are an aging power supply that doesn't provide enough stable current for your system until it warms up or it's a boot sector virus that is on your hard drive and your floppy.

I brought a 286 into my computer lab earlier this year that was having intermittent boot problems. By the time I checked for viruses, I managed to infect 3 floppies and a 486 in my lab with stoned variant. Problems went away after removing the virus.

Reply 6 of 9, by flynth

User metadata
Rank Member
Rank
Member

I start to think a virus possibility is quite high. I'm not sure about the psu. I can't remember if it happens only on cold boots, or if it happened after I powered down and up after it run for a bit.

Fprot was mentioned as an antivirus to try. I'll do that.

Reply 7 of 9, by flynth

User metadata
Rank Member
Rank
Member

I haven't had a chance to test for a virus yet, but I think this is definitely fdd related (or controller). I swapped in a different motherboard. A 386DX-40 and here it didn't freeze on boot, but occasionally on 1.44MB floppy access. It seemed to be affected by certain files (not executables). On trying to access them the machine would freeze completely. Next I'll do a virus scan, but if it doesn't help I think I might need to try swapping a disc controller/cables etc.

Edit: Fprot found no viruses 🙁

So back to square one. However at least it appears it is floppy related. I wonder, can any component(floppy drive, cable, controller) cause such lock ups ctrl-alt-del doesn't work?

Edit2: It was a floppy cable. After replacing the problem appears fixed.

Reply 8 of 9, by flynth

User metadata
Rank Member
Rank
Member

Well, it turns out it wasn't a floppy cable... Or at least not just a floppy cable.

I had multiple more of those lockups that were very weird. Specifically when I was trying to copy one particular file to a floppy, then sometimes on floppy access in general. I had multiple lockups in a row trying to copy one file from hdd to a floppy. I copied other files fine. I copied this file onto another place on the HDD, but when copying on the floppy it would lock up.

Then I have software called Star Manager to copy files off an old Commodore 64 disk drive using an LPT cable. When saving files to HDD it too would occasionally lock up.

Now I'm running a better memory test (memtest 86) to exclude the possibility of a faulty RAM/CPU issues.

If this is fine then it must be either my CF adapter, a CF card, or HDD controller. Does anyone know any DOS tools to stress the HDD/controller continuously for a long time to see if it locks up then?

Also, how does one find IO resource clashes between irqs and address spaces, just by looking at msd output, or are there some better tools anyone knows?

Reply 9 of 9, by douglar

User metadata
Rank Oldbie
Rank
Oldbie

I would try speedsys 4.78, Checkit 3.0 or Norton 8.0 for diagnostics.

You had mentioned that these things usually have trouble around boot time is probably relevant. Resource conflicts or a failing drive would be more random no?

This still sort of sounds like it could be related to old failing capacitors somewhere.