That's fantastic. I can't get close to as weird as that!

I'll just share my short story from another thread of yours yesterday, about the ill-fitting CMOS battery causing system hangs.

Pierre32 wrote on 2020-06-21, 08:12:

You reminded me of an issue I had recently on my Win98 box, which would randomly lock up completely. I ran through all the usual fault finding, reseating, even replacing components to no effect. Finally my eyes went to the CMOS battery, a button battery in a holder. I noticed the battery was a thinner one than intended, so was not quite as snug in the holder as it should be.

I raised the question online, not really thinking that could be it. And the general consensus was no, that will not be it. But it was due for replacement anyway, so I put a properly sized one in, and the lockups went away.

I could only surmise that the battery was making poor contact, which had some obscure downstream effect on something somewhere!

I've had some weird system issues but wow nothing even close to that. Excellent job hunting down the problem and solving it.

It's probably not too weird in the grand scheme of things but it seemed pretty puzzling to me at the time.

When I was in high school, one of my classmates asked me to help with his PC. It was working fine but sometimes would fail to turn on, especially after being moved, and banging on the case was usually enough to solve the problem. Well, until one day when my classmate banged on the right side of the case really hard and the computer stopped turning on permanently.

Long story short, one of the through-hole capacitors on the motherboard had a leg that was left untrimmed and was intermittently touching the case, shorting the 5V rail to the ground. When the owner of the PC banged the case really hard, something shifted inside and the leg was in permanent contact with the case - that's why the PC stopped turning on at all.

Another issue from the old days: my aunt had an old DOS PC that worked fine immediately after booting up but after running several applications it would start throwing "Insufficient memory" errors. Her AUTOEXEC.BAT and CONFIG.SYS looked fine if a bit messy. I couldn't find anything wrong either with the hardware or the software.

A couple of weeks later, I was reading one of "PC user guidebooks" that were very popular at the time. It was in the pre-Internet days and the book included descriptions of the most common computer viruses. One of the viruses matched the symptoms perfectly, my mother brought a relatively recent version of Dr. Web (which was a popular antivirus in Russia at the time) from work and we used it to "fix" my aunt's PC. The virus in question was staying resident in the memory but it was buggy and every time an infected program was run, it gobbled up more and more memory which led to the "insufficient memory" error messages.

Spent actual months trying to track down highly intermittent stability issues in my dual pentium pro build. Tried reducing overclock, changing memory, changing processors, etc. Turned out the fix was disabling a chipset feature in CMOS called "Read Around Write". The setting only needed to be disabled with both processors installed. Running at as a single processor was stable either way

This isn't a PC, but I have a Power Mac 9600 that, for the longest time, would just freeze in the middle of trying to boot off an OS install CD, with no real explanation as to why. It turned out that, of all things, the rubber on the CD spindle had turned sticky and must have been affecting the rotation of the CDs.

Do unsuccessful troubleshooting experiences count?

I've bought an HP box with Pentium 3 CPU from a shop few months ago and the vendor has shown that the PC was working correctly. After arriving to my house and trying to turn it on, the machine immediately produced weird graphics artifacts with HP logo and would immediately freeze. I have cleaned up all the dust, removed all non-essential components and even later recapped the board, but the best outcome I could get was to make it POST and successfully enter BIOS every 20th attempt; during every other attempt, it will not POST and stop at a completely random POST code. In fact, on several occasions, immediately after turning on it'll start with 06 post code, then slowly increment it one by one - 07, 08, 09 and so on, until it reaches FF and then wraps back to 00 (yes, I've actually waited for one hour and half with naive hopes that it'll display something...) It really seems to be a possessed zombie board.

That is Asus CUW-AM motherboard and Pentium 3 1Ghz, so seems to be a very nice system, but unfortunately, it currently belongs to my dead motherboard collection along with SOYO motherboard that won't cold boot and a very nice Asus motherboard for Pentium 4, which has a corroded CPU socket after previous owner being a heavy smoker. 🙁

This isn't me personally but I remember reading at the time (socket 7 timeframe?) in one of the common magazines - PC World, PC Magazine, etc., about a mysteriously slow board and it turned out one of the jumpers used to configure it (cache, I think?) was missing the metallic part inside and was just a non-functional piece of plastic.

appiah4 wrote on 2020-06-24, 11:46:

Until yesterday, when the weirdest thing happened. I would go into DOOM Setup and whenever I clicked on anything, the PC would reset. At first I thought it had to do with me replacing my ES1868 sound card with a CMI8330, because I was trying to go into the Music options when this happened, so I reverted to the old sound card but the issue persisted.

Great story, however I don't see how going into Doom's music setup would result in the beeper beeping and the computer resetting?

Oetker wrote on 2020-06-24, 16:26:

Great story, however I don't see how going into Doom's music setup would result in the beeper beeping and the computer resetting?

Maybe it was trying to play sounds through the pc speaker?
I wonder if something was wrong with the speaker and it drew too much current or something and that was what caused the reset.
Would it be the power supply that would induce the reset if that were the case?

I’ve had several motherboards where the act of loosening one screw holding the board to the case made it post and run stable

ykot wrote on 2020-06-24, 14:46:

SOYO motherboard that won't cold boot

Tried replacing the caps? My 6BA +IV did that from time to time prior to recapping, and it went away after.

Oetker wrote on 2020-06-24, 16:26:

appiah4 wrote on 2020-06-24, 11:46:

Until yesterday, when the weirdest thing happened. I would go into DOOM Setup and whenever I clicked on anything, the PC would reset. At first I thought it had to do with me replacing my ES1868 sound card with a CMI8330, because I was trying to go into the Music options when this happened, so I reverted to the old sound card but the issue persisted.

Great story, however I don't see how going into Doom's music setup would result in the beeper beeping and the computer resetting?

Doom's setup menu makes a short beep through the PC Speaker whenever you select a menu item. Funny, isn't it? You forget about it until it matters.

rmay635703 wrote on 2020-06-24, 16:54:

I’ve had several motherboards where the act of loosening one screw holding the board to the case made it post and run stable

Most likely something on the underside of the board was shorting against the motherboard tray. This is usually due to using the wrong height motherboard spacers..

I have a WANG 386 sx-25 computer that if left idle at the DOS prompt will eventually throw an error and lock up.

When it is actively being used it works just fine.

Tried swapping out the RAM and that didn't help so not sure what the problem is.

It isn't a system that will see much use except for testing so I haven't really spent much time trying to figure it out.

appiah4 wrote on 2020-06-24, 17:07:

Doom's setup menu makes a short beep through the PC Speaker whenever you select a menu item. Funny, isn't it? You forget about it until it matters.

Looks like my PC speaker is broken then, haha.

PcBytes wrote on 2020-06-24, 17:05:

ykot wrote on 2020-06-24, 14:46:

SOYO motherboard that won't cold boot

Tried replacing the caps? My 6BA +IV did that from time to time prior to recapping, and it went away after.

My Soyo P4S Dragon would do similar things. With blown out CPU caps it was stable until the CPU hit 100% load. After replacing those it wouldnt post until it warmed up, would run 100% stable, then wouldnt turn back on after turning it off until like 3 days later. Recapped all the secondaries and it's rock solid finally.

3Dfx Voodoo II SLI not working with official or Fastvoodoo drivers.

This was really weird. Had two mismatched cards, tried seemingly all possible driver combinations, bought new SLI cables, tested those cards on another motherboard...bought a THIRD card...nothing worked.

Until I tried some obscure modded drivers from 3dfxzone.it that made all of those three cards play ball with each other. 😁

schmatzler wrote on 2020-04-30, 20:29:

I've finally found out why none of my three mismatched Voodoo II's would work reliably in SLI mode regardless of the combination […]
Show full quote

I've finally found out why none of my three mismatched Voodoo II's would work reliably in SLI mode regardless of the combination: It's the FastVoodoo drivers.

They claim to support mismatched SLI out of the box, but I only got that working one time for some dubious reason. In all other cases my screen just turned itself off and that was it.

I now installed coolsmoky's drivers from 3dfxzone.it. They have the same behaviour by default (screen turns off, nothing to see), but when I set the registry key SSTV2_MISMATCHED_SLI=1 they work reliably with every game I have.

This is awesome. I'll stay far away from FastVoodoo now and use the reference drivers with coolsmoky's patched glide.dll files now.

(And I really thought I'd have to buy a fourth Voodoo II. Thank god I don't.)

I think I've told this story before:

Some time around 2005 or 2006, I was working at a mom&pop computer shop in the service department when a guy brought in his AMD Athlon system, I think he had a MSI K8 Neo 2.

Guy complained that the sound card he recently bought from us wasn't being detected.

It was a slow business day so I decided to pop his computer up on the bench and it was true - the sound card wasn't being detected - instead it was being randomly MISIDENTIFIED as the PCI Vendor_ID and Device_ID strings were becoming slightly mangled and not providing stable results. Of course, force-installing the correct drivers didn't help much when the ID strings would scramble after a full power cycle.

Me: "That's REALLY weird. Let me pop that board into our test system" ... and the sound card identified fine, reboot after reboot.

Me: "I bet a BIOS re-flash will fix this. Just a gut feeling... Lemme download all that..."

I downloaded everything needed to re-flash the BIOS for the precise version of this motherboard, flashed the system AND... BRICK.

Every BIOS recovery thing I tried failed.

Me: "Did I remember to have him sign our service liability form? ... oh shit, I was being fast and NOPE..."

Me to customer: "I don't suppose the motherboard is under warranty still?"

Customer: "Nah..."

So, since in that case management wouldn't back me up... out of my own pocket I ended up buying the guy a replacement MSI K8 Neo 2 motherboard...

Learned a valuable lesson that day: always have the customer/client sign before getting hands on the system!

But that motherboard randomly misidentifying a newly added PCI soundcard was some of the weirdest shit I'd ever seen, still to this day. And these days my gut feeling is that it was a sign of deeper issues than what a BIOS re-flash might solve... exactly what, I couldn't guess... 😀

Also got some strange issues.
First one: Wanted to connect an IDE-ZIP250 to a Soyo 4SAW-486-mainboard (onboard IDE) as secondary slave with a CD-Rom as master. Primary single drive is a CF-IDE-adapter.
Problem is, that Iomega Guest detects the ZIP-Drive but uses the same drive-letter as the CD-Drive and ZIP can not be accessed. The latter is eve true, if the Zip is the only drive (correctly jumpered of cause).
In the end i updated the board's bios with the newest inofficial version (found here on Vogons) with my newly arrived TL866-programmer and now it works like a charm (Bios even supports booting from Zip).

Second one was just yesterday. I got a bunch of old computer-stuff cheap (3.5-inch flopy-drives, Amiga-external drive-housing, CD-Drive, 5.25 and 3.5" cleaning discs and an external SCSI Zip100 drive) . I wanted to test the Zip in a 486-computer equipped with an Adaptec 2842-VL-Controller, which worked flawlessly with a ZIP250. So I connected the Zip to the Computer with correct ID-setting and termination and powered the machine up. Zip is correctly detected in Bios-scan of SCSI-controller, but DOS startup hangs on loading ASPI-manager for controller. System does this, even if I disconnect the Zip from the computer, so something in the Adaptec no longer works as it should...
This problem, I have yet to figure out...

By the way, the Zip drive itself works flawlessly in another SCSI-equipped machine (Siemens Nixdorf PCD-4ND-Notebook in its docking station with Adaptec AIC6somewhat controller-chip ).

Mostly things like drives that will only work in some really arbitrary cable twisting or jumper combination. Stuff like, these two drives will only work if one is set to CS, the other to slave, and they're on separate channels, but the CD-ROM has to be on the first and not the second ATA connector. I was amazed the first time I witnessed a Macintosh refuse to boot without a fresh battery, but that kind of thing seems commonplace now.

I had a group of nice Epson SD-650 and SD-700s. One of them worked well, the others, although looking perfect, wouldn't turn the light off or be recognized. So I threw them away! Last year, I was tinkering with my old system which I put the good SD-700 into, and it was giving me the same stuff. I just started disconnecting, reconnecting and moving the cable around (switch to the other slot and so forth) and then it randomly started working again. I'll always wonder if the other drives just needed one more chance (to be fair I probably spent a long time doing just that, but still).

That's the #1 thing, arbitrary recognizance of drives. Other stuff, I don't know, drivers mysteriously uninstalling themselves.

I have another that I wrote up but it's not "retro" so I guess I won't put it here.