IMO they just think you're a n00b and try to jerk you around, but when you show some resolve, and threaten them with the 150W soldering iron, they cut the crap and play nice.

Maybe I should try some of the cards I have piled on my bench then. Perhaps they will have come to their senses without needing to be repaired at all. Haha.

I had one 286 motherboard like that. Bought, received, tested - doesn't want to POST - diagnostic card didn't show anything - tried different memory and no success. Left it on the table. After almost a year I was cleaning up my room/desk and I see it again. I retest it and this time it boots like nothing. Now it's been like 6 months and it still works fine 😀

Ugh, all the time. Earlier I set the 112mhz setting and it was like so you’ve chosen 100 mhz. Later it was all like congratulations you have set 112 mhz! I think oddly a bios setting did it despite being set by dip

Kahenraz wrote on 2022-02-24, 04:25:

I always test any vintage hardware that I buy to make sure that it works upon receipt. Over the years I have had several instanc […]
Show full quote

I always test any vintage hardware that I buy to make sure that it works upon receipt. Over the years I have had several instances of testing where something seemed defective but after repeated power cycles and slot reseats the problem magically goes away.

Today I was testing a Sound Blaster 32. At first there was some issue where unisound reported that it could configure the card but the EMU8K wouldn't initialize. Running diagnose.exe displayed various errors of not being able to set DMA or a problem with the configured port. The exact error isn't important because it seemed random.

I shut the system off, reseated the card, booted, and then everything seemed to work. Unisound could detect the EMU8K just fine and there were no issues with diagnose.exe. Several reboots and reseats and I could never reproduce the problem.

I have had similar issues with video cards where they would appear dead with the system beeping as if no video card was inserted. But reseating or power cycles would make the problem go away.

I have suspected that the issue may be oxidation on the card edge pins or maybe the repeated power cycles reforms the electrolytic capacitors enough to spring back to life. But due to the inconsistent nature of the problem and then having it disappear entirely has always put it to the back of my mind with no clear answer.

Has anyone else experienced this or something similar in their own testing?

the secret is how you poke you tongue out of your mouth while musing over the errant behaviour of the hardware.

Get the tongue angle just right and boom the hardware realises whos boss and starts working, having a 12 pound sledge hammer close to hand also works.

Such hardware should always be place in a pyramid of power for several months to let the universe recharge the device with its energy.

Intermittent problems are the worst.

I had a few items I'd ordered in the winter months that showed up apparently dead. But after a few weeks of sitting at a relatively stable room temperature, they came right back to life and hummed along perfectly ever since. I think sitting in a freezing shipping truck for a few days just sent the thing out of whack. But who really knows.

I leave things wrapped up when they get super cold, until they come up to room temp, otherwise you can get humidity condensing on them.... then have to wait for them to dry out.

Yes, temperature/humidity is a factor. My old (c. 2005) backup server is much harder to power on in winter than summer. I guess contacts are tighter.

Murmuring incantations to the machine god also helps

I have a sacred Trident VGA of healing. More than one motherboard that looked dead to the world was able to post once the card was inserted. Without changing the IDE cable around!

Actually, a sacred card of healing might be worth making... something that draws 90% of spec on all the lines, to pop all the shorted tantalums and help reform electrolytics with a few power cycles.

BitWrangler wrote on 2022-02-24, 16:41:

Actually, a sacred card of healing might be worth making... something that draws 90% of spec on all the lines, to pop all the shorted tantalums and help reform electrolytics with a few power cycles.

Oh Great Magic Smoke God !

My Shinko CHC445 and hp 733 Celeron take about 10 minutes of being plugged in then the power switch starts working

Yeah I've got a Packard Bell 286 that was going that way last time it was set up, I figured I gotta do some recapping.

BitWrangler wrote on 2022-02-24, 22:42:

Yeah I've got a Packard Bell 286 that was going that way last time it was set up, I figured I gotta do some recapping.

Dont forget to check for weak solder joints while you are in there !

Oh back to that Trident VGA, it's worth trying ISA VGA in pre-PIII PCI boards that are refusing to boot, because I think they can get themselves in a state of resource confusion with the APCI hubs and "Just can't PCI" until they sort themselves out by booting with no PCI devices.

I have an old Matrox PCI card that I recently bought because it is super long and has a ton of RAM slots on it.

First system I tried it on all I got was a bunch of jumbled characters. Figured it was just dead.

Tried it a few weeks later in a different board and it worked fine. Tried it in another board and it worked fine as well.

Old hardware can be finicky if it hasn't been used in a while. Old caps are a good bet but I also suspect that sometimes the transistors in circuits can get stuck in one position sometimes.

Oh, there's a static thing that can happen, it's not static damage per se, most of these chips stand super high voltages, but just small charges can back bias transistors and make things gridlock. So just grounding across all contacts, or leaving it for long amounts of time for charge to leak away naturally, can cure this and make things work again. A particular frequent example was the DDC getting confused on CRT monitors, they'd get a lot of surface charge from the monitor itself building down the cable, and once in while the monitor would blank screen... so you'd have to plant a slightly damp thumb across the all the contacts on the VGA connector to ground it out and then it would behave for another 6 months or so.

edit: I think this is the thing responsible for when you assemble system, try booting, nothing happens... so leaving it assembled and plugged in, you go and grab a coffee or something to sustain your suspected to be prolonged troubleshooting session... but when you com back, you flip the switch to see exactly what it's per-tick-you-lar malfunction is and it springs to life and works perfect.... yeah with everything plugged in, the charges could drain over a few minutes through all the tie up and tie down resistors in the completed circuits and go to earth.

In December, the bios chip failed in my SuperMicro X9DAE motherboard. I replaced the chip 3 times, and the last attempt with those bios chip services on ePay. Board never came back to life. Few days ago I was cleaning, and had the memory / processors handy, and decided to try again. Board booted just fine and has ran stable as a rock since. I have it doing some bitcoin mining just to stress test it.

its a freaking Christmas miracle.