Heat is an enemy here. Make sure there is proper air circulation (many older cases are really challenging in that regard) and specifically make sure that heat generating components like VRMs or some chips have good airflow around then (to avoid hot spots).

What I do is run the board for a few hours and then check every chip and component on the motherboard with a laser thermometer. Anything that I find is too hot to touch when your finger is left on it for thirty seconds, I add some form of heat sink. There is not much you can do about capacitors, but if worst came to worst, they are easy enough to replace. Another think to look into is the case temperature. Keeping this as low as possible will make the board and all other components last much longer. Just my 2 cents.

Tom

This is slightly off-topic, but DRAMs are prone to die through heat or overvoltage.

Which actually is a good thing. DRAMs saved the other nearby chips that way (acted like a fuse/short circuit).

So whenever you find faulty DRAM in an old motherboard/device, don't be upset. Be crateful. 😀

Once I get another "old" board, I check with eyes for any damages, remove batteries, replace if needed caps and wash the board, dry it with warm air. Test it and store in antistatic bag in a box. Most likley caps are number 1 failure or on old stuff like 386/486 some tantal caps.

If in a usable case/system, store it dry and check eyery now and then the caps for optical damages...

A few things can be done.

1) Proper ESD control - most of the time ESD will damage internal IC's but not necessarily flat out kill it. This leads to future failures when heat+current is ran through that damaged circuit over time. Use a grounded work area, wrist strap, and store components in ESD safe material.

2) Proper storage. As mentioned, store components in an ESD safe bag. Store them flat with the least amount of stress on the board; never stack components directly. (I use motherboard size boxes with foam on the bottom for each component) Broken solder joints and traces can occur over time with boards flexing. Keep the storage area environmentally controlled. This will prevent mold growth, oxidation, corrosion, etc. If stored in a basement, run a dehumidifier or central AC to keep the humidity lowered. Utilize shelving to keep boards or computers off the ground in case a flood happens.

3) Maintenance of boards. Replace electrolytic capacitors every 10-15 years. Some capacitors last longer than others, some puke their electrolyte onto the board, and some just don't last as long as the manufacturer claims. It also helps to energize components every 6 months or so to reform the capacitors. Remove batteries from boards for long term storage. Apply something like DeoxIT to connectors, sockets, and gold fingers - this will provide a protective barrier as well.

4) Ensure your test equipment is also up to par. Utilize power supplies with good voltages and low ripple. Invest in a multimeter. Check for shorts on components before powering them on.

itsgallus wrote on 2021-08-28, 15:43:

Hi there, I have already 2 Slot 1 Motherboards from 1998 that died I'm no expert in Hardware modding / soldering whats so on. But think once the Mobo is dead (makes no post anymore) and even a Debug PCI card won't show life anymore , its very hard to trace the error it could be litterly so many components, ICs , Traces , or bad Caps.

What is the best way to prevent a working Mobo from dying? Would a instant cap replacement help a lot? What other components are fragile? voltage ics?

Best thing is thorough diagnosis. A lot of components degrade over time, but the fact that caps generally go bad does not mean that is what is causing your board not to boot right now.

In terms of best pro-active practice, it's just the basics:
- remove batteries (*any* batteries, even coin cells leak eventually) before storage
- respect ESD precautions (don't go overboard, but keep electronics away from wool or synthetic clothes or floor covering)
- avoid moisture (not in the garden shed or cellar with stalagtites)
- protect from physical damage
- when you do hook them up, ensure you use known-good parts (particularly PSUs) and avoid short circuits or incorrect configuration.

That should keep good stuff good, but if the caps are subpar and dying, they can still do so in an ESD-safe climate-controlled vault. If ESD has fatally undermined ICs, careful handling later won't un-damage them.

But if it's already dead... apart from caps the one thing to check is for corrupt BIOS EEPROMs. I remember back in the day when there were still big computer fairs, I bought boxes of 'dead' motherboards. A couple turned out to work straight away, some had bad caps (more in 2001-2003 period), there were always a few FUBAR, but the biggest single group was the ones I could recover with a BIOS flash. Then I messed around with hotflashing, but these days I just use my TL866 - as soon as a board has boot issues but no obvious physical cause, I pop out the (E)EPROM and re-flash with the best image I can find. It's not a magic bullet, but it works often enough that it's a standard part of my workflow.

Just stacking them in a cardboard box out in the shed has worked for for over 20 years.

Ok guys thx a lot for your inputs. Luckily I could ressurect my motherboard at least for now its working.

I had a spare motherboard that's broken also but I took the Award BIOS Chip from and inserted it in my dead motherboard. after a few boots its runnin like normal again.

So maybe a dead Bios Chip? I should get may a USB EEPROM Programmer to read it out or then reflash it

itsgallus wrote on 2021-09-03, 20:10:

I had a spare motherboard that's broken also but I took the Award BIOS Chip from and inserted it in my dead motherboard. after a few boots its runnin like normal again.

So maybe a dead Bios Chip? I should get may a USB EEPROM Programmer to read it out or then reflash it

<rant>
The darn "Intel algorithm"...
originally it was specified that EPROM cells had to be charged for 20ms.
But then some people felt this too slow.
Then the "Intel Algorithm" was developed to charge the cells only partially, just so much that they retained the data.
But even (E)EPROM capacitors discharge over time, and so many BIOS chips with some forgotten zeroes make old stuff appear broken after decades.
</rant>

Edit:
Yes, it absolutely makes sense to "refresh" (E)EPROMs, like DRAMs... just a longer refresh cycle 😀

Regarding ESD, this is a true killer.
I religiously touch the electronics ground (case, ATX connector metal, mounting screwholes,...) before anything else to put myself on the same potential as the electronics.
The very few hardware kills I had in the past decade mostly happened when I forgot to do so, and discharged my potential by gripping at the electronics themselves. Sometimes these were instakills, so I know the cause 😀

Another killer is to underestimate heat: even if it is bad for the optics, some chips (GPUs, north- and southbridges in particular) run very hot and die early. Putting a cooler on them keeps them alive.

Make sacrifices to the old gods, keep their days holy, and mark the protected ones with the sacred runes.

BitWrangler wrote on 2021-09-03, 21:34:

Make sacrifices to the old gods, keep their days holy, and mark the protected ones with the sacred runes.

And hope the said old gods do not share Rawhead Rex's, euh, fetish, I guess . 😉

https://youtu.be/JjLPwuAHW3c

More seriously, always take your time and avoid shortcuts or compromises, however tempting or inoccuous they may seem . This is a general and possibly obvious thing , but in heat of excitement when trying out a new idea, build or testing a new part, mistakes can happen. Double checking things like chip and connector orientation, voltage settings heatsink contacts, etc will definitely not hurt .

darry wrote on 2021-09-03, 22:00:

More seriously, always take your time and avoid shortcuts or compromises, however tempting or inoccuous they may seem . This is a general and possibly obvious thing , but in heat of excitement when trying out a new idea, build or testing a new part, mistakes can happen. Double checking things like chip and connector orientation, voltage settings heatsink contacts, etc will definitely not hurt .

But the opposite may hurt !
I am lazy and stack my parts in ESD bags in various ways. Motherboards w/0 boxes put in shelving that are stacked Vertical not flat on top of each other, been doing that way for near 30 years and have never had a warped/bent board.
Vid cards, IO, scsi, etc cards in ESD bags in cardboard boxes laying flat (they are not prone to warping like motherboards). All in places that never freeze or get above 100 maybe 110f.
Am not a "got to have the wrist bracelet grounding" things but also never work on boards or parts on any carpeted areas and generally am barefoot inside (no fuzzy slippers ;p) and the work area is mostly concrete and linoleum like a jail 🤣

Caluser2000 wrote on 2021-08-29, 19:13:

Just stacking them in a cardboard box out in the shed has worked for for over 20 years.

I did that also long time, no damages observed yet.
Remember, cardboard always keeps humidity, and thus is (somewhat) conductive.

And, as @wiretap said, keep an eye on the PSU.

This shows +5V and +12Vof an old PSU, with 2V/div.
Nice ripple... what do you think this will do to to your valued old hardware?

(Image source: https://forum.classic-computing.de/forum/inde … n-und/&pageNo=1 )

BTW, this PSU was "working", and appeared completely "normal" from the outside.
So it might be a good idea to grab a screwdriver and visually check the PSU's inside for leaky caps, as such PSUs can become serial killers.