• Please review our updated Terms and Rules here

What, exactly, goes wrong with a motherboard?

C

cmc

Experienced Member
Joined
Feb 23, 2014
Messages
107
Location
South Bend, IN
Hi VCF folks,

I have an ENPC EP-PT11 motherboard from the Pentium 1 era. It's dead. The question is *why* is it dead?

The board isn't worth much and probably I should just pitch it, but I am too curious a person to not at least have a theory about what killed it.

1) There is no obvious physical damage to the board.
2) There were not capacitors swollen or leaking, but I recapped it anyway.
3) I thoroughly cleaned every socket, thinking maybe it was an electrical problem.
4) Tested with both a P1 133 and an AMD K6 233.
5) Tested with several different ram chips.
6) Quadruple checked all jumper settings against both the manual and what is printed on the motherboard itself.
7) Tried two different PSUs.
8) As a last ditch try, I baked the motherboard at 375 degrees F for 10 minutes, to reflow it.


So... what typically goes wrong beyond these things? At this point I'm not even trying to fix it really, I am just wanting to understand at a deeper level what went wrong.

Behavior is that it turns on, runs for five seconds, then turns off. No beep.
 
Last edited:
Heat over time, maybe? Had it been behaving a little erratically for a while and maybe seeming to get worse? If some IC has been running too hot for too long, my understanding is that it damages it over time and it will eventually fail. According to my understanding, overheating causes arcing on the silicon. Each time there's an arc, it builds a little stalactite/stalagmite thing between the lines on the silicon, sometimes causing a crash but sometimes not, and becomes more likely to arc later when things overheat the next time. Eventually, if the arc point bridges together, the chip is good and dead.

Maybe they didn't plan on the board being in service for this long, and didn't allow for enough overheating tolerances. The tinfoil hat part of me sometimes thinks that they do that on purpose as part of planned obsolescence. It sounds to me like you've tried about everything. I think I'd just pitch it at this point. Things of that era are still too common and too hard to work on to get too all up ons about unless you're just fixing it for fun, imo.

Of course, I may be completely full of poo, but that's my understanding anyway.
 
cmc said:
8) As a last ditch try, I baked the motherboard at 375 degrees F for 10 minutes, to reflow it.
Click to expand...

If it already wasn't working, then this killed it.

375 degrees is nowhere near enough to "reflow" anything. All you did was cause outgassing of toxic substances from the board into your oven and house, and kill components. Better go scrub your entire oven down, inside and out before you make any food in it again, and air your house out.

Electrolytic capacitors are not rated for that kind of temperature either, you either destroyed all of the new capacitors you installed on the board, or significantly reduced their life, as well as altering their characteristics.

Tossing any electronic device in any sort of home oven appliance designed for food is a big no.
 
I also thought of the overheating over time issue. Perhaps just failed a little over time until it died. But... what component would die first? Chipset?

As to the oven, the fearmongering and lecture is predictable but not interesting, thanks.
 
cmc said:
I also thought of the overheating over time issue. Perhaps just failed a little over time until it died. But... what component would die first? Chipset?
Click to expand...

Power components usually go first, or something shorts to ground.

cmc said:
As to the oven, the fearmongering and lecture is predictable but not interesting, thanks.
Click to expand...

It's not "fearmongering", it's fact. 375 degrees is only about half the temperature required to melt solder, and since there's minimal or no flux left in the solder, heat will just make the solder brittle and subject to fracturing. 375 degrees is also more than enough to cook electrolytic capacitors, which are usually only rated for 185-221F maximum. Going over this alters their characteristics, like capacitance, ESR, leakage, etc. because the electrolyte starts to boil. The life of the capacitor will be significantly diminished after such an event.

375 degrees is more than enough heat to liberate nasty substances from the plastic parts and residual fluxes and cleaning compounds left on the board from manufacturing. You do not want to breathe this stuff, let alone eat it.

If you want to reflow some surface mounted chip, you need a hot air station and flux, not your home cooking oven.
 
The pros and cons of the "oven reflow" have been discussed at length in many forums. Summary is that it can work as a last ditch effort, but don't try it unless you're going to throw it away. Anyway, I am interested in why a board may fail in the first place, not rehashing this old topic. Any thoughts?
 
cmc said:
The pros and cons of the "oven reflow" have been discussed at length in many forums. Summary is that it can work as a last ditch effort, but don't try it unless you're going to throw it away. Anyway, I am interested in why a board may fail in the first place, not rehashing this old topic. Any thoughts?
Click to expand...

Almost every single part on the board is require to work properly to have the board run. The exception is that they will almost always run with many of the rail bypass capacitors removes. In other words, most any part could have failed.
Now, when you say, it turns on for a while and then turns off. Does this mean that it boots and then shuts off or does it just mean that some light comes on?
If it is booting, your problem is likely in the power supply and not the mother board.
Although, many seem to think replacing capacitors without having the slightest idea what is wrong has worked for some, it is close to some of the quack cures for Covid-19 you'll find one the web. If it works, after doing that, you still don't know if it made any difference. Some people get over Covid-19 on their own, with or without the quack cure, the same is true for computers sometimes. You wiggle or move something that fixes it.
Some of the power supply electrolytics are common failure points but not all of them. The ones that are more critical are the ones that filter the input DC from the rectifiers and the ones that filter the rectified switching current to the output rails.
Dwight
 
Thanks Dwight, that is interesting. On the capacitors, is it not true that many mainboards failed in the 97-2005 range due to bad capacitors? And that many of these failed without obvious symptoms i.e. swelling or leaking?

I am fairly sure it is not the PSU. I have tried two, one of which I know works fine in another machine.

The behavior is that the board turns on for exactly five seconds then turns off. No beeps, no lights. It is hard to be more detailed because there is no diagnostic information.

I suppose any part could have failed, but is there any reason to suspect one part over another? I.e. the bios chip or the chipset? For example I have heard that many nvidia chipsets from 2005 tended to overheat and fail prematurely.

The next question is, if I suspected the chipset, say, how would I go about testing that theory?

This is a academic of course, not about actually fixing the issue. I am just using this as a chance to learn more.
 
Components, such as ic's, fail. It probably was the most common cause of computers ceasing to work, back when it was relatively easy to diagnose such problems, and cheap. I wouldn't suggest a close 2nd, but next on the list was bad solder joints, more commonly in the power supply. That's been my experience. Willy nilly replacing capacitors ... umm, yeah it's popular amongst internet junkies. But just because someone posts a remedy on the internet doesn't mean it's authoritative. I would never stick a board in an oven. I would attempt to reflow individual solder joints. Or, and l never actually had the pleasure, reflow the solder side of a board in a wave tank. Not too many of them around these days I guess.

Unless you have considetable troubleshooting skills and a bit of equipment, well even then it could be extremely difficult to track down a specific fault. If a specific chip was at fault, attempting to cool it could perhaps temporarily remedy the problem. They used to spray chips with freon to get them cooled down. I don't suggest you try that.
 
cmc said:
Thanks Dwight, that is interesting. On the capacitors, is it not true that many mainboards failed in the 97-2005 range due to bad capacitors? And that many of these failed without obvious symptoms i.e. swelling or leaking?
Click to expand...

More like the 1999-2009 time frame, but it was quite literally everything electronic, not just motherboards. Monitors, power supplies, DVD players, etc.

The vast majority of the failure was because of a stolen electrolyte formula from Nichicon that was incomplete, which was used by the Taiwanese and Chinese to pump out billions of faulty caps. These caps were the ones that leaked or exploded. Later on, other shady companies got a hold of a different electrolyte that wouldn't cause the capacitor to vent, it would just silently fail. Usually they'd go high ESR and blow something else up on the board in circuit with them, or go open and cause complete failure.
 
Also check the +-12v rails. Many power supplies have what I call a kick starter to get the regulation started. They then switch to the 12V to run the rails for the regulator's analog circuits. If there is a short on the +12 it may start up for a few seconds but once that bleeds off the capacitors feeding the regulators, the supply will shut down. Do some ohm meter checks of your supply rails to the supplies ground return, on the mother board.
Dwight
 
cmc said:
As to the oven, the fearmongering and lecture is predictable but not interesting, thanks.
Click to expand...

Lecturing it may be, but any given reflow profile for almost any electronic circuit board has to consider 3 things, the initial preheat & ramp-up time of the board/components (i.e.---thermal shock), the maximum time the leads and body of said components can stand reflow temperatures (i.e.---potential burnout), and the rate the board/components cool down (again, thermal shock). Components are designed to accept temperatures in a reflow oven, given the proper parameters...but simply putting something in an oven for an extended period of time will (as mentioned) only degrade both the life and acceptability of the solder and the components.

https://www.johansontechnology.com/ipcsoldering-profile
 
I had a P4 mobo starting to act weird, needed a couple of times to boot correctly etc. Then it died, won't boot at all. I figured since it was an ECS that I'd find a bad cap or two and that would be that. I did try the swap the power supply and pulled the drive cables before removing the mobo from the case and was sure the cap's were going to cure the problem. Nope. Turned out the video card shorted and was the cause all along. Assume nothing! Check the voltages. Swap memory and video cards. Maybe try a PCI boot test card and see if you can find where it dies. If that doesn't even show any life then time to check IC's for shorts or other failures.
 
Dwight Elvey said:
Also check the +-12v rails. Many power supplies have what I call a kick starter to get the regulation started. They then switch to the 12V to run the rails for the regulator's analog circuits. If there is a short on the +12 it may start up for a few seconds but once that bleeds off the capacitors feeding the regulators, the supply will shut down. Do some ohm meter checks of your supply rails to the supplies ground return, on the mother board.
Dwight
Click to expand...

Thanks all for the input.

That's a really interesting idea. Kind of goes along with what others said about a short occurring. So hypothetically, where would a short happen? Inside an IC somewhere? I understand it may be impossible to know in practice, but hypothetically what steps would one go through to isolate a specific component?
 
First measure with an ohm meter. If it is not shorted, that is not the issue.
If shorted, connect it to a bench supply with the current limit, starting at about 0.5 amps. Check around the board with the back of your hand looking for anything that is getting hot. If that doesn't find anything increase the current. Keep the voltage setting to less than .5 voltage and observe proper polarity.
Dwight
 
Chipsets can blow up for whatever internal reason (they will get very hot probably from a BGA connection that has cracked and acts like a heater). Voltage regulators can short out and burn. The usual bad caps.
 
You must log in or register to reply here.
Back
Top