Troubleshooting dead PC killed by lightning

July 26, 2010 at 08:36:07
Specs: Windows XP
Received a dead pc killed by lightning recently. I trouble-shooted that the PSU was fried, everything else A-OK. So I replaced the power supply which fixed the problem...for a short time. The PC is now dead again, PSU does not respond to the paperclip power-up, so something has caused it to fail. Is it possible that the motherboard may be faulty causing the power supply to fail? That seems to me like damn poor design. Also makes me very wary to test the motherboard with a functioning PSU save that end up as landfill too.

Assuming the motherboard is busted, what's the likelihood of the CPU being broken too? Would you risk buying a replacement motherboard and slotting in the existing CPU or would you say it would be more sensible to buy a fresh pair, new CPU, new motherboard. In other words, what I'm asking is: could a fried CPU kill a motherboard (similar to the first question)? What about RAM and and cards connected. Basically, how good is fault isolation in a pc setup?

It seems troubleshooting PCs can be an expensive business...

Muchos grassy arse for reading and +internets for any responses to yet another tit on the internet with hardware problems.

Although I'm sure it won't make much of a difference, for protocols sake, the beast in question is an Acer Aspire M1640 Desktop, that's an E4700, generic OEM MSI D33008 motherboard, 2GB DDR2 of the good stuff and a Ge9300 megagrafixcard takes residence in the PCI-express.

See More: Troubleshooting dead PC killed by lightning

Report •

July 26, 2010 at 08:59:40
It's unusual for a motherboard to cause a PSU to fail, and if it does happen there's often an accompanying aroma of burning electrical bits on the motherboard. I'd suspect you just had a faulty PSU.

Having said that, It's a good idea to keep a decent PSU for swap-out purposes, they have more protective circuits so you don't run the risk of a bang from a faulty board.

Also, if you're doing this on a regular basis I would recommend an ATX tester - much better than the paperclip method plus it will check all outputs - better ones have a voltage readout.

"I've always been mad, I know I've been mad, like the most of us..."

Report •

July 26, 2010 at 09:14:02
are u sure u replaced ur psu with the right power requirements? since u said it worked for some time and then died again
and ur mobo can never hurt a psu unless the mobo is fried

Report •

July 26, 2010 at 09:16:53
Unfortunately, a power surge or spike caused by a lightning strike can damage anything connected to the computer.
Also, sometimes a high voltage spike caused by a lightning strike can jump across switches that are off, and get past a surge / spike protector device.
The only sure way of protecting the computer from that is to, while a lightning storm is going on, unplug the computer, and unplug everything connected to it that connects to AC power directly or indirectly (via an AC to DC adapter) , AND unplug the cable that connects you to the internet if that applies (e.g. the telephone line, or the coax cable, if the computer is connected to a dial-up provider, or to high speed modem or router/high speed modem combo by means of a network cable).

Sometimes a power failure event just makes connections poor; sometimes it cause damage you can find.

Unplug the case/power supply, or switch off the AC power to it otherwise.
Power off your monitor.
Open up the case by removing the left side panel as seen when you're looking at the front of the case.
Check all the connections of the wiring to make sure they are all the way onto their pins and into their sockets, especially the main connector from the power supply. The wires close to the mboard going into the main power connector/socket should be more or less perpendicular to the mboard surface rather than at an angle.
Make sure all cards in slots are all the way down in their slots.
Make sure the ram is all the way down in it's slot(s)

While you're in there, if the cpu fan/heatsink has mung (dust, lint, etc.) on it, clean it off, but DO NOT use a vaccuum cleaner to do that (they produce a tremendous amount of static electricity when running, and anything connected to them can discharge that to your components) - use canned air, or an air nozzle if you have access to an air compressor, or an artist's brush that can be used in small spaces, etc. It may be difficult to clean the top of the heatsink under the cpu fan - the most likely place to have mung on it - and the bottom side of the cpu fan blades unless you remove the fan. If you have a case fan, clean that too if it needs it.

Also check for mung on the video card fan and heatsink if it has that, and the power supply's openings / fan.

With the cover still off, restore the AC power, start the computer and make sure the cpu fan spins - if it doesn't spin, if you're sure the power supply is working okay, don't use the computer until you have replaced it.
If it spins too slowly, and/or if it makes rattling or screeching noises, most likely to be noticed when the computer has cooled to room temp, has not been used for a while, and then is started up, the cpu fan's bearings are failing - the cpu is likely to overheat as a result of that if it can no longer spin it's full speed - replace it as soon as you can.

If that doesn't help,

Unplug the case/power supply, or switch off the AC power to it otherwise.

Remove all wiring connectors one by one and check for evidence of damage - damaged contacts or pins, carbon deposits - and install them again one by one. Clean up any carbon you find.

Do the same with all cards in slots, and the ram.
Wipe off the contacts on the cards and the ram with a tissue or soft cloth before you plug them back in. Make sure the notch(es) in the ram modules contact edge lines up with the bump(s) in the slot, the modules are all the way down, and the latches at the end of the ram slots are against the ends of the modules.

The following is about electrolytic capacitors that were not made properly failing eventually, but you can see the same damage from a power failure event, especially one caused by a lightning strike.

Some mboards develop this problem - electrolytic capacitors were installed on them that were not properly made, and they fail eventually - the mboard manufacturer didn't know they were improperly made at the time the mboard was made.

NOTE that many MSI mboards have this problem.

Open up your case and examine the mboard to see if you have bad capacitors, and/or other findable signs of mboard damage .

This was the original bad capacitor problem - has some example pictures.
History of why the exploding capacitors and which mboard makers were affected:

What to look for, mboard symptoms, example pictures:
Home page that site
- what the problem is caused by
- he says there are STILL bad capacitors on more recent mboards.

Pictures of blown capacitors, other components, power supplies, Athlon cpu's, etc.:

Report •

Related Solutions

July 26, 2010 at 09:30:28

Granted, it was a cheap PSU, but it had only run a couple of times, and to me, the probabilities point towards an external trigger given lightning's destructive nature, perhaps the motherboard, perhaps the surge protector.

Nonetheless, you have given me the confidence to test the motherboard with a working PSU of good quality that I own. I'll put the system through its paces; if it starts and lasts I'll attempt to RMA the faulty PSU for a more reliable model. Otherwise, a new MB I think.

Thank you for your insight and advice regarding the ATX tester. I'll try and post the results for information.

Report •

July 26, 2010 at 09:54:53
@toxict51 Power requirements were all in order. I replaced a 250W with a 400W model (boosted the power because its only cheap). Both are 230VAC and neither can be adjusted. Current on the ATX rails differ between the units but I'm assuming that is not a problem as this is true of all power supplies.

@Tubesandwires That is one jumbo response, thanks for your time. I had checked out the board before for any signs of meltage, but it looked ok, as were all the caps. However, I may have spotted what could be a burnout on one of the integrated boards, only the bloody realtek crab chip. Here are some pictures for your viewing pleasure, fresh off the camera:

Look, no mung.

Check out the realtek chip, the 8 and 2 of the printed code look to be a little blackened and it doesn't seem to rub off.

What do you think?

Report •

July 26, 2010 at 11:30:28
When only some of the characters are discolored on a chip, that quite certainly indicates that chip is damaged. The actual IC package is in the center of the chip, so that's the most likely place to see damage. When you see that's damaged, that indicates the voltage spike or surge got through to the mboard circuitry - anything on the mboard could be damaged - you most often can't see any physical damage on something electronic that no longer works.

The ram and cpu may still be okay, but the only way to tell for sure is for you or a technician to install them in a mboard they are compatible with.
Look up whether your ram brand and part number can be used with (is listed for) the mboard you try it in.
Your cpu must be on a list of supported cpus for the mboard model.

The drives, cards in slots if you have any, may or may not still be okay.

I've found that boards on floppy drives tend to fail most often when exposed to power failure events, then boards on optical drives, then cards in slots, in that order, and hard drives, ram, mboards, and last of all cpus, are much less likely to be damaged.
But - on the other hand, a power failure event caused by a lightning strike on the power grid can be a lot more damaging than any other type of power failure event.

You could remove the cpu fan and heatsink and examine the top of the cpu - example pictures of burnt cpus at the last link in Response 3 - but sometimes you can see no physical damage in any case.

Report •

July 26, 2010 at 12:24:42
Great stuff tubesandwires. Adding all the factors together, the slight discolouration on the chip, the failure of the PSU on installation and the nature of the surge, plus all the tit bits you guys have imparted, I think I'm going to go for a new motherboard and PSU, but keep the RAM and CPU and hope they remain okay.

Thanks also for divulging your experience in terms of components and their resilience to surges, interesting. But, funny things can happen. After all, the PC I'm using at the moment has a failed IDE Bus due to shorting and a non-functioning sound driver post a lightning surge - regardless the SATA works perfectly, as does the new sound card I installed! You can't fault the nForce 650i for holding up to that sort of abuse and still limping on! None of this MSI/Acer toss.

Thanks again.

Report •

July 26, 2010 at 14:27:04
My nephew has a mboard that only the USB controllers were found to be damaged on after a power failure event.

A friend lived in an extended care center that was having power problems - when the power went out the emergency generator came on, which may have caused power surges or spikes as well. That probably damaged the board on a Samsung combo TV/monitor connected to the computer, though it was at least a month later it malfunctioned (b&w TV when it was color before) , and shortly after the event I discovered the PCI-E x 16 slot circuits were damaged on her fairly new desktop mboard - despite the fact absolutely everything connected to the computer, including the telephone line, was plugged into either of two anti- surge protectors, which were still indicating they were protecting afterward. That mboard then worked fine with a decent PCI video card, but the PCI-E video card was an ATI AIW and she wanted to use TV on the computer, so I got another mboard. Then I found the AIW card was damaged, and got a recent HD AIW card.
(I would have made a warranty claim, but could not do that because the anti-surge power bars were made by two different brands - if they had been the same brand the damage would have been compensated for.)

Report •

Ask Question