Can you help me figure out what the lighting fried in my computer. : )

This last week our shop power pole was struck by lighting twice. The computer is one that I built and I had installed a rs232 serial port card, which was used to drip feed our cnc's. I also had a usb to rs232 cable. The chc machines to that were connected to the card had their rs232 ports fried, along with the card and the cable in the computer.

The computer was on a surge protector, but the protector exploded ( sparks, loud boom,) when I touched it to unplug the computer two days after the last strike.

The computer is almost normal. It is really really slow, and seems to freeze or lock up for a while then it will go again. We have some big software that is really heavy graphics that is almost unusable because it drags so much. I have looked all over in the case for popped capacitors or any burnt chips or such. I replaced the rs232 card and checked around while I was in there.

I am thinking that there is something that is was just tweaked a little bit, and is going to fail shortly. I would like to order what ever it is that needs to be replaced before it dies completely.

Thanks I have solved most of my tech problems here, great help. : )

I am just guessing but the power supply is proabably weak they are also easily returnable so I would start there

It may very well be the power supply being so weak that it's slowing down the computer, as Komondor mentioned. I've seen it happen myself, but unfortunately, that costs money to find out, so I would go with the other possibility that can cause similar symptons, and that's the hard drive.
This may be time consuming, so only do it if you have an hour or two to spare.
Open "My Computer", right click on C drive (assuming that is the drive with your OS) and gop to properties.
In the next menu, select "Tools"
On that page, choose the top selection (Cheack Now) for doing a disk check.
It won't allow you to do it then and there, as the drive is in use and it will offer to schedule it at the next reboot.
Check mark both options, then reboot and go do something more interesting than watching the grass grow.
If the PC is suddenly back to its former glory, then wonderful. If not, the PSU is certainly a strong option.

Ok I will run the disk check this evening when it isn't in use in the shop. Is there a way we can check the psu, it is a few years old so I can't return it. My husband can check voltage and such I am sure.

I am wondering about it being the psu, because it seems to have the most problem in the graphics heavy program. And we had issues with a smaller psu because of the graphics card before. But then again I can see where it could be the hd too.

I am running xp on it just in case that matters too.
Thank you for your quick responses. : )

Quote from 01txaggie : Ok I will run the disk check this evening when it isn't in use in the shop. Is there a way we can check the psu, it is a few years old so I can't return it. My husband can check voltage and such I am sure. I am wondering about it being the psu, because it seems to have the most problem in the graphics heavy program. And we had issues with a smaller psu because of the graphics card before. But then again I can see where it could be the hd too. I am running xp on it just in case that matters too. Thank you for your quick responses. : )

There are PSU testers [amazon.com] available for purchase, or the other option is swapping it out with the PSU from another machine, just to try it out, depending on how much time you have on your hands.

I've seen many RS232 ports suffer only partial damage from surges. They'd still work but only at slower speeds, like under 10 K baud, and give lots of errors at 50K - 100K baud. The damage was always in either the TTL-RS232 buffers or the switching voltage regulator that generated the +12V and -12V RS232 signal voltages from the +5V power source. Almost all RS232 ports made since 2000 use that kind oof regulator, even when the hardware has +12V and -12V available to it. The regulator chip itself usually survived the surge, but the two tiny ceramic capacitors for its charge pump would blow, sometimes violently enough that they'd completely disappear from the circuit board.

Apparently communications problems can slow a computer down a lot, not because of the slow data transfers themselves but because the system resets itself a lot (doesn't cause computer reboots). I've seen this even with PS/2 keyboards and mice. Another thing that can cause such resets is dirty power, not necessarily due to a bad PSU but due to damage on the motherboard itself.

At another website, somebody mentioned a CPU being damaged after a lighting hit and not being able to run reliably at its normal 3.2 GHz speed but only 1/4 that speed.

Quote from 01txaggie : Ok I will run the disk check this evening when it isn't in use in the shop. Is there a way we can check the psu, it is a few years old so I can't return it. My husband can check voltage and such I am sure. I am wondering about it being the psu, because it seems to have the most problem in the graphics heavy program. And we had issues with a smaller psu because of the graphics card before. But then again I can see where it could be the hd too. I am running xp on it just in case that matters too. Thank you for your quick responses. : )

Quote from RockySosua : There are PSU testers [amazon.com] available for purchase, or the other option is swapping it out with the PSU from another machine, just to try it out, depending on how much time you have on your hands.

1) I have had little success from psu testers, often psu's will pass on the detector but fail when it has a load on it.
2) you can also do the same as a tester with a multimeter as seen here [about.com]
3) if you have another computer you can swap parts in to that computer to test them, you can also wire the powersupply to the shop pc's motherboard to test it. I would NOT test the PSU on another desktop until after you tested it with a multimeter.
3) for future use get a higher end UPS with AVR (cheaper one's don't do AVR), surge protectors only do so much and slow to react as you just found out.

Quote from mrbobhcrhs : 3) if you have another computer you can swap parts in to that computer to test them, you can also wire the powersupply to the shop pc's motherboard to test it. I would NOT test the PSU on another desktop until after you tested it with a multimeter.

...with the AC unplugged, first using the ohms function to check for leakage between the high voltage side and low voltage side. Keeping those sides isolated from each another is a holy standard for power supplies, but lightning can connect them, such as by shorting transformers.

Here's a recent 1996 article about testing computer power supplies:

http://www.soloelectronica.net/fu...Repair.pdf

STOP and backup the hard drive first. What you're describing is exactly what happens right before a drive dies. If the data is valuable and not backed up, you need to copy it right away.

DO NOT run a disk check before you backup.