You are invited to Log in or Register a free Frihost Account!

Odd hardware issue

My old rig has 2 DDR400's (2x256MB) in dual channel, with Pentium 4 Prescott.
Sometimes, with no apparent reason, the PC crashes and the screen turns off (no signal etc). Many times it happened when I accidentally hit the table slightly with my foot, or even touched the case... I mean, such actions can't usually crash the machine. After I turn off the PC (power button), and turn it on again (restarting never helps), 90% of the time the screen still waits for a signal.
I did some hardware checking using software like memtest and others, but everything seems to be fine.
The next thing I did was taking out one stick of memory, and the PC worked fine for some time (few days) with one stick of memory, but crashed again later.
I inserted the second stick again, just to try, and it worked, this time for several weeks without any problems, until it started crashing again.
Anyway, I don't really think it's a memory issue (although I'm not sure), but rather a CPU issue...
CPU temperature is fine though, so it's not a heat issue.

To describe the crash: usually I see some artifacts of the mouse pointer, like laggy movement, and visual artifacts. Sometimes the whole screen gets messed up (colors, shapes etc). Shortly after the artifacts, it stops responding and the screen stops receiving the signal.

How can I check for sure the origin of this problem?
Do you have integrated graphics? Maybe your motherboard is going bad. Dead graphics would explain the random artifacting. Try putting in an old pci graphics adapter and plug your monitor into that. I'm not sure what else could be the problem besides a completely dead motherboard or a crazed power supply.
I have 9600XT. No way it's a graphics card issue because here I am typing. If the card was screwed, I wouldn't be able to see properly what's on the screen.

Earlier today my PC crashed yet again, so I just switched the memory sticks, and since then it works for few hours.
I also did some connection checking (shook the parts and I/O connections a little bit), and the connections seem to be find. No crash.

Anyway, I'm still not sure whether it's the CPU or the RAM modules...
There is one computer in my company that acts the same manner. To me, I feel it has to do with the motherboard. All you need is a slight adjustment of the CPU or an accidental tap on the casing and the computer may just crash and refuse to reboot until I open the casing and readjust the hardware. Have you tried clearing out the dust? If it's a RAM issue, you'll most likely hear the beep when you try to restart your computer.
I hear no beeps :O
And there's no dust in there, I clean it regularly.
As I mentioned earlier, I checked for loose connectivity by manually shaking the parts inside, and nothing happened.
Oh well, I guess I'll just buy new RAM anyway, if that will solve the problem - good, if not, then I'll just let it die Wink
well, from what i read. Have you searched your pc for a worm virus?? If you changed the memory dimm and after some days the system crashed, if your cu temp is fine and if the tests are ok. then maybe your prob is a vius. of course it is not usual to just touch the case and loose your monitor signal, but maybe it is a random shutdown of your memory, you vga, or ........
It might be everything m8.
However, run msconfig. and look the services that are running. if you see something strange, search it in a google search and clear it .

With my best regards
Alright, I will run additional scans, but I'm pretty sure it's not a software issue.
Badcaps? Certainly not a software issue.
By badcaps you mean bad capacitors?
I've checked the capacitors, and they seem to be fine. I'll check again though... :S
Sudden random instability and crashes are a typical symptom of motherboard failure in general and bad caps in particular.

By the way, is your motherboard mounted correctly in the case, i.e. is the board itself separated from the case wall with those golden mount screws, or is it mounted directly to the case wall / in contact with the case wall one way or the other? If you're not sure, first try to remember if the motherboard was mounted by a professional?
It's mounted correctly.
I had the same exact problem with our dedicated firewall server at the office - every time you would bump the server rack a tiny bit it would reset. Ran a BSD distro, and that was the only thing it did (three zone firewall plus a DMZ), so it was not a software issue at all. We replaced it, I took the board home, a few of the larger can caps were leaky.

You might also want to check and make sure that all the power switch connections on those headers are connected all the way.
Just before the crash you were getting video artifacts, then no signal? It's a video problem. It could be a bad AGP (or PCIe) bus, bad socket, bad power connector, bad driver or bad card. But it is almost certainly video related somehow. Just because the card doesn't act up 100% of the time, that doesn't eliminate it as the cause. The memory isnt acting up 100% of the time either. The artifacts 'could' be caused by bad RAM, but the signal error is not bad ram, thats no video signal from your card. Perhaps it is overheating. Do you have it overclocked? (the vid card I mean)
Nope, but it is quite hot where I live. I have those crashes in the winter too, so it can't be the heat IMO.
There are a few things, some already mentioned that could cause these problems.

One thing, it could be your RAM...they might just be going bad or be bad in general. You can check this by running memtest86 for a few hours to see if any errors are generated.

Another thing it could be is your graphics card. I know you said it is fine, however, I had the problem with my Network Gateway server and it was just a matter of taking out the graphics card and reseating it, making sure that the card was in firmly and I screwed it in.

It could also be your motherboard. If it is an old motherboard, it's possible that, depending on how you checked the caps, that they are going bad but there is no physical evidence of it. If you checked it with a volt meter and it still reads normal on EVERY cap then that isn't it.

Check your PSU, it is possible that there is a poor connection that is critical to keep the system running. Make sure you check all power connections and make sure that the pins on all connectors are in good condition, no corrosion or stuff on the pins.

Lastly, check for anything that might move when you giggle the case or knock into your desk. Loose wires, something behind the motherboard that might move. In particular look for a wire that has the protective rubber maybe cut or scraped away a little. It could be that by knocking the case it causes a normally ok wire to touch something and groud itself out which would explain it shutting down suddenly.
Thanks for the TheGeek, I will do some re-checking.
But the PC works just fine for few weeks now :O
I will update as soon it crashes again.
Related topics
How do I disable scandisk in XP?
server 2 has problems?
corrupt data on dvd
Computer won't route
Slow wireless connection but excellent signal
Windows XP blues
G3 and Tiger Raid issue, Help
Why cant i read dvd9?!
Computer restart
Graphics Problem
Problem connecting my new laptop to my home network & DS
XP problem
Problem with one of my computers....
Reply to topic    Frihost Forum Index -> Computers -> Hardware and Electronics

© 2005-2011 Frihost, forums powered by phpBB.