Zygoptera on 21/8/2009 at 01:57
Having a fairly significant problem. Symptoms are random hard lockups and very occasionally BSODs which get more frequent over time, eventually leading to boot failure with no error beeps, but which can be fixed by simply turning the computer off for ten minutes. On successful reboot an 'overclocking failed' error message (no o/c was attempted) comes up at POST but the computer is then fine for days until the cycle begins again.
Tried reinstalling windows, run diagnostics/ stress tests (Prime95/ memtest/ HD check utility etc) so it does not appear to be RAM, overheating, cpu, HD, virus or graphics card related. Also flashed bios to newest version, updated all drivers, taken apart, cleaned and reassembled the computer etc.
Any possible pointers as to whether the PSU or the MB is to blame, or other possible culprits? I don't have an alternative MB, and the only other PSU I have is an icute 320W, probably worse than what's in there at the moment.
I guess the real question is if buying a new PSU (a decent one I'd later use in a new computer) has any real prospect of fixing the problem, given that I'm not going to buy a new MB for old hardware and that the rail voltages and stability under load of the current PSU seem to be fine.
Specs, no o/c except the GC
XP Home, fully updated
C2D e6400
2 GB RAM 667
Seagate SATA drive 320GB
Palit 7600GT (factory overclocked)
LiteOn DVD drive
Crappy 400W generic PSU
Asus p5ld2se motherboard with latest drivers and bios
If it is the motherboard I'm even more glad I ended up deciding not to upgrade earlier this year.
bikerdude on 21/8/2009 at 08:53
Quote Posted by Zygoptera
Having a fairly significant problem. Symptoms are random hard lockups and very occasionally BSODs which get more frequent over time, eventually leading to boot failure with no error beeps.
Hmm this sounds a lot like the motherboard, but the only way to be sure is to swap out everything one by one starting with the memory (mem, gfx, sata cables, psu, cpu).
If your in the Uk I might be able to help you with this.
Zygoptera on 21/8/2009 at 21:37
I'm in in New Zealand...
Unfortunately the only other modern motherboard I have access to is an AMD one so not any use for a swapout, and the intermittent nature of the problem makes diagnosing a fix problematic. I'll just have to live with it for the moment, it is perfectly stable 90% of the time.
bikerdude on 22/8/2009 at 02:53
Quote Posted by Zygoptera
I'm in in New Zealand...
Unfortunately the only other modern motherboard I have access to is an AMD one so not any use for a swapout, and the intermittent nature of the problem makes diagnosing a fix problematic. I'll just have to live with it for the moment, it is perfectly stable 90% of the time.
you swap everything bar the cpu and mobo then, and if the problem goes away then you know its either the mobo or the Cpu.... but my money is on the mobo...
Zygoptera on 16/9/2009 at 02:26
Well, it's fixed. If I take the side panel off the case, no crashes. Put it back on, crashes start again.
Weird, but I'll take it.
theBlackman on 16/9/2009 at 06:15
Quote Posted by Zygoptera
Well, it's fixed. If I take the side panel off the case, no crashes. Put it back on, crashes start again.
Weird, but I'll take it.
That is not a "Fix". Check that no wires or other things are being pinched by the side, but it sounds like HEAT.
HOT HOT HOT. Check your CPU FAN and Vcard Fan, as well as the cooler fans for the case. If it takes time to heat up and quit, and it runs without the side (more air flow) it is most likely HEAT buildup.
If it is a system you have had for a long time, check the fan blades with the side off.
If they are turning properly, then shut down the system, and look at the blades of each fan. If they are gunked up, use a toothbrush to clean the blades.
Then get a can of air and blow the case clean (or if you can reach in with a Vacuam cleaner wand use that).
When done, double check all connections, put the side back on and test run the system.
If you have a System monitor (there are free apps that will run on your desktop) check the CPU temperature. Sometimes it's only available in the BIOS.
TBE on 16/9/2009 at 07:21
Yeah, sounds like over-heating. If you've had the computer for a few years, your processor's cooler may have moved around if you've moved the computer. The thermal paste between the cpu and cooler becomes hard sometimes, and a little brittle. You'll need to look at your cpu temperatures in the bios with the side of the case on, and see how high they go. If the temperature isn't above 70 degrees Celcius, you should be ok. If it hovers around 70C or above, you are too hot. Ideal temps should be around 45Cto 55C. So if the processor is running hot, you should pull off your cooler, and scrape off the thermal paste til you get back to stock processor. Do the same on the cooler. Put some new thermal paste on the processor. Most people recommend about a drop the size of a BB right in the center. If you don't know what BBs are, they are the little metal balls you shoot out of airguns. Like a .177 caliber pellet size. Anyway, when you get that drop of paste on the processor, push your cooling fan back down onto processor, slightly twisting clockwise-counterclockwise. The paste should smudge. If you have too much paste, it will start coming out of the sides like a peanut butter and jelly sandwich with too much jelly. I personally put a ball of paste on there that's a little bigger than a BB. I've had good experiences with cpu cooling in this fashion.
Another thing to watch on some motherboards is that they are set to turn off if your processor temp gets to a certain temperature. That may be set artificially too low on the BIOS, and you can raise the shutdown temp to about 75C. My BIOS on a P4 board was set to shut down at 48C, and it would get that hot when I was gaming, or doing video conversions.
Zygoptera on 16/9/2009 at 21:40
It almost certainly isn't overheating as it will happily run stress tests overnight without crashing and all the temperatures show perfectly normal range- eg 42C idle, ~58C under load for the CPU- and the one thing I could almost guarantee wouldn't cause crashes was stress tests (prime95, memtest, 3dmarkxx, etc. though it hated the windows version of memtest) or playing games. Cleaning the fans/case is something I'm pretty good at, and I've cleaned and reseated pretty much everything twice since I started having the problem.
I would have said overheating too from a brief description, but I can't see how it is one of the usual suspects. There are two other heatsinks on the MB which could be the culprit, I suppose, but only one looks user serviceable.
I did wonder about the wire pinching thing, or even that the case top might be shorting something, but all the leads are nicely bundled and the case top touches nothing other than the rest of the case.
At this point I'm just happy that it's working.
Al_B on 16/9/2009 at 22:04
Another thing to consider is that it may be a mechanical problem. i.e. taking the side off the case may reduce stress on the motherboard which allows a dry joint to make a good connection. It's a bit of a stretch, particularly if lightly fitting the case side still causes the problem but it could be something to watch out for.
theBlackman on 16/9/2009 at 23:26
Does it run the tests with the side ON?
And the comment about the side stressing (torqueing) the MOBO. is worth looking into. Try the side on but not screwed down. If it works that way put in one screw at a time and see what that does.
I had a situation where the bottom screw of the two for the side panel torqued the cutout for the peripheral boards just enough to lift my sound card partially out of the PCI slot.
