Is it my power supply, mobo, or CPU?

Two weeks ago my system (specs as follows):

E2100allendale (was running at 2.8ghz), Abit p35-e, 8800gt Cooler Master Extreme power 650w PSU

started becoming unstable after 1yr of running fine. It started BSODing or rebooting with increasing frequency until it got to the point that it was locking up in the BIOS menu before I could change settings. I put the clocks to stock and was still having trouble. MY CPU voltages were reading way low (1.1 instead of 1.35 where it was set) so I decided to try my backup power supply. It's an ancient (but high quality) antec true control 550. When I pulled the old PSU, sure enough, the caps were leaking all over the place.

Installed the new PSU and every thing went back to hunky dory for another two days. Ran prime95 for a good 6 hrs and survived another 4-5 of gaming. Then, in the middle of playing GTAIV I got the dreaded BSOD. Now, the computer is back to its old ways again and often can't even get to windows before hardlocking. What's weird is that it was like flipping a switch. It was stable, then BAM BSOD and completely unstable.

So, any ideas as to what happened? I have two sticks of RAM, and trying them alternately does not make a difference. Doubt its the vid card since the computer seems no more likely to hardlock in games as opposed to anything else. That leave a few options:

The second PSU is also having problems. Seems like an odd case of lighting striking twice.

The mobo is bad and causing instability. If that's the case, why did replacing the PSU get it working at first?

Somehow fried the CPU with overvolting (maybe the voltage readings were way off?).


Thoughts, ideas? I can't afford to just buy a new PSU/Mobo/CPU and I don't have a spare mobo or CPU to trouble shoot.

What were the exact contents of the STOP error, and which OS are you running? What's the VID of your chip (using RealTemp--> Settings page--> Max. VID), and w/ CPU-Z open what's the current Vcore (Vcc) while running Prime or SuperPI?

What were the exact contents of the STOP error, and which OS are you running? What's the VID of your chip (using RealTemp--> Settings page--> Max. VID), and w/ CPU-Z open what's the current Vcore (Vcc) while running Prime or SuperPI?

Click to expand...

The stop errors are varied. In fact, it's just about a different one each time. I've also tried booting from a different hard drive with a different copy of windows to make sure it's not software related. The results are the same.

Unfortunately, it's at the point that I really can't get any diagnostic utilities running before the system locks up. When I ran prime initially after switching PSUs the voltages didn't budge noticeably while running prime.

In the BIOS under PC Health or Hardware Monitor, ensure that the current Vcore is at least 1.35v.

The bios is reporting normal voltages, though I don't know if they stay normal once the CPU start's working.

The following template is for a IP35 Pro, but should be somewhat similar to your board. Fill it out the best you can w/ all of your current settings...

I've tried several settings (including all default)

Original speed (with OC) was
350mhz external clock
CPU: DRAM 1:1

Right now clock speeds are on auto
CPU core set at 1.4 (raised a tad in an attempt at stability)
DDr2 is at 1.8 (default )
CPU VTT 1.25
Some of those other voltages are not on the ip35e

Temps are in the 30s-40s range
Memory set to default

I was able to get it priming (and oddly enough it ran for an hour before BSODing) and voltages did not waiver.

nealric said:

I've tried several settings (including all default)

Original speed (with OC) was
350mhz external clock
CPU: DRAM 1:1

Right now clock speeds are on auto
CPU core set at 1.4 (raised a tad in an attempt at stability)
DDr2 is at 1.8 (default )
CPU VTT 1.25
Some of those other voltages are not on the ip35e

Temps are in the 30s-40s range
Memory set to default

I was able to get it priming (and oddly enough it ran for an hour before BSODing) and voltages did not waiver.

Click to expand...

Neal,

You need to list your hardware completly and as Redduc stated, your BSOD messages with any files associated with it.

You need to list your hardware completly and as Redduc stated, your BSOD messages with any files associated with it.

Click to expand...

BSOD messages change each time- sometimes does not BSOD and just reboots or hardlocks.

I beleive my hardware is listed completely (spread out along the posts)

nealric said:

BSOD messages change each time- sometimes does not BSOD and just reboots or hardlocks.

I beleive my hardware is listed completely (spread out along the posts)

Click to expand...

Couple items:

Check your caps on that Antec, see if they are bulged or leaking too.

Check voltages with a DMM in a free molex plug, don't rely on software or BIOS.

What are your default RAM voltage requirements? You listed 1.8v default in the BIOS but you're not stating WHAT brand of RAM and specs. Not many RAM pairs need only 1.8v except value RAMS, performance RAMS need a little more. That's why I asked to list your hardware so it can be easily seen.

Test your RAMS one at a time with Memtest86+ from a boot disk at default speeds, no overclock. If they pass individually, test them in pairs full passes for several hours each test. There should be ZERO errors if they are good at default.

Not many RAM pairs need only 1.8v except value RAMS, performance RAMS need a little more. That's why I asked to list your hardware so it can be easily seen.

Click to expand...

Sorry, its actually more on the value ram side. Brand is OCZ, but it's their lower spec stuff. Sorry, I don't have exact model. It was running at that same voltage for the year before, so I'm pretty sure it's not being undervolted.

nealric said:

Sorry, its actually more on the value ram side. Brand is OCZ, but it's their lower spec stuff. Sorry, I don't have exact model. It was running at that same voltage for the year before, so I'm pretty sure it's not being undervolted.

Click to expand...

I checked the OCZ value type RAMs and they are 1.8v. Wouldn't hurt to kick them up a little to 1.85v if your BIOS allows, it won't hurt them. This may not be the problem, I don't know, but it's the easiest test to start with and many times is the culprit. Take your time, let the test run for a while each time. You can check voltages with a DMM while Memtest is running too.

I ran memtest and it did find errors. But I think it could still be the mobo causing them.

Seems weird that both sticks would fail simultaneously. It does not seem to change the stability much whether I am running one or the other. Tried upping the ram voltage and it didn't help.

I'm pretty sure it's not the CPU because overclocking it makes no difference in stability.

nealric said:

I ran memtest and it did find errors. But I think it could still be the mobo causing them.

Seems weird that both sticks would fail simultaneously. It does not seem to change the stability much whether I am running one or the other. Tried upping the ram voltage and it didn't help.

I'm pretty sure it's not the CPU because overclocking it makes no difference in stability.

Click to expand...

If you found errors at stock testing individual sticks, that's your problem. You can try other RAM slot pairs if you have four slots. Otherwise time for RMA. Rarely is it the CPU.

Just wanted to share that it turned out to be the mobo. $50 and a newegg refub later everything is running like a champ