PC stability issue, CPU to blame?

[russianboy]

So I had a mishap with my PC a while ago and to sum it up, I am on my third Biostar motherboard.

The problem with my PC is that it restarts by itself. Usually it takes about half an hour or so, I got lucky when it finally lived through the installation of Windows 7.

Sometimes it's a clean restart out of nowhere. The screen goes black and the system reboots. Other times the screen goes black and hangs, the computer is unresponsive, the power light on the PC often times will shut off and the monitor will indicate no input. The motherboard has four "phase" LEDs which indicate power consumption/overall load, and in the time that the system is not functional, all four phase LEDs will be lit. Two troubleshooting LEDs on the board extinguish, indicating a CPU/chipset related issue.

Temperatures of the chipset have at times been very hot, I would not be able to touch the heat sink for very long without it burning my finger. At this point Speedfan indicates no issue with temperatures and the chipset is not hot during normal operation. However, one of the system temperatures reads -75 degrees Celsius, leading me to question the validity of the program.

Memtest passed, however CPU stress tests do not pass. Voltages seem all within reason. At the moment, I am not sure what component is to blame, and I am wondering what course of action I should take. Are there any thorough tests that can pinpoint an issue? Please advise.

Thank you.

Specs as follows:
Biostar 790GXE 128M, replaced for the third time
AMD Phenom x4 9600
On board video (3300 chip), have a 5500 available
Seagate 160gb PATA drive
Thermaltake 400 watt PSU, Ultra 500 watt PSU available
2x 1 gb of a-data DDR2 800.

 

[russianboy]

Anybody?

 

[No3Dalefan]

Try "CPUID HWMonitor" to see if you can get some reliable temps for your CPU. It's a free download.

Remove one stick of memory at a time and see if the problem goes away.

Select "Safe" defaults in your BIOS, maybe something is tweaked a little too much.

Some things to try.

 

[Conflict0s]

Try doing another check to make sure that everything is where it should be and connected correctly to the motherboard etc.
Try running Prime95 for a couple of hours whilst keeping an eye on your temperatures with HWMonitor like No3Dalefan adviced.

Before hand I would check your BIOS and see if you have a setting that turns your PC off if the CPU hits a certain temperature. Note what it is set at and see if your CPU hits that point when it is being stressed. If it does, then I would either look into borrowing or getting another cooler or seeing if you can send the CPU back as the default cooler should be able to manage the temperatures given at stock.

 

[russianboy]

Removing one stick of memory did not yield any difference. I used HotCPU and during the first 40% of the first CPU stress test (about 15 min) the system died then restarted. BIOS is at stock. I'll download the software listed here and determine the temperature.

 

[twilyth]

This borders on being off topic, but an interesting fact that be useful.

I have a 1090t in an xhair IV. With stock cooling on the old m/b I could get it to 4Ghz (nearly) and stable. With the new board I used Ceramique. At first I could get past 3.8 but I had to push more voltage than before and it wasn't stable. Now though I can run at 3.9 with much less voltage, much lower temps and it is completely stable.

Ceramique I believe is the type of paste that has to "settle" to be most effective. So if you're using something similar, it might take a couple days and several times of cycling off, cool down, cycle on, before you get the best performance.

That's probably not the issue, but good to keep in mind if you use that sort of TIM.

 

[robn]

Go into the BIOS when it restarts and check chipset temperature there, should be more correct. If the chipset feels that hot it could well be your issue.

This reminds me of an AthlonXP system I had which kinda half turned off when pushed - reset itself without actually restarting, just blank screen, with fans running - and that was entirely due to a safety cut out feature of the chipset which was overheating. I changed CPU cooler and PSU before I worked it out. The new CPU cooler made it worse since less air blew across the motherboard. Fixed my issue by replacing the hot heatsink with a Zalman aftermarket one, and modding an exhaust fan into the back of the case. Wouldn't mind, but my sister's identical machine ran cooler!

Anyway I would suggest pointing a fan at your chipset heatsink.

 

[JATownes]

I use OCCT on small data set to test the CPU clock stability, then
large data set to test remainder (i.e. CPU/NB voltage, etc.) I use HWMonitor for temp testing. Try these and post a screenie running OCCT with CPUZ and HWMonitor running in the shot. Rebooting could be caused by overheating, but the only way to tell is accurate temp reading. Also. have you tested your PSU? A faulty PSU could also cause the problem.

 

[Trickster]

Are you running Windows 7 64-bit? IIRC, certain AMD Phenom 9600 CPUs had problems with 64-bit Operating Systems -- BSODs and reboots, etc. I'm not saying that's the case with you but maybe you should investigate that prospect. You may want to install a 32-bit OS and see what happens.

 

[russianboy]

Are you running Windows 7 64-bit? IIRC, certain AMD Phenom 9600 CPUs had problems with 64-bit Operating Systems -- BSODs and reboots, etc. I'm not saying that's the case with you but maybe you should investigate that prospect. You may want to install a 32-bit OS and see what happens.

Actually I am running a 64 bit copy of Win 7, but the issue was with 32 bit XP and even during the installation of Windows. Hmm...

 

[russianboy]

Also, memtest on one stick ran for two hours and passed, something that never occurs in windows.

 

[Magikherbs]

So I had a mishap with my PC a while ago and to sum it up, I am on my third Biostar motherboard.

The problem with my PC is that it restarts by itself. Usually it takes about half an hour or so, I got lucky when it finally lived through the installation of Windows 7.

Sometimes it's a clean restart out of nowhere. The screen goes black and the system reboots. Other times the screen goes black and hangs, the computer is unresponsive, the power light on the PC often times will shut off and the monitor will indicate no input. The motherboard has four "phase" LEDs which indicate power consumption/overall load, and in the time that the system is not functional, all four phase LEDs will be lit. Two troubleshooting LEDs on the board extinguish, indicating a CPU/chipset related issue.

Temperatures of the chipset have at times been very hot, I would not be able to touch the heat sink for very long without it burning my finger. At this point Speedfan indicates no issue with temperatures and the chipset is not hot during normal operation. However, one of the system temperatures reads -75 degrees Celsius, leading me to question the validity of the program.

Memtest passed, however CPU stress tests do not pass. Voltages seem all within reason. At the moment, I am not sure what component is to blame, and I am wondering what course of action I should take. Are there any thorough tests that can pinpoint an issue? Please advise.

Thank you.

Specs as follows:
Biostar 790GXE 128M, replaced for the third time
AMD Phenom x4 9600
On board video (3300 chip), have a 5500 available
Seagate 160gb PATA drive
Thermaltake 400 watt PSU, Ultra 500 watt PSU available
2x 1 gb of a-data DDR2 800.

Does that ps have 2x 12v ? 400w, seems a little weak for a descent board like that. Id also try manually setting the cpu/nb in the bios. Whats your cpu voltage? My old Phenom x4 9150e was default/max @ 1.125v. Nb was 1600mhz. Also make sure the cpu is seated proper/clean. What pattern does the thermal paste leave when you remove the HSF?
good luck !

 

[Mussels]

from the symptoms, i'd say PSU.

lower end thermaltake PSU's are utter shite, and not quality at all.

 

[russianboy]

Tried Ultra 500w PSU, no improvement. Turns on, no post, no anything. Two diagnostic lights are off, signaling CPU/chipset error. Four phase lights are all lit.

How is AMD with replacements/RMA procedure?

 

[Magikherbs]

Tried Ultra 500w PSU, no improvement. Turns on, no post, no anything. Two diagnostic lights are off, signaling CPU/chipset error. Four phase lights are all lit.

How is AMD with replacements/RMA procedure?

Have you tried the cpu in a different board ? or different cpu in your mobo ?

 

[russianboy]

Have you tried the cpu in a different board ? or different cpu in your mobo ?

I have only one other computer available and I can't risk both PCs going down, I don't think that is an option.

 

[Magikherbs]

I have only one other computer available and I can't risk both PCs going down, I don't think that is an option.

How else do you expect to find out ? there is very little risk involved. Just make sure to use good/enough thermal paste.
Are there any bent pins ? What mobo power connectors are you using ? Also check the amps on your 2 12v rails, they should be at least 15-20A each.