new system : random BSOD

[flashgeist]

Ich just bought new PC parts for my girlfriend : mainboard, CPU, RAM
Windows XP sp2 has been installed.

The problem is that she gets very often BSOD:

IRQL_NOT_LESS_OR_EQUAL

BAD_POOL_HEADER

PFN_LIST_CORRUPT

PAGE_FAULT_IN_NONPAGED_AREA

I joined some minidump files, I have more if needed

I checked the RAM with GoldMemory (excellent program), and it's fine. I tested the RAM on another PC and it was also running fine.

Can anyone tell me more be viewing the minidump files ? (I tried myself with Windbg, but I'm not good enough to interpret the results)

 

[howard_hopkinso]

Hello and welcome to Techspot.

All your minidumps are crashing with a different bugchecks, 50/8E/19/E3. Two of them reference hardware.

I believe you have a ram problem.

Go HERE and follow the instructions.

Regards Howard :wave: :wave:

 

[flashgeist]

thank you for your reply !

Unfortunately, RAM has been checked thoroughly. It's been even swapped, and the problem remains the same.

I forgot to mention that there are not only BSOD, but also different program errors (Outlook, ICQ, IE, Media player...). The errors are often different, but what my girlfriend noticed is that there are often several applications crashing at the same time (and I could checked that myself in the "event viewer / application").

Yesterday I went through a lot of testing : I ran sisoft sandra's burnin test during 4 hours (100 % cpu utilisation all the time), at the same time, I launched media player, outlook, word, msn messenger, icq (all what my girfriend usually uses). There hasn't been any system crash, any error. The CPU and board temperature remained both below 35°C (I don't know how much it is in Fahrenheit or Kelvin, but it's very low for a cpu at 100% load, it would even be low for a bath ).

So today I downloaded MBM 5 (Motherboard Monitor 5), and I noticed the following :
There seems to be a problem with the voltages :

I got the following informations :
core 0 : current=1.40V low=1.36V high=1.40V (seems normal to me)
core 1 : current=1.50V low=1.47V high=1.50V (don't know what this second core is, probably not used ?)
+3.3 : current=2.50V low=2.50V high=2.51V
+5.00 : current=5.12V low=5.10V high=5.15V
+12.00 : current=(changes all the time)V low=4.58V high=13.36V average=9.76V
-12.00 : current=-11.75V low=-11.75V high=-11.63V
-5.00 : current=0.00V low=0.00V high=0.00V

the +3.30 is quite low, the -5.00 appears to be not acurate at all (no value reported).
But the strangest is the +12.00 which is constantly (and quickly) varying, from 4.58V to 13.36V.

If these informations are correct, the problem either comes from the mainboard or from the PSU. Does this conclusion make sense, or am I completly wrong ? Is there another possibility, which I haven't thought of yet ?

 

[flashgeist]

UPDATE : I don't think I can rely on MBM's voltage measures, because my motherboard isn't officially supported by this program. I tried to change the sensor settings and I got other values than the ones I just posted.

So I don't know what to think...

 

[howard_hopkinso]

Does this conclusion make sense, or am I completly wrong ?

Makes perfect sense to me.

I suspect your psu may be faulty.

Regards Howard

 

[howard_hopkinso]

Download the free Everest programme from HERE

When you run the programme click on the large computer icon in the righthand pane, and then the sensor icon.

It will list your temps, and your voltages.

Regards Howard

 

[flashgeist]

this Everest programme is really great, thanks for the tip !

It displays more accurate informations than MBM, there's just one cpu core displayed and it's showing 30°C (86°F).

As far as voltages are concerned, I've got the following :

CPU core : 1.39 V
+2.5 V : 1.50 V
+3.3 V : 2.50 V
+5 V : 5.08 V
+12 V : constantly varying (like in MBM), from about 6 V to 12.5 V
+5 V standby : 4.92 V

So there seems to be really something wrong with the voltages. I will try to assess this with a voltmeter (or is it unsafe to do this ?)

 

[howard_hopkinso]

I will try to assess this with a voltmeter (or is it unsafe to do this ?)

If you know what you`re doing with a volt metre, there shouldn`t be a problem.

Maybe a new psu is on the cards?

Regards Howard

 

[flashgeist]

I tried with 2 different PSU, it's worse: It freezes on the mainboard start screen (just after the memory check). It's not even possible to go into the BIOS setup.

So I guess it's either mainboard or cpu which is faulty.

 

[howard_hopkinso]

Mmm I don`t know why trying a different psu should make things worse.

What were the psu`s you tried?

Check your mobo`s capacitors. Look for bulging, or leaking caps. See HERE for further info.

Regards Howard

 

[flashgeist]

Problem (almost) solved :

I must have done something wrong at the first place when trying the other PSU... I tried again and it worked. But the BSOD problem was still here.

So I thought I try testing the RAM again. And this time I used Memtest86 instead of GoldMemory, and it made the difference: 660 errors have been detected on the first pass.

You were right from the start, it was a RAM problem. :grinthumb

The problem concerning the wrong voltages can come from a faulty sensor or from the monitoring software itself.

So I changed the RAM for a new one, but I still get errors in Memtest86 (just 10 errors this time). Maybe I've got very bad luck with RAM sticks, or can it be the mainboard which has problems to handle the RAM ?

I found a way to make the new RAM work by lowering the CAS latency and other timing settings. The recommended settings were 3.0-3-3-8 @ 400Mhz and I changed them in the bios to 3.0-4-4-12. Since then I haven't had any error in Memtest86 and so far no BSOD either.

Still, I would find it nicer if I could use the RAM to its full potential... I guess I'll have to go back to the shop and ask once more for a new RAM stick to try.

 

[flashgeist]

Problem REALLY solved : RAM wasn't faulty.

I did something I should have tried at the beginning : I swapped the RAM from slot 1 to slot 2. No more errors, even with the timings 3.0-3-3-8 @ 400Mhz.

I then brought the mainboard back to the shop, it's been tested and it was faulty. They didn't have any PC CHIPS M871G any more (that was the name of this cheap mobo), so they gave me an ASUS K8V-MX. It was still very cheap (70 $), and the performance (especially display - onboard VGA) is much better than the PC CHIPS.

No more RAM errors, no more BSODs :grinthumb

 

[howard_hopkinso]

Glad your problem is solved.

Thanks for letting us know.

Regards Howard