I was playing Assassin's Creed last night on my Q6600, GF 8800GTX, P5N32 SLI mobo & ST75F PSU.
However, @ ~3am (what can I say, it's an addictive game I received a blue screen of death, which stayed up for only a few milliseconds before the system rebooted.
I thought nothing of it at the time, as it was late, and the game is reknowned for being abit unstable.
However, I came to boot up my PC this-morning & was halted on the POST screen with an error warning me that the motherboard has detected a temp/fan/voltage fault.
The BIOS is reporting 4 of the voltages as bad :-
Vcore - 4.16V (not sure what this relates to, or what it should be?)
3.3V - 4.08V
5V - 6.85V
12V - 16.32V
I've measured the 5V & 12V lines on a molex, they're coming back as 4.2V & 10.8V - though it's an analogue mini-tester, so am not particularly confident of the accuracy. (though 10.8V is far enough from 16.32V for me to question the accuracy of the mobo's value...)
Any suggestions what I should do next to isolate whether it's the mobo, PSU or something else at fault?
Incidentally, the PC boots fine if I ignore the faults, though I havn't tried putting it under any load yet, nor am I keen to do so yet.
More about :bsd crash mobo reporting overvolt fault
December 4, 2008 10:43:46 AM
Just to clarify, while the mobo is SLI - I havn't yet put in a 2nd 8800GTX in it. (I was going to get myself one this xmas!)
My first thoughts are that if it is pushing those voltages through, then your CPU is almost definitely dead, however I doubt it is. Try clearing the CMOS or seeing if your BIOS has a windows-style event logger you can clear.
Although Mi1ez suggestion was as good call, It might be premature to say "merry Xmas"
What are the votages being reported by your bios now. Your inital readings were wacked as the system would not have even let you in to read, much less boot if they were as reported which is why mi1ez call was geat. BUT why was the bios corrupted, may have been the crash caused by either the game, OR an out of tolerence voltage from the PSU.
I'm alittle concerned over the meter readings, And yes your anolog Voltmeter might also be in error. You can check your meter by measuring 4 1.5 volt batteries placed in series (ie pick up a $5 4 cell battery holder from Radio shack and 4 NEW 1.5V batteries). Your meter should read 6.2 volts (1.55 x 4) This was an old standard for calibrating anolog voltmeters. You can also measure a 12V lead acid battery as it should read 12.6 V (Note do not measure a 12V car battery right after charging as it will be closer to 13 -> 13.6 V).
After you have verified your meter, then measure your +5 and +12V UNDER LOAD at a molex pin. The 5V MUST be greater than 4.75V (Min level for TTL logic) and 11.5V (some say 11.3V) for the +12V.
You can load the +12V by running ATITools 3D View.
December 4, 2008 2:51:43 PM
Thanks for the replies, i'll investigate further.
Incidentally, now that the bios isn't reporting garbage values, SpeedFan shows these voltage values when idle.
I've got no idea what most of them relate to, so can't tell if any are erroneous.
First - Ignore the +3.3 = 0, -12V = 11.84 and the -5V = -8.78. The negative 5 and 12 are hosed on most software reporting programs and I believe are not even used on newer systems. The +3.3 V should read properly, but speedfan is not.
Your +5 = 5.05 is great, the +12 = 11.84 is OK AS LONG as it does not drop below 11.5V UNDER Full load.
Recommend you down load CPUID Hdmonitor and monitor Min voltage for +12V under Full Load. www.cpuid.com/hwmonitor.php
I use ATITools' 3DView to load the +12V (Note - Does not work for setting GPU clock/fan speed for ATI 48x0 cards - But the 3DView works on my 4870) www.techpowerup.com/downloads/436/ATITool_0.26.html
I may be using the beta version 0.27. Added - But you can use any app that will fully load your GPU.