Are my PSU rails messed up?

[fuzzybabybunny]

I have a Fortron 700W PSU and I get some major computer problems when the weather heats up a little bit.

When the weather heats up, sometimes the system will simply shut off by itself. Then when I reboot, sometimes it will hang at a black screen even before POST. After a few more restarts it'll get into Windows, but then do a hard lock after a few minutes, meaning the keyboard doesn't work, the mouse doesn't respond, etc.

I'm running watercooling and none of the components in my system are getting above 40C, including video card, CPU, Mobo, and HDDs. I don't know about the RAM or the PSU because these don't have temperature sensors that I can read via software.

My PSU rails are thus, measured with SpeedFan:

+3.3V: 0V - what does this rail do?
+5V: +4.87V - looks good
+5V: +5.08V - looks good
+12V: +12.16V - looks good
-12V: -16.97V - uhhhhh
-5V: -8.78V - uhhhh

What do the negative voltages do?

Also something confusing. I'm running a dual core Pentium D 805 overclocked to 3.66GHz on both cores (from 2.66GHz), but the Vcores don't match.

Vcore1: 1.54V
Vcore2: 3.26V

Anything look wrong?

 

[pkrush]

I think it's got the sensors arranged wrong, what you have for Vcore2 should be for the 3.3 volt rail. I think that you're suffering from some temperature related problems because PSU problems tend to result in consistently random hangs that go away after the machine's been rebooted, not after it's had a chance to cool down. Try measuring various spots on your motherboard with an IR thermometer. My guess is that it's the motherboard power regulators, but I can't guarantee that's what it is.

 

[Aluvus]

You need to measure voltages with a real digital multimeter, software readings are routinely inaccurate.

The negative rails are mainly used for very old hardware (some ISA cards, for instance) and in a modern system are usually of no real importance.