I realize HWMonitor sucks for reading voltages, but I don't understand why it's suddenly reading lower voltages than it did 6+ months ago. When I first used HWMonitor, I'm positive it was showing higher voltages, nearly identical to what the BIOS displays. Here's what the BIOS showed a few minutes ago:
12.3V in bios is a bit towards the high side but still acceptable, i would always trust BIOS readings over software readings. The voltage measuring chips are not consistent between boards and its not well documented what each board uses so the software tends to take a guess and hope to come out close.
To make sure you are accurate use a digital multimeter, i would probe a molex connector, yellow to black will give you 12V.
The voltages displayed in the BIOS/CMOS Setup Utility is informational only. It was never meant to be adjustable so there's really no way to calibrate the BIOS voltages displayed to match those that your multimeter shows.
Just as info, there's a fan inside the tower (probably the CPU fan) that likes to throttle up and down -- just enough to annoy anybody using the computer. It does this for about 5 seconds at a time.
Per HW Monitor over the last few hours, the CPU temps have fluctuated 38-46C for core 0, and 39-49C for core 1. They both average about 41-43C. The fan throttling up and down has no correlation to CPU temps or workload. Sometimes happens when the computer is cold, other times when warm. AMD Cool-n-Quiet is disabled in the BIOS.
Also, is it normal for the PSU case to be noticeably warmer/cooler to the touch? Sometimes it's just barely luke warm, other times it's as hot as the motherboard heat pipe.
The noise could be from a case fan. If it has the cheap case fans that use brass sleeve bearings, the lubricant has probably all evaporated. Try finding the offending fan, clean it and lubricate the bearing.
When was the last time you blew the dust out of the case?
Need to take two readings on PSU. First one @ idle, The 2nd one with a Heavy load on the +12 V. A drop of more than 0.4 V indicates poor regulation and/ or dieing PSU. While the low spec for the +12 V is 11.4 V, I normally ditch the PSU if it dropps below 11.6 V. I normally run Furmark (in a window) or Prime 95 to load the +12 V. For High end GPUs furmark will load +12 V more than Prime 95.
Voltage noted in bios are reasonably accurate, But Very low load on +12 V.
Normally you can tell if HDMonitor is not correct because the readings will be considerably off. If you have not changed the version of HWMonitor and it was correct, then dropped that would indicate a drop in the real voltage. I normally verify the displayed reading with a Voltmeter (you indicated you measured it with a DVM @ molex, Do this while running Prime 95 and HDMonitor, complare the readings.
Just something worth noting people, I just ran HW Monitor and it said my 12v rail was under 11v, 10.98 according to HW. In Everest it was 12.10, I'm pretty sure HW is wrong because my system would be down at that voltage.
Yes, Was that under load, or idle.
CPU might function as this goes thru a VR(s) that knock it down to vcore BUT would stress the VRs, GPU is questionable along with HDDs. While the spec is For the +12 v is +/- 5 %, many circuits will operate with a +/- 10 % (the extra 5 % is a margin - But that does depend on the circuit and other factors).
Would verify with a voltmeter (Orange wire to black on molex).
Added +12 V as this is not true for the +5 V if it is used to power TTL logic circuits which is set to 4.75->5.25V, However cmos circuits may work at 4.5 V.
I swapped out the existing power supply (Antec TruePower Trio 550W) with a brand new PC Power & Cooling Silencer 370W. I had purchased the latter for a HTPC and forgot I had it.
Almost everything is the same, except the random reboots are gone. That's good news. HWMonitor is reporting virtually the same (low) voltages, so there's proof the software sucks.
I'm scratching my head about a new problem though. When the computer is under a load, and I attempt to use Google Earth, it is substantially more sluggish than before. I'll have Thunderbird open, sometimes 10-15 instances of Firefox, several folders, Adobe Photoshop, Adobe InDesign, open simultaneously. With the old PSU I could run Google Earth with no drop in performance. Now, I need to close programs or folders out before it behaves properly again.
It's worth noting that OTHER programs are not affected when Google Earth is run under these circumstances.