Voltage & temp questions

[wiskeychris]

I recently upgraded my system to what you see in my sig. I'm running it off of an Antec 400W PSU. Here is my voltage levels:

+12V = 11.84
+5V = 4.945
+3.3V = 30248

My CPU is running at 32/89 C/F
MB is at 33/91
Vid card is Idle at 42.

Does this all sound about right or do I need to consider getting a new PSU? I'm asking because my system has shut down unexpectedly twice in the last hour.

 

[Ice Czar]

well from your temps, you either:

A. have the mobo's SYS sensor in a hotspot

or

B. have a problem with the CPU temp accuracy

as far as those voltages go they are in spec except the +3.3V

however, cut and paste

closer to the reference value is better than either higher or lower
stuff also doesnt tend to get fried from lower values as much as higher values
anything higher than the 5% range is called an overshoot and that fries most stuff
anything lower is called an undershoot and that makes things unstable


for reference Basic Spec compliance of 5% is
+12V....11.4V to 12.6V
+5V......4.75V to 5.25V
+3.3V...3.135V to 3.465V
but its also a dynamic range
the supply converts power from AC to DC as the load changes from the components
well it also has to keep the reference voltages within the range above while it doing that
so measuring at one point is like watching a single frame in a movie
so logging the voltage is a good idea

that said your BIOS and software is lying to you
what you dont know is by what amount
for that youd need to measure it with a multimeter
cheap ones can be had for $20 or $30 though their accuracy isnt all that much better than the BIOS
they scale up to hundreds of dollars and increasing accuracy

ideally you calibrate your software to a hard measurement (like at idle)
then log till your confident that your within spec at all times

a high reference voltage at idle means that while the supply might be regulating the load within its own parameter,
under some circumstances it could be letting overshoots through
now there are advantages to having a higher reference value if your overclocking
but only if your not allowing overshoots
thats one of the reasons overclockers like supplies with tighter than spec load regulation
so instead of a supply with a range of 5% regulation it has 3% or 1%
and can thus have a higher reference value without letting through overshoots

most components can deal with higher than 5% values and are commonly tested to survive split second 10% values, but these wear on the components

since IC Chips (integrated circuits) vary from pristine to just barely passing the functional test, that means a little "wear" (electromigration) could cause immediate failure or not be noticable degredation for years

more to the point though, the odds are greatly increased for some failure somewhere by the sheer number of IC chips throughout the computer as a whole.
some are behind buffers like the CPU and RAM which have the voltage modified by the motherboards voltage regulation modual, while other lack any buffer at all, like the drives

its best to test them under a real load so in place powering your computer


a rail that doesnt have a load on it will likely be fairly off
which is why you can ignore the negative rails as they arent typically employed anymore
(unless your running some antique cards)

 

[wiskeychris]

Ok, Thanks. What should my MB and CPU temps be? Are they too hot?