Sabertooth 990FX/AMD965BE voltage flux and wrong reading in cpuz

[chartiet]

I have a Sabertooth 990FX and a AMD 965BE and I cant get the voltage that's set in BIOS to match up with cpuz monitor in windows. It reads about 0.012 different and jumps from around within that 0.012v. I have it set at 1.41875v and in bios ez mode screen, it reads 1.428/1.440v, in cpuz, it reads/jumps 1.428/1.440 and 1.452v (under load). I have been trying all the auto setting adjustment in digi + power control. any ideas? asus support says its a bad cpu voltage sensor regulator, but others are having similar issues as well. Thank You. Please let me know what other info you need or questions you have.

 

[chartiet]

I cant pass stability test (Prime95) with cpu voltage as high as 1.43125v! Any suggestions?

 

[raja@asus]

The voltage polling is never accurate. There are multiple reasons for that:

1) It depends where along the power plane the via for the super IO is placed.

2) Real time VCC and the applied VID are two different things. Too complex to explain here, but there is offset between the two. Realtime overshoot cannot be measured with a DMM or the super IO - they don't sample fast enough.

3) If any other software polling tools are running with CPU-Z, the reading can fluctuate erroneously.

4) The voltage will modulate somewhat according to the load, this is simple ohm's law (V=IR)


-Raja

 

[chartiet]

ram gets through memtest at current settings fine, but at a high cpu voltage and digi+ settings at where they're at, a worker will fail in Prime95. temps are fine. its a cpu instability issue i presume.

Its probably a setting on auto or similar that is causing issues?

CPU: AMD Phenom II X4 965BE 4.01GHz (200x20@1.43125v)
MoBo: ASUS Sabertooth 990FX R2.0 (CPU/NB 2600MHz)
RAM: G.Skill 8GB (4x2GB) DDR3 1600 8-8-8-24-40-1T @ 1.645v
GPU: MSi Twin Frozr Radeon HD 7850 2GB
HSF: Corsair H80i | PSU: Corsair HX750W
HD: WD Black 1TB 7200RPM 64MB SATA 6.0Gb/s
OS: Windows 7 64-bit | Case: CM HAF 932

 

[raja@asus]

You may just have a poor clocking CPU - these Phenoms were very hit and miss.

To debug you can try lowering CPU-NB frequency, and see if the CPU becomes more stable. Also check load temps of the CPU. Anything past 60C is prone to failing heavy loads on these CPUs.