• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Wild VCore fluctuations...Help

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Super Nade

† SU(3) Moderator  †
Joined
Aug 30, 2004
Location
Santa Barbara, CA
Hello all,

I'm seeing my VCore fluctuate wildly between [1.648, 1.7] V when I play Doom 3. I've set VCore to be 1.675V in the BIOS. Could this be the PSU? My rail voltages are fairly constant (software monitoring) and the 12 V line varies by a maximum of 7%. Here's the speedfan SS. My tempartures are Idle 38C/ Full Load 52 C.

My questions are:-
1. Is this range of fluctuation acceptable?
2. Any factor other than the PSU that I've missed out?
 

Attachments

  • speedfan_fluctuations.JPG
    speedfan_fluctuations.JPG
    34 KB · Views: 787
This looks even more scary. Look at a spike shooting well above 1.70 V!!!!
What is going on!!!!!
 

Attachments

  • speedfan_fluctuations_2.JPG
    speedfan_fluctuations_2.JPG
    35.6 KB · Views: 786
Most of my MOSFETS are sinked with big Copper RAM sinks. Not all of them are sinked though. The RAM Sinks wouldn't fit on the ones very close to the CPU.
 
Well i know this happens with everyone on the Epox board's for S754. When you set the vcore to say 1.65v, during IDLE it stays thier but during Load it goes up to like 1.70-1.75v.

I havent heard this causing any problems i just ignore it
 
CandymanCan said:
Well i know this happens with everyone on the Epox board's for S754. When you set the vcore to say 1.65v, during IDLE it stays thier but during Load it goes up to like 1.70-1.75v.

I havent heard this causing any problems i just ignore it
There is a somewhat fix for this over at aoaforums.com. I know you have posted there but have you tried the modded bios with the overwriting of the bootblock? http://www.aoaforums.com/forum/showthread.php?t=26658 Its in about the middle of the page and it has really helped me stabilize my oc. I dont get nearly as bad overvolting as I used to during load and my temps have dropped over 5C during load.
 
What is the time scale of that voltage vs time chart shown?

Digital volt-meters (DVM) or digital multi-meter (DMM) are measuring the DC voltage component, which is by definition no fluctuation and varying very "slowly" (determined by the time constant of the specificiation). DVM/DMM won't be able to capture the sudden change in CPU voltage and current surge which can occur in micro-second (or even less) due to load and task changes.

Those software monitoring programs may not be as accurate (in absolutely %-wise) as a DVM/DMM in absolute value, but they are able to show (visually and comparatively) how voltage varies with time up to the time resolution allowed, e.g. voltage fluctuation of different PSUs, different CPU load conditions.

The best way is to use a high speed sampling oscilloscope which can show accurately how voltage and current change in a short period of time during load switching of a CPU. These are usually done in labs and in a controlled environment.
 
Last edited:
If you don't have a voltmeter......

Remember that fluctations are usually caused by the motherboard. Some are known to overvolt, whereas others have a tendancy to undervolt excessively. So there is a good chance that your Asus is the culprit. However, you might benefit from a more powerful psu such as an OCZ, which is sure to provide you will a more stable 12+ volt. Nonetheless, if your computer is stable then you have nothing to worry about. But if you really want to fix this dilemma then chances are you're probably going to have to replace something.

deception``
 
My fortron is a 530W PSU. Why would I need something more powerful?

And I think you can obseve spikes on a DVM if its set to detect AC and if the spikes are periodic with a wide time scale spacing. Ofcourse as hitechjb1 said, an oscilloscope is the best way to observe fluctuations. I'm gonna try my backup stock PSU and see if this is a problem with the PSU or the Mobo.

Strangely I have not observed more than 2 spikes !
 
Ohho..more than two spikes almost 1.75V! I'm taking out this PSU.
 

Attachments

  • speedfan_fluctuations_4_Nov23.JPG
    speedfan_fluctuations_4_Nov23.JPG
    36.2 KB · Views: 749
As your 12v line is varying by 7% if you consider your vcore is varying from approx 1.64 to 1.74V this is also about 7% so it is not really surprising if the figures are accurate. However if it did vary by 7% (as indicated in your first post) I would be more worried about that as 7% equates to approx 0.8v on the 12v rail or a variance of say between 11.6 and 12.4 - altogether more worrying if you ask me - you sure about the 7% variance?
 
A64skate said:
There is a somewhat fix for this over at aoaforums.com. I know you have posted there but have you tried the modded bios with the overwriting of the bootblock? http://www.aoaforums.com/forum/showthread.php?t=26658 Its in about the middle of the page and it has really helped me stabilize my oc. I dont get nearly as bad overvolting as I used to during load and my temps have dropped over 5C during load.


Yea but it undervolts when i change the multi i already tried this method thanks a bunch tho for trying to help :thup:
 
Yes I have never been at 12.0V. Always less than 12.0 V about 11.2V. I'm gonna RMA this PSU ASAP. What do you guys say?
 

Attachments

  • speedfan_fluctuations_Nov24.JPG
    speedfan_fluctuations_Nov24.JPG
    58.9 KB · Views: 729
How do I measure the voltages? Do I connect the leads to an unsued line or do I need to read off something from the Mobo?
 
Super Nade said:
How do I measure the voltages? Do I connect the leads to an unsued line or do I need to read off something from the Mobo?

To measure the 12 V of the PSU, you may just measure the voltage between a 12 V wire and ground, preferably from a wire chain of all usused molex connectors. The red probe of the DVM/DMM to the 12 V and the black probe to ground.

Using a wire chain of all unused molex connectors is to avoid any potential voltage drop due to current. If you cannot find a free unused molex chain, and you don't want to rewire the molex connectors, it may be ok too for initial measurement. Take precaution not to short any of the voltage to any wire or ground which may damagel the PSU.

Same for the 3.3 V and 5 V.

Ground can be just the metal case or the black wire of a molex.


Measuring directly the Vcore from the motherboard is a bit trickier and more prone for shorting and killing a motherboard due to close proximity of connections, mistake and accident.
 
Make sure when you stick the Red wire on the Meter in any line you wanna measure like the Yellow for the 12v. That you stick the black wire in a ground like the black wires on the molex like said above. Also you should stick them in the molex when the PC is off. Doing it while the pc is on can short it out.
 
Back