Voltage Oddity?

Kazeko

So, the Original Voltage for my Q6600, is SUPPOSED to be 1.15v.

That's all well and good, but, for whatever reason, it's under volted. It's been running at 1.10v, despite it's forced setting at 1.15 in the bios. Mainboard still curbs off about .05v. So I turn it up a little bit, and see where it sits at 1.175v - 1.12v. Turn it up to 1.20 - 1.16v.

Forced settings are confirmed by value listings (shows current value) in the bios, and by CPU-Z.

It's really pretty strange. Should I be concerned, or just compensate for the odd values in the bios? (Bios updates checked)

The really funny part, is that it's happy and stable (8+ hours on prime95) at 2.4ghz - 2.7ghz at 1.10v.

Zelig

Software is pretty useless for reporting voltage. If you get a multimeter and test the voltage, it will probably be pretty close to whatever you set in the BIOS.

ChaosMinionX

Its whats called "Voltage Droop, or Vdroop" it occurs under load on all boards to some extent, it is nothing to be worried about.

Zelig

Doing a quick google of vdroop, I don't think it's the same thing. I've got my voltage set to 1.35 in the BIOS, it detects as 1.312, but my multimeter reads 1.351 or so.

H3X4D3C1M4L

It isn't the same thing. Software sensors just aren't that reliable.

A board with a good voltage regulation design shouldn't drop period on load. Maybe it'll drop a little but it shouldn't drop as much as that.

procupine14

yeah from my recent personal experience.....software isn't really good for monitoring anything......kinda makes you wonder why we all even bother

ChaosMinionX

Well software isnt, but the same theory applies to what its reading... Might not be good that excessively, but yes it can occur that way on some mobos with extreme overclocks and poor voltage regs.

Kazeko

Link for a quick guide on measuring voltage with my fancy-shmancy new multimeter I got last x-mas?

The only thing I can come up with is measuring on the back of the mainboard somewhere, but I'll be damned if I know where?

The_Neon_Cowboy

1) multimeters are carefully calabrated
2) voltage chips on mothers boards are not
thus a mobo sensor is never more then "a guide"
becouse it'll never be truely accurate

Kazeko

Right-o... But where would I go about finding how to measure the voltage with a multimeter?

Lot of things I do understand, lot of things I don't. : /

ChaosMinionX

Just google MM points on various voltages your looking for on that particular board, I couldnt tell you as I have never used it.

Kazeko

Well, apparently it's quite the problem on this board, and it is fairly accurate. Folks have even gone to using graphite (Pencil Lead!) to bypass a resistor for more stable voltages.

Apparently the low idle and low loaded voltages are standard fare on this mainboard.


Fortunately, I'm one to ask questions like "Why didn't nVidia's engineers just reduce or eliminate that resistor in their reference design, if this was a problem?"

Likely answer: While voltage doesn't droop as much, the spikes will be far more drastic, IMO. Going to stay away from this one. I'll deal with the Vdroop.

From other forums, for others running 680i Mainboards. Apparently it's a semi-big problem across other mainboards on the nForce680i chipset as well. Hope someone else finds this helpful.

Pencil Mod, and References
Stopping VDROOP dead in it's tracks!

680i (eVGA, XFX, BFG...) A1 / D00 Rev. VCORE AND VDROP DONE! - XtremeSystems Forums

Edit: On the upside of things, the thermals are the only thing stopping me from overclocking this cpu further than I already have, and I found the best place to test the core voltage.

Also, I think with the right OC, I can break a 15k 3dm06 score... for what it's worth.

H3X4D3C1M4L

Isn't the resistor there for a specific reason.... using graphite to short it doesn't seem like such a great idea

ChaosMinionX

Its used generally for extreme benching to eliminate vdroop under high loads.

H3X4D3C1M4L

So it should be left alone then....amiright?

ChaosMinionX

Vdroop is actually an integral part of motherboards today.

Source: Intel Processor Power Delivery Design Guidelines and Specifications: Vdroop Explained - The Tech Repository Forums

Very good reading.

H3X4D3C1M4L

I'll give it a read

So what I was saying was, should the transistor not be left alone

ChaosMinionX

Normal Users - Leave it alone

Extreme OCers - have at it

People with money in their pocket - have at it

Kazeko

I'll be letting this one alone...

But I gather that this mainboard wasn't necessarily the best choice in the world.

Edit: An X38 northbridge may have been a better option, but no SLI, and... blah. Perhaps I should be happy with 3.5ghz on air.