Vcore drooping is by design...

[Falkentyne]

After looking at this webpage
http://www.elecdesign.com/Articles/Print.cfm?ArticleID=7648
(given to me graciously in the P4 EE thread), it seems that boards that droop under load (as much as .07v) is actually based on intel's own specifications for the voltage regulator. That is, if I understand this correctly, and the drooping doesn't actually occur deep inside the processor ?_?.

Some of this stuff is over my head.....
But if a cpu is set for 1.55, gives 1.52 idle, and 1.47 load, that's completely normal. It's supposed to do that, to cut down on heat and power. (of course, it becomes a problem when you overclock....but we all know Intel didn't design the chips for overclocking, but for long term reliability.......).

Something to keep in mind.....
(But boards that droop from 1.55 to 1.35 are a serious problem...).

[woutertal]

After looking at this webpage
http://www.elecdesign.com/Articles/Print.cfm?ArticleID=7648
(given to me graciously in the P4 EE thread), it seems that boards that droop under load (as much as .07v) is actually based on intel's own specifications for the voltage regulator. That is, if I understand this correctly, and the drooping doesn't actually occur deep inside the processor ?_?.

Some of this stuff is over my head.....
But if a cpu is set for 1.55, gives 1.52 idle, and 1.47 load, that's completely normal. It's supposed to do that, to cut down on heat and power. (of course, it becomes a problem when you overclock....but we all know Intel didn't design the chips for overclocking, but for long term reliability.......).

Something to keep in mind.....
(But boards that droop from 1.55 to 1.35 are a serious problem...).


Hey tks for posting this. It makes things more understandable. I always wondered how they could design such bad circuitry; now I know it's done on purpose. Definitely something to consider before doing a droop mod.
:-/

[jenko]

Its the VID right? i believe you can adjust in the bios i have a value of 62, im no expert on this though.

[woutertal]

Its the VID right? i believe you can adjust in the bios i have a value of 62, im no expert on this though.

What's 'VID' ?

[brywalker]

Thats all well and good.

But when you have VCORE set at 1.7 and it bounces between 1.73 and 1.68, then that is a problem in my book.

[Stratcat]

What's 'VID' ?
VID = Voltage ID.

Your mobo socket detects a pattern in your cpu's pins, as to whether they're tied 'hi', 'low' 'shorted', etc. Based on this pattern, your mobo will 'ID' the correct Vcore to be supplied to the currently installed cpu.

Some ppl w/duallies or other mobo's that don't have much in the way of Vcore adjustments in thier BIOS, can alter thier cpu's pins to change thier mobo's detection of the installed cpu's VID, and therefore change (raise) the Vcore for overclocking.

Strat

[woutertal]

VID = Voltage ID.

Your mobo socket detects a pattern in your cpu's pins, as to whether they're tied 'hi', 'low' 'shorted', etc. Based on this pattern, your mobo will 'ID' the correct Vcore to be supplied to the currently installed cpu.

Some ppl w/duallies or other mobo's that don't have much in the way of Vcore adjustments in thier BIOS, can alter thier cpu's pins to change thier mobo's detection of the installed cpu's VID, and therefore change (raise) the Vcore for overclocking.

Strat

Thanks strat. I got it.

[larva]

Thats all well and good.

But when you have VCORE set at 1.7 and it bounces between 1.73 and 1.68, then that is a problem in my book.
Time to read a few more books then.

I have preached for years (to basically no avail) that seizing upon the rudimentary numbers available via the motherboard sensors is the worst kind of pseudo-knowledge and is nothing to be acted upon. Obviously we needn't know anything about power supplies, power supply regulation, operational semiconductor voltage tolerance, measurement accuracy, or anything else. All we have to know is that we set 1.5V so if it doesn't read 1.5V, and exactly 1.5v the world will end. Or at least that must be why I'd like my cpu to run 3512MHz but it will only run 3500...

Or perhaps not.

The truth is the numbers made available via the motherboard sensors are the crudest form of dog-and-pony show imaginable. The aren't accurate or precise enough to draw the types of conclusions often drawn from their reading, and even if they were often the conclusions don't follow anything more substantial than wishfull thinking.

The sampling frequency of the Vcore monitoring is not uniform between the various motherboard manufacturers. The exact sampling frequency along with the dampening designed into the circuit will far more likely determine the exact stability of the reading, but not necessarily reflect the quality of the votage regulation of the supply being monitored. Get out an oscilloscope and look at the actual truth of the matter before you leap to the conclusion that the very practiced power supply engineers at Asus (or most anywhere else) have left stones unturned for you to flip over.

[Falkentyne]

Time to read a few more books then.

I have preached for years (to basically no avail) that seizing upon the rudimentary numbers available via the motherboard sensors is the worst kind of pseudo-knowledge and is nothing to be acted upon. Obviously we needn't know anything about power supplies, power supply regulation, operational semiconductor voltage tolerance, measurement accuracy, or anything else. All we have to know is that we set 1.5V so if it doesn't read 1.5V, and exactly 1.5v the world will end. Or at least that must be why I'd like my cpu to run 3512MHz but it will only run 3500...

Or perhaps not.

The truth is the numbers made available via the motherboard sensors are the crudest form of dog-and-pony show imaginable. The aren't accurate or precise enough to draw the types of conclusions often drawn from their reading, and even if they were often the conclusions don't follow anything more substantial than wishfull thinking.

The sampling frequency of the Vcore monitoring is not uniform between the various motherboard manufacturers. The exact sampling frequency along with the dampening designed into the circuit will far more likely determine the exact stability of the reading, but not necessarily reflect the quality of the votage regulation of the supply being monitored. Get out an oscilloscope and look at the actual truth of the matter before you leap to the conclusion that the very practiced power supply engineers at Asus (or most anywhere else) have left stones unturned for you to flip over.


PREACH IT, Larva !!!!!!