alot of people says the issue with certain asus board.. in particular the VDroop prevents successfull overclocking.. is that so?

and according to asus website..

Source : http://support.asus.com/faq/faq.aspx?no=BB...SLanguage=en-us

can some1 clarify?

i think almost all of intel board experience tis Vdrop, it is absolutely normal
u can mod the board to prevent any Vdrop

Its a drop of CPU voltage when the CPU is taking a lot of power. Happens when the voltage regulator on the motherboard is not well designed. Usually for socket 775 processors a 4 phase or better power regulator is recommended.

For boards suffering from vdroop, sometimes overclockers voltmod them for higher voltage to compensate for the voltage drop. Setting a higher voltage in bios is also possible, but often results are unsatisfactory. But increasing the voltage has its problems as well, as when the processor is idle, the voltage will be quite high (no voltage drop), which may damage the processor.

---

Added on June 24, 2007, 4:17 pmAnd yes, it also affects other boards, not just Asus. Asus is often mentioned because they used to often use really small mosfets in the voltage regulator which were prone to overheating (which causes mosfets to conduct less current)

This post has been edited by lohwenli: Jun 24 2007, 04:17 PM

yeah... it's normal... tat's y ppl want more powerful PSU ma... if nt Enermax, Tagan and other lords of PSU will eat grass to survive lo...

i don't think its becoz of PSU.. its got to do with the voltage regulators on the motherboard

even a powerful PSU cannot prevent the Vdrop

another myth...

Vdroop is a design, read through this article if youre interested:Vdroop

Most Asus board (C2D) have terrible vdroop, even for asus commando iirc, when cpu is loaded, Vcore will drop because the power circuit design, thus lowering the Vcore than specified in bios, hence, cpu not getting enough juice and restarts. So there be pencil mods for vdroop, for extreme overclockers.

HTH

This post has been edited by anangryorc: Jun 26 2007, 12:32 AM

Not a chance. Even with a PSU with adjustable rails there will still be vdroop if the mobo voltage regulator is not well designed.


Cool article. Downside of pencil mods and other hard mods is they don't solve the vdroop problem entirely or introduce problems of their own. But anyway only very hardcore overclockers will care...

HAH!!! Of all the makers to publish something concerning VDroop, it had to be Asus. Brrrrrrr........You should measure the VDroop with a DMM for the P5B series, it'll give you the shivers

And seriously, what is written in that article is rubbishing. If you really want to know about VDroop, read this instead: http://www.thetechrepository.com/showthread.php?t=126

No doubt it's more technical but it explains the real truth

This post has been edited by almostthere: Jun 26 2007, 06:47 AM

^^

ok.. i've browse through the tech repository.. and i dun understand it.

so this vdroop is part of the design spec? and not a dreaded problem?

This one,i think is something related to P=VI where your P is at a certain valur while your V is dropping,I will increase the value to compensate the loss of V. THus the compensation make the mosfet area become hotter.

It IS a problem, and has to be taken into account during design, as its pretty much unavoidable. Although its important to minimise it, its not cost effective to use a very complex design, so manufacturers make do with simpler, cheaper designs that work and compensate by adjusting the voltage to a value where vdroop is still acceptable. The problem starts during overclocking, as higher power consumption in overclocking will mean more current, which will cause the vdroop to be more severe. There will be minimal vdroop when the processor is at idle, but severe vdroop when the processor is fully loaded, which is bad, as overclockers may increase voltage to high levels to compensate for vdroop, and the absence of vdroop in idle will mean the voltage may be dangerously high (processor damage possible).

You better double check that, the formula should be V=IR, where V is the resulting voltage drop, I is the current taken by the processor, and R is the equivalent series resistance in the voltage regulator circuit from the sensing point to the processor.

P=VI is to calculate power usage, specifically conversion from electricity to heat (which applies in almost all electrical devices). Although your description does explain a downside of vdroop, it does not explain accurately why it happens. Yes, a poorly designed board suffering severe vdroop will actually make the problem of overheating MOSFETs worse and worse, as the voltage drop (using V=IR) across overheated MOSFETs will result in a power loss and they will overheat more. And the more the MOSFETs are overheated the worse the voltage drop-a vicious cycle. Contrary to your explaination, current (I) can only be increase up to a certain point until the MOSFETs are working at full capacity-which would be limited by overheating (a safety feature of MOSFETs).

This post has been edited by lohwenli: Jun 26 2007, 10:15 AM

interesting explanation... so how do we know the MOSFETs are working at full capacity...?

Vdroop happens when CPU draw current from the CPU Voltage regulator. At certain times CPU can draw as much as 100A. So Vdroop is unavoidable. There is a vdroop spec and most board designers will design so the droop will not violate the Vdroop specs.


For this Vdroop problem, there are a few ways to improve it and it requires you to rework (change components) the board. It will void your warrenty and so try it at you own risk.

1. Put higher capacitance caps or better quality capacitors around the cpu socket.
2. Fill in the empty capacitor slots around the CPU with simillar capacitors.
3. Do some compensation tuning to the cpu voltage regulator. You will need to know the voltage regulator very well and expert in control theory. Not recommended.

There's no sure way of finding out unless you are very familiar with the voltage regulator circuit of that particular motherboard. At best, you can guess the peak current of the mosfets during the max duty cycle of the voltage regulator by estimating from the power usage of the processor. If the number you come up with is higher than the rated current for the mosfets, then the mosfets are overloaded. However, finding out the exact current used by the processor is a major problem, even the best estimates can be off by quite a bit unless you have proper equipment and are prepared to hard-mod the motherboard.


Sucess rate of 1 & 2 highly depends on how well the voltage regulator circuit was designed in the first place. If the manufacturers decided to cut costs by using fewer or lower quality caps, then you'll improve on it. However, if the circuit was already well-designed to start with, you may just screw everything up, with potential problems like PSU damage (due to surge current at startup), voltage fluctuations (due to the different circuit harmonics),

3 (AKA hard-modding) on the other hand, has been done by overclockers with some success, but like you said, need to be very familiar with the circuit. I did it once before, seriously its not for the faint of heart.

What lohwenli said is true, you will need to hardmod the board to put a current probe the measure the current exactly. I have seen people done before and its not an easy task. The end result is a messed up board after the measurement. But if you dont want to hardmod the board, a good indication if you mosfet is work hard is check the temperature. Usually the at ghigh current draw, the mosfet temperature will be around 70-90 celcius

This post has been edited by ljs2000my: Jul 12 2007, 11:03 PM

Question on this Vdroop thingy....

Say that I set 1.45v vcore in BIOS. In windows, vcore is 1.41v and at load it drops down to 1.38v (using any app e.g. CPUz or speedfan)

If that is the case, if I compensate for the vdroop, would this put too much strain on the proc? e.g. 1.50v in BIOS, 1.45v idle in windows, and vdroop 1.41v.

I mean, if the CPU is running at 1.41v at load due to vdroop, then theorectically the CPU is only receiving that much amount of voltage rather than the vcore at idle, right? (which is 1.45v in windows or maybe correctly at 1.50v as set in BIOS)

Anybody can clarify this?

Note : I realize that software monitoring is inaccurate but for the drop in voltage, I guess that would be quite accurate right?

This post has been edited by kmarc: Aug 13 2007, 01:04 PM

there is nuthin' to do with psu..
its the board design itself & the worst still, asus board vdroop amongst all board maker

I hafta gree with that...i have tried a number of ASUS boards from 965p to 975...both have serious vdroop at high OC

So far, the P35 is the best....very very low vdroop or can consider as no vdroop at all.