get more vcore under load: vdroop pencil mod for p5b-deluxe (pics)

[satandole666]

Alright, I'll add another success story here.

I have the P5B vanilla PCB. I followed the picture here.

Before the mod:
Bios idle load
1.35 1.296 1.288
1.50 1.440 1.424

Average bios loss: .057V
Average droop: .012

Also, my board would not supply over 1.512V at idle and 1.488V under load no matter what bios setting before the mod.

After the mod:
Bios idle load
1.35 1.328 1.320
1.50 1.472 1.464

Average bios loss: .025v
Average droop: .008V

The mod improved my droop and bios loss across the board. It also improved my max vcore substatially. It has reached 1.568v so far and I haven't tested any higher.

Thanks for the help. Now I'll get my E6600 back over 3.6ghz.

 

[jchunter]

Somebody needs to take away the punchbowl, here.

That pencil mark is a graphite resistor of unknown value that you are placing in parallel with a surface mount component. If you are lucky, that component is a resistor and you are just reducing its net value by adding another resistor in parallel. This will result in more current and higher voltage passing into other (unknown) components that may or may not be able to handle it. The resistance of the mark is a determined by the width and thickness of the mark and the uniformity of the graphite particles (not to mention different thermal expansion characteristics) and so will be guite variable and so will Vdroop.

In addition to those obvious problems, there is nothing bonding those graphite particles to the component and if you hit or jar the computer, you could shake some of them loose and two things could happen.

1) Vdroop could change (up or down), which might push a heavily modded CPU over the edge into instability. You might not notice for weeks, until you stress the computer...

2) The graphite particles that came loose are quite conductive and could land anywhere on the circuit board and cause any number of other problems. Tiny fireworks, anyone?

Bottom line: this is bad engineering. Don't do it. Your run the risk of transforming a reliable, predictable computer into an unstable POS.

 

[Hazaro]

Somebody needs to take away the punchbowl, here.

That pencil mark is a graphite resistor of unknown value that you are placing in parallel with a surface mount component. If you are lucky, that component is a resistor and you are just reducing its net value by adding another resistor in parallel. This will result in more current and higher voltage passing into other (unknown) components that may or may not be able to handle it. The resistance of the mark is a determined by the width and thickness of the mark and the uniformity of the graphite particles (not to mention different thermal expansion characteristics) and so will be guite variable and so will Vdroop.

In addition to those obvious problems, there is nothing bonding those graphite particles to the component and if you hit or jar the computer, you could shake some of them loose and two things could happen.

1) Vdroop could change (up or down), which might push a heavily modded CPU over the edge into instability. You might not notice for weeks, until you stress the computer...

2) The graphite particles that came loose are quite conductive and could land anywhere on the circuit board and cause any number of other problems. Tiny fireworks, anyone?

Bottom line: this is bad engineering. Don't do it. Your run the risk of transforming a reliable, predictable computer into an unstable POS.

Nah, I blew on it and have hit more times then I have fingers.

Best solution ever!

*While I may agree it's not the best way, it does work, and for the moment I have no problems.

 

[graysky]

@jchunter - I have been pencil modding both MBs and processors since... well, forever (1993 I think). I have never had a board or processor go bad on me yet and this is after doing it to several dozen machines.

 

[jchunter]

Gray,
I've read your guide to overclocking and it helped me a lot. Thank you for writing it. The only thing that disagreed with was the reference to using a pencil to control Vdroop. In fact, I almost fell out of my chair…

Glad to hear that you have been lucky with your mods. However, your experience does not prove that the method is either safe or reliable. IMO, it is neither.

Electrical Engineers have made steady progress in this field by controlling the precision and reliability of the components that they use to construct complicated systems. We've made things better, faster, and cheaper every 2-3 years by painstaking control of quality. This is the reason that you and I can buy a quad CPU chip with several tens of millions of interconnected transistors, all able to sing and dance together at frequencies of several GHz.

Our presumption has to be that these engineers are extremely competent:

1) Vdroop is there for a reason. The board designers had an important purpose in mind. My guess is that it plays a role in voltage regulation. Nobody knows enough to declare that it has no purpose and that it can be “fixed” by changing the values of certain components.

2) No one has proved that Vdroop is "bad" or that voltage regulation would be better if Vdroop is minimized.

3) Until a qualified engineering team decides on points 1 and 2, IMO, nobody should be field-modifying an expensive motherboard.

 

[Hazaro]

Gray,
I've read your guide to overclocking and it helped me a lot. Thank you for writing it. The only thing that disagreed with was the reference to using a pencil to control Vdroop. In fact, I almost fell out of my chair…

Glad to hear that you have been lucky with your mods. However, your experience does not prove that the method is either safe or reliable. IMO, it is neither.

Electrical Engineers have made steady progress in this field by controlling the precision and reliability of the components that they use to construct complicated systems. We've made things better, faster, and cheaper every 2-3 years by painstaking control of quality. This is the reason that you and I can buy a quad CPU chip with several tens of millions of interconnected transistors, all able to sing and dance together at frequencies of several GHz.

Our presumption has to be that these engineers are extremely competent:

1) Vdroop is there for a reason. The board designers had an important purpose in mind. My guess is that it plays a role in voltage regulation. Nobody knows enough to declare that it has no purpose and that it can be “fixed” by changing the values of certain components.

2) No one has proved that Vdroop is "bad" or that voltage regulation would be better if Vdroop is minimized.

3) Until a qualified engineering team decides on points 1 and 2, IMO, nobody should be field-modifying an expensive motherboard.

But it's our motherboard, so who cares?

Secondly it's there to protect (If it's protecting anything) the stock CPU's from shooting up into the high high temps if the people are running Pentium 4's or something. Even then you get thermal throttling.

It's not for overclockers. I can't have my processor idling at 1.54V to get 1.45V under load to reach 3.5Ghz... Not good for the chip, with the mod I can do 1.44V idle and get 1.45V under load, thus enabling more over clocking room. Some other motherboard scale the Vdroop, bad enough where it was 0.7V from one member's experience.

 

[jchunter]

Anyone who is contemplating performing any Vdroop modification should read the following http://www.thetechrepository.com/showthread.php?t=126 because (quoting from the paper):

"you're much better off learning why Vdroop exists (and why it's needed), rather than attempting to hack up your expensive, well-engineered motherboard in a horrific effort to "correct" the problem."

 

[Hazaro]

Anyone who is contemplating performing any Vdroop modification should read the following http://www.thetechrepository.com/showthread.php?t=126 because (quoting from the paper):

"you're much better off learning why Vdroop exists (and why it's needed), rather than attempting to hack up your expensive, well-engineered motherboard in a horrific effort to "correct" the problem."

Two sentences before the he states this: the axiom - "bigger is always better" is certainly not true in this instance. For example, the addition of extra capacitors, the favorite VR circuit modification of the over-zealous albeit misguided basement-hobbyest, often does more harm than good.

The first question that may come to mind is why droop voltage at all. Truthfully, in most cases the designer may determine that a more cost-effective solution can be achieved by adding droop. Droop can help to reduce the output-voltage spike that results from fast load/current demand changes. The magnitude of the spike is proportional to the magnitude of the load swing and the ESR/ESL of the output capacitor(s) selected. By positioning the no-load voltage (VNL) level near the upper specification limit (bound by the Vccmin load line), a larger negative spike can be sustained without crossing the lower limit. By adding a well controlled output impedance (RLL), the output voltage under load can be effectively 'level shifted' down so that a larger positive spike can be sustained without crossing the upper specification limit (such as when the system suddenly leaves a heavy load condition). This makes sense as the heavier the CPU loading the smaller the potential negative spike and vice versa for lower CPU loading/positive spikes. The resulting system is one in which the system operation point is bound by Vccmin and Vccmax at all times (although short excursions above Vccmax are allowed by design).

The first question that may come to mind is why droop voltage at all. Truthfully, in most cases the designer may determine that a more cost-effective solution can be achieved by adding droop.

It seems the Vdroop is in place to safeguard the processor from crashing, generally from a quick idle to load change.

Lowering the Vcore means that the spike is also lowered, hence lowering the chance of any problem.

The problem with this is that most of our CPU's here (OCF) aren't going to take the 0.xx drop and keep stability with that drop in voltage. There is also no way I'm going to be running 1.54V idle to keep my processor 3.5Ghz stable at 1.46V.

Heck I bet we run our processors higher than the voltage spike should be.

Please correct me on anything.

 

[satandole666]

We could speculate why vdroop is around. There are plenty of viable reasons. I'm willing to bet almost all of those reasons don't apply to the overclocking crowd. Do we care about the lifetime of our chip? To an extent, yes. Do we care about loaded temps? Hell yes, but if they are within an acceptable range we'd like to push the mhz/vcore even more.

For the enthusiast crowd to accept vdroop they would have to have it somehow magically increase my max overclock. Otherwise I'm going to do a pencil mod to remove the majority of that vdroop, increase my load voltage and decrease my idle temps.

Its been over a week since I've done the mod and I've packed my computer up and moved to another apartment. I even dropped the computer box (talk about a moment) about a foot and half and the mod is still in place and working fine.

 

[jchunter]

Hazaro, Satandole,
Try enabling CPU Voltage Damper in the Bios (Advanced Tab below Vcore setup). It brought my Vdroop down to 0.008v. @ 3.3GHz, Vcore=1.28. At 3.5GHz., Vcore=1.376, Vdroop is 0.016v.

 

[bing]

Hazaro, Satandole,
Try enabling CPU Voltage Damper in the Bios (Hardware Monitor section). It brought my Vdroop down to 0.008v. @ 3.3GHz, Vcore=1.28. At 3.5GHz., Vcore=1.376, Vdroop is 0.016v.

There is no such feature in the P5B-Deluxe BIOS.

 

[Hazaro]

Or P5B vanilla

 

[jchunter]

I have it in my P5K-E. It would be strange not to find it in the P5K. I'll have to restart to locate it.

Edit: Sorry, it is in the Advanced Tab, way down on the bottom below the Vcore setup.

 

[Hazaro]

P5K =/= P5B

P35 chipset and a G/P 965 chipset

 

[Randyman...]

What do you think the "CPU Damper" setting is doing? Sounds like a "BIOS enabled Pencil Mod" to me

V-Droop might be fine and dandy for stock systems - but as numerous others have stated, heavy OC'ers have much to gain by lowering Idle Vcore while maintaining a solid Vcore under load. It just makes sense...

 

[jchunter]

What do you think the "CPU Damper" setting is doing? Sounds like a "BIOS enabled Pencil Mod" to me

Excellent thought! Maybe it is. However, you can bet your bipi that this mod is made of finer stuff than your pencils. I'll bet that it even has a known resistance that doesn't fluctuate with the phase of the moon or the temperature of the board...

 

[Randyman...]

Excellent thought! Maybe it is. However, you can bet your bipi that this mod is made of finer stuff than your pencils.

Like a Carbon Resistor? ( )

 

[jchunter]

With all Vdroop's bad rep, Asus was probably too embarassed to call it a Vdroop Minimizer...

BTW, the P5K-E manual explains the operation of the CPU Voltage Damper as follows: "Enables or disables the CPU Voltage Damper." You have to love these guys.

 

[Reefa_Madness]

With all Vdroop's bad rep, Asus was probably too embarassed to call it a Vdroop Minimizer...

BTW, the P5K-E manual explains the operation of the CPU Voltage Damper as follows: "Enables or disables the CPU Voltage Damper." You have to love these guys.

For what its worth, the Asus P5K Deluxe was the first motherboard to have that Vdroop implementation in BIOS. The option exists in the BIOS of the "vanilla" P5K, but is non-functional.

When I read your first post about not doing these type of mods I was going to point you in the direction of some threads discussing the CPU Voltage Damper...until you posted that was the board you used.

I found it sort of funny that you were warning everyone about doing this very common mobo mod and all the time your own board has this built into the BIOS.

 

[jchunter]

Reefa,
There is an enormous difference between an engineering implementation that strikes a balance between droop and voltage regulation and a goofy "penciled in" stripe of graphite of unknown resistance that could easily change over time due to thermal and mechanical stress.