P5B Deluxe Vmods

[GMX]

I googled decoupling capacitance and i understood what you said.
the thing is that WE can minimise the vdroop effect just by decreasing one resistance. Couldn't they do just this? I am not saying to use completely different parts to diminish vdroop but to at least minimise it as we do.

do you still think that I don't get it?

Its been there for years. Nothing has been done about it, except by us. I doubt the situation/story will change.

[Millyons]

Its been there for years. Nothing has been done about it, except by us. I doubt the situation/story will change.

ya one thing i dont understand is all of a sudden on the forum everyone is crying about the vdroop, and mostly bashing Asus for it....for years since i have been into computers (before celly 300a days when OCing took off) all mobos have had a vdroop ..... only way i can explain this is the ppl that cry the most and bash Asus is that they probably started OCing with the Ultra-D that really had a very slight drop or none, and now migrated to c2d and dont even know that vdroops are normal on mobos....and lol its not a Asus "feature" like some are calling it, it has always been there on all mobos...well except the ultra-d but that is why it really was an exstraodinary peice of hardware.......

[celemine1Gig]

I googled decoupling capacitance and i understood what you said.
the thing is that WE can minimise the vdroop effect just by decreasing one resistance. Couldn't they do just this? I am not saying to use completely different parts to diminish vdroop but to at least minimise it as we do.

do you still think that I don't get it?

You won't like to hear it, but yes.
What you still don't seem to understand is that there are engineering rules. Changing the droop function has more effect on the vcore regulation than just the one different resistance. This resistance controls the droop function, it is not the droop function.
As I already said, they want to safe money, as they want to make money at all. I don't think that the manufacturers earn a fortune designing motherboards for enthusiasts. We are not the main source of capital in this industry. When you wanted a board that was perfect (speaking about regulation and parts), I guess you'd have to pay more than for example the latest Asus boards and still the board wouldn't sell good, as most people who buy these boards (even the so called enthusiasts) don't have a clue about vdroop etc. That's to say colored LEDs and mechanical and thermal features often are more eye-catching and buying reason.

So why honestly would any manufacturer in his right mind listen to what you say? Nothing against you, but think about it.

[pumbertot]

You won't like to hear it, but yes.
What you still don't seem to understand is that there are engineering rules. Changing the droop function has more effect on the vcore regulation than just the one different resistance. This resistance controls the droop function, it is not the droop function.
As I already said, they want to safe money, as they want to make money at all. I don't think that the manufacturers earn a fortune designing motherboards for enthusiasts. We are not the main source of capital in this industry. When you wanted a board that was perfect (speaking about regulation and parts), I guess you'd have to pay more than for example the latest Asus boards and still the board wouldn't sell good, as most people who buy these boards (even the so called enthusiasts) don't have a clue about vdroop etc. That's to say colored LEDs and mechanical and thermal features often are more eye-catching and buying reason.

So why honestly would any manufacturer in his right mind listen to what you say? Nothing against you, but think about it.

of course but hey wasnt the DFI venus designed designed to be like the expert but made purely for o/c. they could release P5B-Deluxe Extreme Edition with 0.001Vdrop, 2C Vcore, 1.9VMCH etc.lol just a thought.

lol@me big update. yep mobo survived my not so great mod skills and then I killed here by trying to flash to older BIOs(why do you ask, because always looking for a better overclock). tried 24 BIOS reset(battery out etc etc) but no good no boot.
so have to RMA. prob is getting the mods off with crap iron I have. when I removed vmch wire the solder flowed between the 2 legs on the IC and no matter how I try with braid and iron it wont budge.2 legs are shorted.bah!

any tips on how to get it off?

[celemine1Gig]

of course but hey wasnt the DFI venus designed designed to be like the expert but made purely for o/c. they could release P5B-Deluxe Extreme Edition with 0.001Vdrop, 2C Vcore, 1.9VMCH etc.lol just a thought.

lol@me big update. yep mobo survived my not so great mod skills and then I killed here by trying to flash to older BIOs(why do you ask, because always looking for a better overclock). tried 24 BIOS reset(battery out etc etc) but no good no boot.
so have to RMA. prob is getting the mods off with crap iron I have. when I removed vmch wire the solder flowed between the 2 legs on the IC and no matter how I try with braid and iron it wont budge.2 legs are shorted.bah!

any tips on how to get it off?

You don't need to RMA the whole board when only the BIOS is the problem. Ask someone from the forums to rewrite the EEPROM for you. I bet someone near you (I mean in your country ) will be able and willing to help you. It's no big deal and sure cheaper and faster than RMA.

And concerning the DFI Venus:
a) It was, as I said, very very expensive.
b) All the difference was, AFAIK, the all solid caps and another BIOS. If it was a bit more complicated, I apologize.
c) The Asus P5B-E Plus ist sort of Asus's Venus. Yet still there's the droop "problem" you all moan about and I doubt that Asus will ever do anything about it. Reasons stated above.

[funkflix]

Try one of that things:





And try to boot with the Asus-CD was in the package of the board, maybe it will flash the old original bios. Or try a disk with flashprogram + biosfile. Sometimes the boards will boot this way!

@celemine1Gig

EEPROM is not in a socket, it is soldered to the board.

[pumbertot]

Try one of that things:





And try to boot with the Asus-CD was in the package of the board, maybe it will flash the old original bios. Or try a disk with flashprogram + biosfile. Sometimes the boards will boot this way!

@celemine1Gig

EEPROM is not in a socket, it is soldered to the board.

I think I will try that suction thing, already have wick/braid and it wont budge.lol if I just RMAed as is they would be like wtf is this?

[funkflix]

Anyhow, are u sure that ur P5B bios is damaged? How the board affects if u power the pc on? Just black screen, but ur screen getting a signal? Or beeps?

Three days ago i flashed 1004 bios, and i needed about 2 hours to get my P5B back to life.. Thought it was dead too.

[celemine1Gig]

...

@celemine1Gig

EEPROM is not in a socket, it is soldered to the board.

WOW! All I can say is damn Asus. Intel has no socketed BIOS on BadAxe and many other boards, but they got good recovery functions. Is there none on the Asus?

[funkflix]

WOW! All I can say is damn Asus. Intel has no socketed BIOS on BadAxe and many other boards, but they got good recovery functions. Is there none on the Asus?

Some guy on hwluxx wrote, that he get back his P5B to life, when he insert teh Asus-CD from the OVP. Then the board flashed back the 605 bios to the EEPROM and it worked again. Know that function from my old lanparty nf3, it flashed new bios from disk, after i had a corrupted flash.

I would give that all a try before i make the RMA.

[celemine1Gig]

Ahh, looks like they integrated ATA support to the bootblock. Sounds like a chance.

[pumbertot]

Anyhow, are u sure that ur P5B bios is damaged? How the board affects if u power the pc on? Just black screen, but ur screen getting a signal? Or beeps?

Three days ago i flashed 1004 bios, and i needed about 2 hours to get my P5B back to life.. Thought it was dead too.

I tried battery out for 24 hours. will try boot with cd like that other guy did but I need to remove this solder short first as cant power it on,lol.

yeah it was black screen.no ginal, no beeps.

[Gam3Ra]

I need pencil vmod for the memory, please

[eva2000]

Did you do it?
If so, what is the result?

Slowly finding the balls to be a bit more adventurous and decided to do the Asus P5B Deluxe vdroop pencil mod I used a 2B pencil and about 4 strokes across resistor pictured with yellow line through it below:



------------------------------------------------------------------------------------
Before vdroop pencil mod
- resistance was DMM measured at 80 ohm
- 1.325v bios set vcore = 1.280-1.288v bios monitored, 1.29v idle/1.27v load Asus PC Probe2 reported.

------------------------------------------------------------------------------------
After vdroop pencil mod
- resistance was DMM measured at 37.4 ohm
- 1.325v bios set vcore = 1.296v steady bios monitored, 1.30v idle/1.30v load Asus PC Probe2 reported.

I'm happy with that for now a reduction in vdroop by ~0.03v Asus PC Probe2 reported under load. Not going to be too greedy

------------------------------------------------------------------------------------
Before vdroop pencil mod @1.325v bios set
CPU Prime95 v25.1v small ffts load:


CPU Prime95 v25.1v small ffts post load idle:


------------------------------------------------------------------------------------
After vdroop pencil mod @1.325v bios set
CPU Prime95 v25.1v small ffts load:


CPU Prime95 v25.1v small ffts post load idle:


------------------------------------------------------------------------------------

[celemine1Gig]

@eva2000:
Nicely done! Just two things:

1.) Didn't you measure Vcore with a DMM? You wrote that you just measured the resistance before and after the mod, but no word about measuring vcore while testing. I wouldn't really trust the software readings too much and I think you know that, too.

2.)You do know that droop is a function of the current drawn by the CPU, do you? Honestly, I doubt that 2.933GHz at ~1.3V is your normal overclock, thus your droop-mod results won't help you much. If you want to minimize vdroop, then you have to do it again for each new operating point of the circuit. That's to say when you raise your vcore now and overclock some more, vdroop will occur again, although you did the mod. The reason is the higher current draw. If you are clocking at very high vcores for example, I'd always advise you to do the mod adjustable (with a trimmer), as it could be that you adjust vdroop to a minimum at high vcore, but then get increased idle volts at lower vcore settings, as a result.

Just a little reminder.

[eva2000]

Yeah to be honest i totally forgot to DMM measure the vcore (didn't even bother looking for vcore measure pt in this thread) - was so preoccupied with the resistance measurement hehe

Yeah there's a bit more droop at higher vcore i.e. i get 0.008v droop under load at 3807mhz at 1.5875v bios (1.56v idle/1.552v load)

Idle


Load



@3600Mhz use to take 1.45v (1.38v Asus PC Probe2 loaded), now with vdroop pencil mod 1.4125v (1.38v Asus PC Probe2 loaded)



Also my P5B Deluxe is setup in a part of the room where it's hard to get down on my knees to hold a DMM steady for measuring vcore under load hehe

As i said, I'm not going to be too greedy here... I'm happy with the improvement so far...

[K404]

From what I understand, Droop is built into Intels spec for CPUs, but AMD dont worry about things like that. The entire power circuit set, although obviously having similar raw components and basic purpose, is totally different.

To compare the Ultra-D (and NF4 LP series) to the C2D boards is crazy.

Quick question: The droop pot mod is applied directly across a resistor...is it safe to say this mod works off voltage divider ruling, not tricking the sense pin of an IC?

[Millyons]

eva...did u try a quad yet with the vdroop....you will probably have to adjust it since with quads there is even bigger vdroop

[4Qman]

eva2000 Superb mate. I was ready to strip my entire rig just to do this mod. No need now i seen your results mate

[eva2000]

thanks guys..

haven't tried Q6600 kents yet but it did 3600mhz at 1.45v bios set vcore before vdroop so can only be better now with penciled vdroop as is anyway

final outcome 4hrs solid @3807mhz at 1.5875v bios (1.56v idle/1.552v load)

Load


Post Load Idle

[eva2000]

@eva2000:
Nicely done! Just two things:

1.) Didn't you measure Vcore with a DMM? You wrote that you just measured the resistance before and after the mod, but no word about measuring vcore while testing. I wouldn't really trust the software readings too much and I think you know that, too.

2.)You do know that droop is a function of the current drawn by the CPU, do you? Honestly, I doubt that 2.933GHz at ~1.3V is your normal overclock, thus your droop-mod results won't help you much. If you want to minimize vdroop, then you have to do it again for each new operating point of the circuit. That's to say when you raise your vcore now and overclock some more, vdroop will occur again, although you did the mod. The reason is the higher current draw. If you are clocking at very high vcores for example, I'd always advise you to do the mod adjustable (with a trimmer), as it could be that you adjust vdroop to a minimum at high vcore, but then get increased idle volts at lower vcore settings, as a result.

Just a little reminder.

Guess you were spot on... At higher vcore above 1.6125v the initial pencil modded of drooping resistor from 80ohm to 37ohm resistance wasn't enough to prevent further vdroop. So whipped out the 2B pencil and dropped resistance down to 14.5 ohm. And now it's all good at least from bios perspective

Final outcome



Improvements after vdroop pencil mod are amazing! Before the mod, 10x 400FSB = 4000Mhz would need 1.7v max vcore set in bios for 1.62v idle and 1.59v load. I could pass single Super Pi 32M but Systool dual 32M Pi would reboot at 50-80% mark - Systool dual 32M Pi is a good indicator of whether a setting is capable of dual Super Pi 32M.. Now with vdroop pencil mod, with vcore set to 1.65v in bios I get 1.616-1.624v idle and 1.624v load and now capable of Systool dual 32M Pi

Idle


Systool dual 32M Pi load

[Millyons]

thanks guys..

haven't tried Q6600 kents yet but it did 3600mhz at 1.45v bios set vcore before vdroop so can only be better now with penciled vdroop as is anyway

[/IMG]

OT but man thats a nice quad

[pumbertot]

eva how are your temps so low with a storm? whats your full loop consist of?I thought its roughly same as mine but you load 10c less than me at same volts. wait have you got airconditoning at home and have a low ambient?

[Slay0r]

Eva,

your excellent results inspired me to do the pencil mod.

load

idle


mobo used to vdroop by 0.03/0.05v from idle to load at the same clocks and voltage set in bios (1.5000v) not to mention that at idle it sat @ 1.450 ish

very quick and worthy mod imho

[eva2000]

eva how are your temps so low with a storm? whats your full loop consist of?I thought its roughly same as mine but you load 10c less than me at same volts. wait have you got airconditoning at home and have a low ambient?

Yes i have 1HP portable air con but still room temp hovers between 24-28C on cool day and 28-37C on hot day with air con on

loop = Swiftech G4 Storm/Thermochill 120.3/MCP650 + Kayl 100W Custom PSU with T-line

The reason - it's the cpu, this X6800 i have runs 15-20C cooler than any core2 cpu i have used at the same volts! all my other E6600/E6700 run 3600mhz at 1.35-1.4v in same room at 60-68C load versus X6800 @3600mhz at 1.4v at 42-48C load!

it could be due to good IHS - notice the corners are lower than the middle!

Eva,

your excellent results inspired me to do the pencil mod.

mobo used to vdroop by 0.03/0.05v from idle to load at the same clocks and voltage set in bios (1.5000v) not to mention that at idle it sat @ 1.450 ish

very quick and worthy mod imho

congrats