Asus P5B Vanilla (non dlx!) vcore+dimm needed

[nicepun]

Vdroop, Vcore and Vdimm have been found a few pages back.

These pics are from a P5B-E, that's why I posted them to see if the P5D vmods apply to me.

[wittekakker]

It looked like it was P5B Vanilla. Vdimm should work. For Vcore I guess it should work to, maybe you should check out how we found Vcore and Vdroop and see if your board has the same resistance and stuff.

[rjw]

Ok...here is what I got

Strangely, with the bios fsb termination set to auto, 1.2 or 1.4 volts

your 4.13 = 3.63 and 1.35 = 1.23 (both places)

with bios set to 1.3 or 1.45, your 4.13 = 3.8 and 1.35 = 1.34 (both places)



In the second picture (below), I listed all the voltages measured on my board.

I circled one pin, where like in the first picture and with the same bios settings, the voltage changes from 3.63 to 3.8




so, it seems theat with auto, 1.2 and 1.4 v I get the lower voltages and with 1.3 and 1.45 I get higher readings.

I tested while in the bios and in windows idling and the readings were the same.

Hope you can figure it out from what I measured

thx persivore

rjw

[persivore]

Can someone check the resistance between the pin measuring 4.42v on the LM324 chip and the pin on the MOSFET above it measuring 4.41v.

Can you also check the resistance between each of the pins I've marked and ground please.

[rjw]

the resistance between the pin measuring 4.42v on the LM324 chip and the pin on the MOSFET above it measuring 4.41v.

995 ohms

the resistance between each of the pins I've marked and ground please.

ON LM324

1.034k left pin

1.024k right pin

[persivore]

Vfsb: Connect a 20k variable resistor between the pin on the left (pin 6) and ground. Decrease resistance to increase voltage.
Check voltage at the MOSFET leg reading 1.35v in post #35

Vmch: Connect a 20k variable resistor between the pin on the right (pin 9) and ground. Decrease resistance to increase voltage.
Check voltage at the leg showing 1.29v on the MOSFET above the LM324 in (post #36)

[rjw]

Vfsb: Connect a 20k variable resistor between the pin on the left (pin 6) and ground. Decrease resistance to increase voltage.
Check voltage at the MOSFET leg reading 1.35v in post #35

Vmch: Connect a 20k variable resistor between the pin on the right (pin 9) and ground. Decrease resistance to increase voltage.
Check voltage at the leg showing 1.29v on the MOSFET above the LM324 in (post #36)

I'll give it a try right now

any idea as to what the max voltages could be before KABOOM ?

and again, thanks so much for the help

rjw

[Vasilii]

Hi evreyone,

I'm new here need some advise, yesterday assembled system based on P5B + celeron 2.5 Ghz, ddr2 677

I have 2 concirns about it, first tempretue, what is standard one, none O/C with boxed heat sink and fun to be expected?

second before I was using athlonXP 1.533 Ghz on asus a7v33, I don't see much of performance boost, I understend that celeron 2.5 Ghz suck but still shouldn't it outperform athlonXP 1.533 noticeably ?

V.

[wittekakker]

It will outperform the Atlon, but no big boost. 1533MHz Atlon XP -> 1800+ I guess. Got a 2,5GHz Celeron Prescott here, with a cheap mainbord I got it on 3,2GHz with stock heatsink and stock voltage. Didn't try higher, so maybe there is more left.

[rjw]

I would like to personally give a great BIG thank you to Persivore, who just added the Vfsb and Vmch mods for our board (P5B vanilla)

My setup is now totally stable at over 4.5 Ghz. I included my latest OC effort below. Is this board any good?...check it out

[nicepun]

Got the vdimm and vdroop done, they worked great. I'll do the vfsb and vmch next.

rjw, Could you please post pics of your vfsb and vmch vmods. I need to know which point to use as ground soldering points. Thanks.

[wittekakker]

I need to do some work, I will try the mods later. Big thx Persivore!

[rjw]

Got the vdimm and vdroop done, they worked great. I'll do the vfsb and vmch next.

rjw, Could you please post pics of your vfsb and vmch vmods. I need to know which point to use as ground soldering points. Thanks.

When I did the Core, droop & dimms mods, I ran 2 4 conductor flat cables to a drive bay, where I had the 3 vr's montd on a board. So I had 2 grounds already.I just ran another 4 wire flat cable and used the grounds that I already ran. Other wise there are plenty of ground points on the MB.

Good luck

[nicepun]

When I did the Core, droop & dimms mods, I ran 2 4 conductor flat cables to a drive bay, where I had the 3 vr's montd on a board. So I had 2 grounds already.I just ran another 4 wire flat cable and used the grounds that I already ran. Other wise there are plenty of ground points on the MB.

Good luck

What voltage you have for the vmch and vfsb?

[rjw]

What voltage you have for the vmch and vfsb?

I've been up and down with them, but to start with, try 1.4

I've done my highest overclock at 1.55 to 1.65, but it made not difference. Both were stable.

I may have gotten there at 1.4 as well, but haven't had time to check that.

The key with my board is that bios will strangely set both to either 1.2xx or 1.3xx as per one of my previous posts.

auto, 1.2 and 1.4 are at the lower seting, while 1.3 and 1.45 are at higher setting.

So, if you set them to 1.4 to start with, at least they remain stable and at a value just above where the bios would set them.

according to the formula fsb = 166.67 x Vmch + 170, I should need well over 2 volts for 516 FSB, but as I get over 1.6 to 1.7, the NB starts to get a little tepid, plus I saw no difference between 1.55 and 1.7.

I have a copper heatsink and fan on my NB, so while the heat is being removed, I have found that in my case, I seem to be limited by cpu voltage more than NB.

My E6600 is only completely stable at 1.75v or higher (when over 500 fsb)

Also, 1.65v causes the cpu to run hotter,, but then from 1.65 to 1.8, it doen't really change.

I still have a long way to go but there seems to be some oddities once over 500 FSB.

I think that it has to do with the loose timings that are set by the bios.

Once I am in windows, I can use memset to change mem timings drastically, but it won't start windows if you try to tighten them up in the bios.

Which brings up another observation.

when I first started to exceed 500 fsb, I had to set the bios to mem controlled by spd, or (sometimes) no windows. my mem is rated at 5-5-5-15 @ 1064. (Corsair pc8500 dominator) at 2.2volts.

With these settings, I made it to 513 fsb (barely)

lately, I am using 6-6-6 along with loosening all the other timings in the bios and I can boot into windows at over 516.

But then the machine is a dog, even tho cpu is over 4.6 Ghz.

So memset comes in and I can set the timings down to 4's or 5's and things pick up again.

I think that the bios is totally confused at such high fsb and cpu speeds.

At 550 x 7 , it did fine.

Just some observations on my board that may help others

[SMa]

I did the vdimm mod on my P5B-E.
But something seems to be seriously wrong here.
When the board goes into the BIOS vmem starts dropping to ~1.85v, when I reboot after setting up the BIOS vmem drops to 1.20v.

So 2 options:
measuring point is different on P5B-E
vdimm mod is different and I'm overvolting something else

[nicepun]

I did the vdimm mod on my P5B-E.
But something seems to be seriously wrong here.
When the board goes into the BIOS vmem starts dropping to ~1.85v, when I reboot after setting up the BIOS vmem drops to 1.20v.

So 2 options:
measuring point is different on P5B-E
vdimm mod is different and I'm overvolting something else

I did it and it worked fine for me? What do you have the vdimm set in bios? I set my to 2.1vdimm before doint the vdimm mod.

[SMa]

I did it and it worked fine for me? What do you have the vdimm set in bios? I set my to 2.1vdimm before doint the vdimm mod.

I set it auto.
I'll try 2.1 too.

If 'auto' causes the problem,
then vmch mod doesn't work either?

[nicepun]

I set it auto.
I'll try 2.1 too.

If 'auto' causes the problem,
then vmch mod doesn't work either?

I have not done the vmch mod yet, I will do it this weekend when I have time.

[wittekakker]

Sma, before doing the soldering you should have checked out if the readpoint was actually the good one. You can still undo the soldering first, or put your VR on it's maximum value so you can not do any damage. Check the Vdimm with 1,8V and 2,1V set in BIOS and see if the voltage changes on the suggested point.


How I use it: I change Vdimm in BIOS to 2,1V. Then I use the hardmod to give 2,4V. After doing benchmarks and testing I just set my voltage back in BIOS so that I don't have to turn the VR 10 times every time I'll want do benchmark.
Example: 2,1V set in BIOS gives my 2,4V with the hardmod. If I switch to 1,8V in BIOS I get actually 2,1V which is a safe value for my memory.

[nicepun]

How I use it: I change Vdimm in BIOS to 2,1V. Then I use the hardmod to give 2,4V. After doing benchmarks and testing I just set my voltage back in BIOS so that I don't have to turn the VR 10 times every time I'll want do benchmark.
Example: 2,1V set in BIOS gives my 2,4V with the hardmod. If I switch to 1,8V in BIOS I get actually 2,1V which is a safe value for my memory.

This is exactly what I did, except I went up to 2.5v, I'm thinking about doing 2.6v just in case I want to do some benching/testing.

[rjw]

[QUOTE=

How I use it: I change Vdimm in BIOS to 2,1V. Then I use the hardmod to give 2,4V. After doing benchmarks and testing I just set my voltage back in BIOS so that I don't have to turn the VR 10 times every time I'll want do benchmark.
Example: 2,1V set in BIOS gives my 2,4V with the hardmod. If I switch to 1,8V in BIOS I get actually 2,1V which is a safe value for my memory.[/QUOTE]

I agree

Also, never use auto. on my board, in auto, the bios or mb tries to lower the voltage, so you are fighting the bios.

has anyone else tried Vmch.

It stabilized my board, but even at 1.7v I could't get higher fsb. Somewhere around 1.75 or 1.8. the board won't boot.

[M@ster]

@ nicepun
Any chance you post some pics of your system with the mods done?
Tks

[nicepun]

rjw, So you would say that the mvch and vfsb is not necessary?

[rjw]

rjw, So you would say that the mvch and vfsb is not necessary?

I haven't seen any gains from pushing them to the limit.

I have been to 9 x 520 with 1.5 to 1.75v on both, but with the mods, it was solid. without, voltages never got close and stability was questionable

I've been playing with a few other asus boards as well

p5b deluxe
p5wdh deluxe
p5wdg2 ws pro

all have vmods, and while I need to do more testing, the p5b and p5b deluxe are the easiest (so far) to get to over 500 fsb.

as someone stated in another post, the 965 chip tries to start windows and will crash later, while if the 975 chip isn't happy, windows will be no show.

on another note, I haven't been able to get my ex 6800 up as high as my e6600, as of yet.

If I needed a chip today, I would probably look into the E6700

once again, the jury is still out on the other boards and chip, till more testing can be done

that's just my experience so far