Joe Camel's 920 / ASUS P5WD2-E

HISTORY:


no matter what i do, i cant even get it to POST!!!

push the power button all i want...NOTHING!!

no fans, no lights, no beeps....NOTHING!!






.




..




...





....





.....





......








the rest will be here tomorrow! (friday)
(its sitting @ the local UPS hub)


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

ill start a new thread soon but i figured id throw you a teaser pic...
(formating/installing OS...)





a few quick things...

the BT mounting brackets dont really fit... the back plate hits one of the MOSFET HS mounting "rivets" and the top plate hits the MOSFET HS (file down the top plate and your good to go)


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


OK now that thats out of the way, i can get down to some STATS:
o-wait, im still installing OS....

any questions so far?

(besides why did i get a SLI card when the mobo is X-Fire ready)


--------

CPU: (stock) 200 x14 / 1.3v
mem: (stock) 4,4,4,18 / 2.0v


sandra:

CPU.......14842 / 7120

MEM......4986 / 4989

M.M. .....31528 / 37411


-----------

EXTREME air...






lets just say the MOSFET's HS is nice and cool now


although ive been into the 1.5v (BIOS setting) >300 seems to be bench-able but not consistently "stable"...


but for the first hours with a new mobo and "gen-1" BIOS...im not too unhappy.


PS THIS FAN is for benching purposes ONLY, ... its just too loud for 24/7

----







ya, thats a little too high in the temps department....


lets water cool this thing......

.

yep, im a hardware FREAK!!
i collect the perfect specimens and exploit their POWER in the never ending pursuit of the perfect OC !!! MUAHAhahahaaa....


G5 storm, twin single pass Fedco heater cores in series, 4x 120+ CFM Nidec 120s, D4 pump running @ 16v (12v=stock) ; pump's mosfets are fan cooled







here are some pics and load temp comparison shot...







water cooling has opened up an entire new realm of possibilities to this setups potential........







this is by NO MEANS FOLDing stable or even P95/sandra...but it finished Pi1M

Note
this was @ 1.55 BIOS Vcore ... CPU-Z shows 1.496 (idle) =\




CPU-Z mod for mem settings:
http://xtremesystems.org/forums/showpost.php?p=1247815&postcount=32




original thread: in the FOLDING section... do YOU FOLD?



.

God damn, nice results. Your load temps are lower than mine @ 1.5v :-\. yay for bad IHS. Excellent results Joe ... seriously.

can you try measuring vcore with a dmm?

only if you can show me where to test...i think i still have the cheep-o-DMM from my pencil V-mods....somewhere around here

ive been up 30+ hours now, i dont know how steady my hands will be

the mosfets around the socket look the same as the p5wd2 premium, on that board you measured the middle pin on the mosfet i drew to in this picture:

got some... editing in....


..............IDLE.........LOAD..........IDLE..........LOAD...........IDLE........LOAD

BIOS.......1.475.........1.475........1.50............1.50............1.55........1.55
CPU-z......1.44..........1.42..........1.46............1.43............1.49.......1.46
DMM.......1.363.........1.339........1.373..........1.358..........1.414......1.385


(load=dual P95)

wow that is odd

does it fluxuate much on load (most important)?

also, do you have a psu with an 8-pin p4 connector to try with?

no more than .03 +/- (load or idle)

PSU is a sparkle 550 SLI
http://www.newegg.com/Product/Product.asp?Item=N82E16817103517

i only have 1x of the p4 sockets used...

the other set came with this cap on them...
(i do have another p4 off the PSU)

take off the cap and plug that sucker in!

Very nice Joe! And that's a great PCMark score, be sure to post it up.

crimedog said:

take off the cap and plug that sucker in!

Click to expand...

Is there also a 4 pin molex next to the 24p atx on that mobo?

/edit .... Doh ... never mind ... i looked at your posted pic and see that there is.

there ya go, 8k






RAS Precharge is where its @ (as far as i can tell...) having it set to 2 makes things FLY!!!!! super "snappy" (and a sweet boost in bandwidth benches)

Gautam said:

Very nice Joe! And that's a great PCMark score, be sure to post it up.

Click to expand...

ran it @ 220 FSB (above mem settings)
3D 01 with a BONE-STOCK 6600GT.... 25340


now... can it FOLD?!!!

holy, 0.O exellent!!!!!

did measure Vcore after getting the other connector pluged in?

you make this look too easy..

im still "on the fence" as to the other plug....

i assumed that was for the WS "flavor" of this mobo. (the 8 phase power thing...)


im sure in the next few days ill try it out

Joe, definitely add the 2nd 4-pin to use all 8 for the proc power, especially on a dual core. It's NOT going to solve your droop, it might not even help if the problem isn't power supply, but regulation (mobo)...that's ridiculous. It looks worse than the P5WD2 Prem. At least it doesn't overvolt at idle and then undervolt on load (which Asus is famous for), so you can probably turn the Vc up a little extra to cover some of it in the meantime.

The only thing that is going to solve the droop 100% is a vdroop mod, but first make sure what you see on CPU-Z is what is really happening...carefully measure that mosfet that crimedog mentioned. Mine (P5WD2 Prem) was dead-on what CPU-Z shows, both with and without the vdroop mod.

The Stilt (XS) should have his -E soon (bouncing around Europe in a UPS truck somewhere the last I heard) and he's a mobo modding freak, so I am sure he'll locate the points on the -E and post I am sure. It looks almost identical, so they may very well be the same as the P5WD2 Prem mods, but I haven't really seen a close up of the -E (SMTs) to tell if they are the same layout or not.

Anyway, plug in the 2nd 4-pin and measure that FET!

I'll be bringing a scope home today and really finding out how stable the P5WD2-e regulation is. It's also a storage scope, so I'll have pleany of pics to distribute. Board is in Parsippany UPS hub and will be here tomorrow. If anyone can find the manufacturer and part number for the Vc switching regulator,I'll download a datasheet and do the first Vmod myself tomorrow night. Probably Allegro or National

Ross said:

Joe, definitely add the 2nd 4-pin to use all 8 for the proc power, especially on a dual core.
-snip-
Anyway, plug in the 2nd 4-pin and measure that FET!

Click to expand...

will do!!... right after work...

after seeing what this rig can do with a (1x) 6600GT....i WANT 2x 1800's in X-Fire!!!!!

this thing has woke up the Bencher in me.... i can feel the need, the need for speed

Ross said:

The only thing that is going to solve the droop 100% is a vdroop mod, but first make sure what you see on CPU-Z is what is really happening...carefully measure that mosfet that crimedog mentioned. Mine (P5WD2 Prem) was dead-on what CPU-Z shows, both with and without the vdroop mod.

Click to expand...

from my chart above it looks like there is a (almost consistent) .08V difference between CPU-Z and my DMM...

this is the cheapest (radio shack) DMM i could find, do you think IT could be off by .08V in the 2V read setting?

i know its more likely that CPU-Z is off, and if i had a nice or even decent DMM i wouldnt question it.... but a $20 DMM doesn't inspire much confidence in me.

http://www.radioshack.com/product/i...4&cp=&kw=digital+multimeter&parentPage=search

proth, I don't even know where the regulation chip itself is on the board. For the vdroop mod on the P5WD2 Prem, it's a jump between 2 SMTs. I guess it's some kind of feedback circuit, but DECREASING resistance stabilizes it instead of increasing resistance like on an IC FB loop, which I found a little odd. I realize it should be able to be done elsewhere on the board (IC), but maybe Stilt did it at the SMTs on purpose...the IC chip may be buried in a tight spot.

Joe, LOL. How about 2x 1900s in CF? Not a whole lot more expensive than x1800XTs and they're affordable...NOT! I hope the extra molex helps with the OC, but I doubt it will with the droop...keep us posted!