ljapilot

Glad it worked our for you haydent.

Thanks for posting the results.

Good luck dont burn it up hehehe

haydent

well after some more reading and finger crossing im running E8500 @ 4Ghz again but stable enough to stay there this time, and run small prime and occt tests (will run longer when get chance)

what really made the difference i believe was putting the optional fan on the northbridge, (so im looking at getting something more serious and quieter for that job)

that thing gets hot enough as it is but when you are upping the frequency and voltage to it, you can burn yourself on it....

some other settings that really helped give back the stability (but ill try scaleing them back over time if i can) were:

NB: 1.56 v
HT: 1.52 v
SB: 1.56 v

(and ones that i also increased again to get 4.2Ghz out of it)

Also my vcore is set to 1.3 v in the bios, reading 1.26 v in windows. Leaving my cores at 41 deg C idle

by increasing those last 4 values i was able to get 4.2 Ghz (stable enough for 3dmark to finish with 1686 marks), and i was even able to post and get into windows upto 4.44 Ghz but not for long it was way to unstable. and the board wont post anything from 4.5 ish up.

but ill stick with 4 Ghz, as i want to get some use out of it atleast seems its brand new and im poor...


btw. - im running a zalman 9900 cooler, 2x 8800gts sli, 4gb of 1066mhz kingston ram unlinked, qfan 850w, and speedfan + rivatuner to manage my temps and fans.

haydent

Well after talking to a friend about this mod and him getting interested and then reading some more about it he filled me in on the published purpose of vdroop, something i hadnt bothered to read.

basically to prevent voltage spikes to the cpu when high load is taken of the system.

its not hard to find people talking about this issue online, and i felt the issue was worth reconsidering because apparently the 45nm chips are more sensitive to voltage than the older 65nm.

one thing id like to find out more about is how high these voltage spikes can be and if they could go above the supposed safe limit of ~1.37v if you are running 1.26v idle with a vdroop mod...

the other thing is if over time it could be safer to run:

A: a vdroop mod, 1.27v and 55 deg C load
or
B: no mod, 1.35v and 67 deg C load ... ???

...

so with thes thought in mind i figured to open up my case again and go the safe way and use as little pencil as possible to just eliminate vdroop enough to maintain stable, but not disable the feature all together... ( see i had been rather generous with my coloring in )

so i started by erasing it off, easy worked well. than i thought perhaps i should down clock to 3.6Ghz from 4Ghz (my stable prior to the mod) or up my vcore so i could at least boot windows to test if i had erased it properly by testing for vdroop with occt.

but i figured hey, i might as well try at 4 and hey what do you know it went in fine, ran a little of prime to see if it wouldnt lock the momont it was stressed then went into occt at ran a 6min test which i was suprised to pass, and see that infact vdroop was back, i had gotten the lead of and it was now bigger than b4 as i am running 4 rather than 3.6.

Results: Ripple: (1.28% @ 4.0 & 1.23% @ 3.6)



as you can see its back and bigger, but the crazy thing is im getting results and its no locking up and im only running ~1.25v .. !!!

so i can hardly so its absolute stable as i really havent stressed it but its way better than no windows at all...

all i can think it is, is from increasing the HT,NB & SB voltages as mentioned in my previous post.. that it was them all along, not the vdroop !!

so ive almost got the best of both worlds, low vcore and temp, the supposed safety of vdroop, and a high stable overclock !

its just the mobo running higher, (and i should mention i also added the NB FAN which could be helping alot considering its normal temp), which i dont mind cooking as much as its worth $100Au less ...

good luck all,


Hayden.

muledeer

Has anyone gotten this board stable with fsb over 1600. My Q6600 is at 3.6GHZ. stable,but I have room on voltages and temps for higher oc, but I can't hardly even post above 1600 let alone stable. I tried for hours and got 3.75GHZ. to boot, but everytime I benched with 3DMARK06 it would crash Any ideas ? I want to hit 4.0GHZ. sooooooo bad

__________________

4 GHz Overclock Club
OCN Water Cooling Club


3DMARK06 22084

http://valid.canardpc.com/show_oc.php?id=709636 max 4.833 Ghz.

haydent

im running 1688 Mhz Rated FSB now stable , and got a post as high as just over 1852, but beyond that noting but error beeps and blank screens.

try settingg your NB/SB/HT voltages as i posted earlier, and run the optional NB fan, its really needed !!!

muledeer

Wow thats a high fsb for this board, but it handles dual cores a lot better than quads. I guess 3.6 is going to be my limit

__________________

4 GHz Overclock Club
OCN Water Cooling Club


3DMARK06 22084

http://valid.canardpc.com/show_oc.php?id=709636 max 4.833 Ghz.

haydent

well that vdroop i had getting down to 1.23v was too low it seems it failed a prime95 test after not too long so i had to up my vcore after all, but no where near as much as another time i tried for 4Ghz without the bridge/ht adjustments...

obviously i need need to do more tests but this is working a better so far and closer to 1.26v overall:

haydent

after your comments i decided to scale my mobo chip volts back, something ive been meaning to do... im now at NB/HT/SB 1.4/1.4/1.5 and doing a running a small fft test

haydent

ran some more voltage tests:

small fft tests: ( NB/HT/SB)

1.4/1.4/1.5 - 1.30 vid - core1 prime95 error after 20min

1.4/1.5/1.5 - 1.30 vid - core1 prime95 error after 15min

1.56/1.52/1.56 - 1.30 vid - core1 prime95 error after 17min

1.4/1.4/1.5 - 1.31875 vid - passed 1hour

1.3/1.3/1.5 - 1.31250 vid - core1 prime95 error after 20min

1.3/1.3/1.5 - 1.31875 vid - core1 prime95 error after 16min

1.34/1.34/1.5 - 1.31875 vid - core1 prime95 error after 18min

1.36/1.36/1.5 - 1.31875 vid - core1 prime95 error after 15min

so i now at the one that passed 1 hour, and will run longer test nekt time i go out

haydent

things are still going well at the bold line above, just a quick note that i tried running with out the northbridge cooler fan the other night as its noisy and i was running the machine through the night. well the next day when i was using it i was getting random program crashes and bluescreens, so i put it back on and it stopped ! so as i thought and now have pretty much proved you really need that cooler to run stable at 4 if your running the mobo volts higher.