Gigabyte P35 Memory Voltage - Where to start??

[Kougar]

Usually I'm answering questions, but this time I need some help as I couldn't google this one up and didn't spot anything over on V-R's site. If someone could please illustrate where I can measure the DDR2 voltages for these boards, or what I should be looking for to identify the correct measuring points then then I would be much obliged!

I have a failing kit of Crucial Ballistix RAM, and I have already learned just how wrong software and BIOS voltage readings typically are... I would like to ensure the board was not overvolting my RAM before RMAing the kit. Thanks in advance for any help!

[zsamz_]

my p35c is overvolting
@ +4 i get 2.34-2.37
@ +3 i get 2.22
i had to turn pc on to measure voltage

the pic is for p35c-dsr3 but you can always try it

[Kougar]

Those volts you mention exactly match what the software has told me... typcially it says 0.07v higher than any setting I use. But even that wouldn't be the difference between life or death of my RAM since 2.35v was my testing limit.

Using that pic it took me some work but I think I found the same analog circuit. The DQ6 unfortunately has a different power layout, and I have a big fat 8800GTS completely blocking me from the chip... time to see if I have a small enough graphics card... Thanks, will post back once I get a measurement.

Edit: I set the stock 2.2v in the BIOS, Speedfan reported 2.22v, and after I found the right circuit I got a reading of 2.28v. I shudder to imagine what the true voltage was when Speedfan reported "2.37v"... thanks again for the help Zsamz!

[GMX]

If you cant find it, best way is to poke around the fets on around the dimm area.

[K404]

from what I've found, the actual VDIMM is around 0.1V higher than BIOS setting.

Heres some pics of the DS4- the PCB is the same as the DQ6
http://www.extremeprometeia.com/foru...ead.php?t=6951

[DMH]

Yes it's true the board overvoltage the dimms but on the vcore if I put 1,325V on BIOS on speedfan it gives 1,3V it's this normal??I'm not very experienced in mods but I'm buying a multimeter to know what's the real voltage going into the CPU can you tell me where to read it??

[SoF]

yes looks like for vdimm it overvolts and for vcore it undervolts - same here.
When I boot @default vcore I will get 1.28V shown by any tool, 1.4V in bios will result in ~1.35V.
Much different from vdimm

btw: subtimings are sometimes a bit slow on Gigabyte P35 I remarked, maybe you should try to test some better subtimings with memset.

[Kougar]

Again thanks everyone! I already had updated my above post but since I am still getting replies I will say again: I set the stock 2.2v in the BIOS, Speedfan reported 2.22v, and after I found the right circuit I got a reading of 2.28v actual. Right now I think setting anything higher than 2.35v for RAM is going to seriously overvolt most RAM.

SOF or K404, is there a good measuring point so I can measure vcore? I'm still leery of poking around with metal objects inside an already crowded board...

K404, great thread you have over there! Will be keeping on eye on it to see what you find when you remove that northbridge cooler assembly. If it helps you any, at 1.4375v 3.5GHz/501FSB I only needed +0.50v MCH and had the FSB at stock. I don't presently recall the FSB/MCH voltage settings at my 100% OC, but I think it was +0.20v on the MCH and +0.15v on the FSB though. The board cannot run above 540FSB without errors even after dropping the CPU multi to 6x and going +2v on the MCH and FSB both.

I did discover something intriguing while testing what was going on with my RAM. At my 24/7 3.5GHz settings I noticed that WITHOUT any extra MCH voltage the system was Orthos Blend stable... but Coretemp would instantly BSOD the system every time when started. Back to using +0.50v on the MCH and Coretemp would never cause a BSOD... I would be interested to know if that is actually repeatable by anyone else that had Coretemp BSOD issues?

[zsamz_]

Those volts you mention exactly match what the software has told me... typcially it says 0.07v higher than any setting I use. But even that wouldn't be the difference between life or death of my RAM since 2.35v was my testing limit.

Using that pic it took me some work but I think I found the same analog circuit. The DQ6 unfortunately has a different power layout, and I have a big fat 8800GTS completely blocking me from the chip... time to see if I have a small enough graphics card... Thanks, will post back once I get a measurement.

Edit: I set the stock 2.2v in the BIOS, Speedfan reported 2.22v, and after I found the right circuit I got a reading of 2.28v. I shudder to imagine what the true voltage was when Speedfan reported "2.37v"... thanks again for the help Zsamz!

your software is correct multimeter and everest speedfan all show same voltage
you better be carefull gigibyte got xtra juice lol

[K404]

Thanks Kougar Im working on RAM settings at the mo- have you found a hole when using the 2:3 Ratio between 370 and 405 FSB? I have, and im wondering what other quirks are waiting to be uncovered properly. More details on EP when I can gather my thoughts and findings

Its harder to measure the VCore on the DQ6 as there are no empty SMC positions, but if you carefully loop a bit of wire around 1 of the choke legs and put a molex on the end (keeps a DMM probe secure) that *should* work.

I increase my VMCH by 0.2 when im pushing my luck, but I too have only got to 500-501 on the FSB. Im hoping that its my chip holding me back. Ive yet to really find the need to jack up the MCH, i'll be honest. Chipset doesnt really seem to need it?

[Kougar]

I don't recall using that ratio at that FSB range, since I stick to 1:1 when possible which gives exactly 800Mhz DDR at 400FSB. I'll admit I've seen so many reports of "holes" that I wouldn't even have a clue what they might be anymore. I simply don't recall running into any, but I never did incrementally raise the FSB until after I hit 501FSB.

If you want to check the FSB just drop your CPU multi to 6 and see what happens? Gigabyte finally fixed their F5L BETA Bios links so I plan to flash mine momentarily and see if I can find any differences.... 7 BETA Bios's, I'm sure a great deal of work was done behind the scenes even if they don't mention it. I hope.

I will say that MCH voltage on this board is rather perplexing, I've not pinned down when exactly the extra MCH voltage is needed, as I can't find any direct cause & effect with increasing it in terms of stability. I had no such issue with my old Rev 1 965P-DS3 though!