Motherboard situation - What do I do?

[Saxamafone]

Ok, heres the predicament I am in:

I have a P35C-DS3R Gigabyte motherboard that I purchased from ebuyer back in December. I went for it because Id read good things about it and have had no problems with it until recently.

Basically I was using some cheap Kingston Value 5400 ram until I upgraded a few weeks ago to the Patriot Extreme Performance 6400. The motherboard's auto settings were running the Patriot ram at default speed (800mhz) but with 5-5-5-16 timings which I wanted to improve. So I manually set the timings to 4-4-4-12 which is the quoted performance from Patriot but the computer would not boot. I realised I had to up the ram voltage as it was running at the board's default of 1.8v - a lot lower than the 2.2 that the ram is supposed to run at. Heres where I started running into problems.

The bios lets me increase ram voltage in increments of 0.1 so I set the voltage to +0.4 thinking that this would set the ram at 2.2v. I booted the computer but when I checked with Everest and Speedfan the readings were 2.05v and when I checked in the "PC Health Status" section in the bios it says it is running at 2.048v. I set the voltage to +0.3v just to check what the readings were and under Speedfan and Everest they were 1.97v and in bois they were 1.952v.

Is this normal? I have heard about Vdroop/Vdrop but dont really know too much about it. Could these readings be accounted for by the Vdroop of the board? I dont want to increase the voltage anymore incase I damage the ram - with +0.4v they are already quite hot to the touch.

Should I RMA this board and consider purchasing something else or get a replacement? Would these readings even register as the board malfunctioning or are they to be expected? Cheers for any help (and for taking time to read a rather long post! )

[mrt9888]

Run them at 2.2V shown in BIOS, they should be able to handle the voltage as long as you run them at their rated voltage they should be within warranty.

[Hunain]

Yes, voltages shown in Windows are somewhat less than those in the BIOS, its okay as far as they are set to a safe level in the BIOS itself.

[Saxamafone]

Fair enough, its just that to get 2.2v I will have to raise the ram voltage in the bios to +0.6v which seems a little excessive. Surely this amount of droop or whatever its called is too much?

[staffsMike]

If it runs and is stable at what your bios is reporting 2.2v I don't see the problem.

Run something like orthos to test stability.

[Thorsson]

I certainly wouldn't advise putting +0.6v. You'd probably get away with it, but why take the risk if your system is running OK? Some RAM is very sensitive to over-volting. I had some Crucial that got blown because my motherboard was over-volting by 8%.

Maybe yours is under-volting, or maybe it is under-reporting. If you can measure the currently accurately with an outside device then you'll be able to make an informed choice. Otherwise if it ain't broke, don't fix it.

[Saxamafone]

Thanks for the advice everybody

Thorsson I think youre right - I really dont fancy setting it at +0.6v.

I think I may go ahead with the RMA anyway and pick up something like the Abit IP35 Pro or ASUS P5K Premium instead. What do you think?

[Max Tractor]

I would not trust any software reading of memory volts in windows, you should go with what you set in the bios, unless there is a known bug/issue then go by bios settings. Vdroop is related to cpu, what would be your grounds for rma, personally I would set the mem volts in bios and then run Memtest86+ v1.70(dos) and then memtest for windows and take it from there

[Saxamafone]

Grounds for RMA would be that the bios readings of the voltage compare rather unfavourably to the settings they should be at which gives me no confidence in the hardware at all. Being able to adjust voltages is a feature of most, if not all, modern boards and I should be able to do this without guess work and the worry of putting too much voltage through my ram (and possibly destroying it).

I will do the tests with memtest though and see if it is stable at +0.3v and +0.4v and go from there. If there is no problems then I will think again about RMA'ing.

[Blitzen]

Grounds for RMA would be that the bios readings of the voltage compare rather unfavourably to the settings they should be at which gives me no confidence in the hardware at all. Being able to adjust voltages is a feature of most, if not all, modern boards and I should be able to do this without guess work and the worry of putting too much voltage through my ram (and possibly destroying it).

I will do the tests with memtest though and see if it is stable at +0.3v and +0.4v and go from there. If there is no problems then I will think again about RMA'ing.

????
There isnt any grounds for RMA in what you are reporting. Its a bit of droop in the voltage thats all and is par for the course on many many boards.
Whatever software you use, its not going to be as accurate as the bios.
Mike was right. Set it to the required V's in the bios and forget about it.

[UseItNow]

Siblings PC I built in August/September is running with the P35C-DS3R using Crucial Ballistix PC2-5300 (D9 Micron - Crazy stuff!), like yourself, after installing found that it had set itself to PC2-5300 at 1.8v with 5-5-5-12 timings. Now the memory runs at a stock voltage of 2.2v for PC2-5300 with timings of 3-3-3-12 but I have it set at PC2-5300 with 1.9v having timings of 3-3-3-9, am pretty sure I could tighten it further but having tested with Memtest and Orthos, everything is stable. Touchwood, it will continue to be the case.

Good luck, how is everything running?

[Saxamafone]

Ok here are the bios readings for ram voltage at the various different settings:

+0.3v = 1.920v
+0.4v = 2.000v
+0.5v = 2.032v
+0.6v = 2.112v
+0.7v = 2.192v

0.7 is the highest setting allowed by the bios which should be taking the voltage to 2.5v. As it is it doesnt even reach 2.2v, although it does get very close. Surely 0.3v difference in droop should be a concern?