A8N-SLI Deluxe: Memory Voltage

[Insomniak]

Ok, not sure if this belongs here or in the general hardware forum, but here goes.

Using Mushkin Basic Green in my system, 2x512 in dual channel. It's rated, according to Mushkin and Newegg, at 2.5v. When I set it to 2.5v in the system and then do some mild overclocking (11Mhz on the FSB/HTT) the system gets a boot failure.

Now, here's the odd part - when DIMM voltage is left on auto in the BIOS, the BIOS feeds the RAM 2.75v, even though it's supposedly rated 2.5

Does this mean the chips are excessive power hungry, or perhaps overdraw to ensure stability?


It is worth noting that even with the chips drawing .25v aboive rating, the system primed for 9 hours without a hitch, and I've run 4 memtests to 100% coverage with 0 errors, 0 warnings. The RAM is obviously not having any problems - I just wonder why it wants so many volts.

Any ideas.

 

[DerKaiser]

I have had to raise the memory voltage to 2.85 to even get to 210 Mhz FSB at 1:1. This is with 2x512 Crucial PC4000. To get to 225-230 FSB requires relaxing memory timings.

I don't know about your particular issue, I only know from my limited experience with the A8N-SLI Deluxe that voltages are seriously messed up. My guess is that the 2.5v setting is feeding significantly less voltage than that.

 

[fishmonger12]

i run my ram at 2.7 volts... corsair valuselect. On some boards, 2.5v isn't enough for running the ram at stock speeds. much less overclocking.

 

[JMag]

I would be interested to see what voltage your ram is actually getting. Do you have a multimeter?
For me the BIOS actually is understating the true ram voltage... If I read voltages form the MOSFET vs Settings in the BIOS here is what I get:

BIOS 3.0V = 3.18 @ MOSFET
BIOS 2.9V = 3.05 @ MOSFET
BIOS 2.8V = ~2.96 @ MOSFET

 

[Insomniak]

Interesting.

Well, the OC run continues. I'm walking my A64 3000+ Winchester (native 1.8Ghz) on up.

I dropped the HTT multiplier to 4X and the RAM clock to DDR333, and started walking. 1.9Ghz primed fine for an hour, ran 3dmark03 through fine, and memtest cleared. Temperatures never broke 40C. I just finished 2.0Ghz and everything went fine again, this time load temperature was 38C. As I go higher, it's gotten cooler running.

My theory on this is that the .09 processors don't need as much voltage as AMD says. When left on auto, my board feeds my proc 1.4v, and it slammed off 9 hours of Prime95 last night no problem, load temp 39C. I raised it to the rated 1.5v this morning to start OCing, and my temps went up at stock, and have been decreasing as I've driven the frequency up. I think this means the processor is now using the "extra" voltage it's being fed, so less is dissipating as heat.

Memory is still at 2.75v, timings 3-4-4-8, and things are going fine.

My target is 2.2Ghz - I wanted a 3500+ but didn't have the $$, so I went with the next best thing. If this processor does 2.2Ghz stable, I'm peachy.

 

[DerKaiser]

I had to switch to 2T to get over 225 FSB at 1:1. After I find the upper limit of my CPU and the FSB, I'll try underclocking my memory to achieve 1T and compare performance. My guess is that it won't be that much different. Right now, I'm testing my CPU (3200) at 2.35 Ghz, FSB at 235, memory at 1:1, 2T, 2.5-4-4-8.

 

[Insomniak]

Well, my OC marches on, and things are looking good. Target speed of 2.2Ghz is good to go. Here's where I'm at:

Processor: Athlon 64 3000+ Winchester (native 1.8Ghz)
Total Core Speed: 2.206Ghz
FSB/HTT: 245Mhz @ 4X multi
Memory: Mushkin PC3200 @ 408Mhz, 3-4-4-8
Vcore: 1.5v
Vdimm: 2.75v

Thing primed for four hours, ran through 3dmark03, memtest cleared, and did two hours of Half-Life 2 at max settings (except water at Reflect World) and 4xAA, 2xAF. Rock solid.

Want to know the real kicker? This thing has NEVER BROKEN 40 DEGREES C! Even under full loud for hours by prime, it's never topped 40C. These new Winchesters run cool as the dickens.

Anyway, tomorrow I'm going to see how high I can take the memory, and I'll also start working with the GPU.

So far so good!

 

[DerKaiser]

I'm testing at 2.5 Ghz now. Every time I up the speed, I think, well that's it, I won't get any more. None of the previous AMD chips I've owned have been good overclockers. I think you should try and up your multiplier and see whan happens.

What kind of HSF have you got on that thing?

 

[Insomniak]

Originally posted by: DerKaiser
I'm testing at 2.5 Ghz now. Every time I up the speed, I think, well that's it, I won't get any more. None of the previous AMD chips I've owned have been good overclockers. I think you should try and up your multiplier and see whan happens.

What kind of HSF have you got on that thing?

You mean my chip multiplier or HTT multiplier? From what I understand, getting total HTT over 1050 or so causes stability problems, and if I went to a 5x HTT, I'd be running somewhere around 1250.

And yeah, I'm really surprised at how great these chips are for overclocking as well.

Right now, I'm on the stock HSF that shipped with the processor. I plan to move up to a Zalman 7000Al before too long, but since I'm not even breaking 40C with a 400Mhz OC, I'm not sure that's neccesary!

So far my idle temp is 29-30C, and my load temp has never gone above 40C.

 

[DerKaiser]

I meant your CPU multiplier. I just keep raising mine and it just keeps going and going and going......

I have my HTT multiplier at 4x.

 

[Insomniak]

I thought CPU multipliers were top-locked on the A64....to be honest, I havent checked it out.

 

[DerKaiser]

Yeah, you're right, I was thinking you had a 3200. Guess you'll just have to keep raising that FSB.

 

[Insomniak]

Originally posted by: DerKaiser
Yeah, you're right, I was thinking you had a 3200. Guess you'll just have to keep raising that FSB.

Actually, I'm going to hold here for the time being...there really isn't anything out there right now that a 3500+ can't handle on max settings...if anything is a bottleneck in my system right now, it's the 6600GT, but I got that to hold me over until R520/NV47 - when they come out, I'll see if I need to move up, or wait for the generation after that.