Ram isn't stable at rated speed

I'm having some issues getting the system in my sig to be stable at DDR500. A few months ago, I was running an OC of 2.8ghz @1.4v,with the ram on a divider for an effective DDR520 with 3-3-3-8 timings(stock are 3-4-4-8, but it was 24hr memtest and orthos stable). So, I dropped my overclock because of power consumption issues, and the fact that I wanted to run the CPU fanless.

I was able to get 2.2ghz out of the chip at 1.15vcore (220x10). Then I decided why not try bumping the ram up too, so I attempted 250x9, keeping the ram at 1:1. Here's the setup,
CPU: 2.25ghz @1.35v(it can do up to 2.6ghz on stock vcore)
Northbridge: 1.9v (stock 1.8v)
HTT multi - 4x
Ram: 2x1gb DDR500 3-4-4-8 2.8v(these settings are all stock)

I can boot to windows, type this message, watch a video, but as soon as I stress test the ram, or try a 3d app, the system either locks, or the program gives an error. I can use Orthos to stress the CPU either by StressCPU or by Small FFT's, but once I try a blend, or a Large FFT test, Orthos crashes. Does anyone have any suggestions?

Try bumping up the vdimm a notch. Could be the ram is being undervolted. If this doesn't help, RMA the ram.

Sometimes a simple fix of adding a bit more voltage to the ram works, as the board may be undervolting. This sometimes happens.

Nebulous said:

Try bumping up the vdimm a notch. Could be the ram is being undervolted. If this doesn't help, RMA the ram.

Sometimes a simple fix of adding a bit more voltage to the ram works, as the board may be undervolting. This sometimes happens.

Click to expand...

The bios and Everest say the ram is being given 2.8v, but I guess I could give it more, is 2.9v safe?

Yup, 2.9 is safe. Try and see. Again if this doesn't help, i suggest RMAing the ram.

2.9v seems to be working. Thanks for the advice. I would have tried it eventually, but I just assumed it should run at 2.8v, since that's their rated voltage. Maybe the board does undervolt and the voltage sensors aren't letting me monitor that.

I guess i spoke too soon. DDR500 was stable with 3d gaming, so I decided to be sure, I'd stress test with orthos overnight. So, started a blend test, and after 26 seconds, it crashed. I then thought, maybe it's my fsb that's unstable, or even the simple 250mhz overclock that was causing the issue. What I like about my bios is that it lets me use a divider for DDR500, I don't even have to overclock the fsb, just set the divider to DDR500, and that's that. So, I did that, rebooted to windows and decided to try HCI's windows Memtest program. I open two instances, as it suggested, and it immediately crashed, causing a reboot of the pc. So, I guess I'm going to have to call Crucial tomorrow. I hadn't even owned these modules for a year, and I hardly ever pushed them beyond 500mhz DDR. Oh well....

It could also be that your IMC isn't receiving enough voltage, Remember that the IMC and the Core are on the same power plane. You may need to up the voltage to get it to be stable. I remember with one of my sempron's, to run DDR500 I needed to actually increase the CPU voltage even when the MHz hadn't been touched .

Moto7451 said:

It could also be that your IMC isn't receiving enough voltage, Remember that the IMC and the Core are on the same power plane. You may need to up the voltage to get it to be stable. I remember with one of my sempron's, to run DDR500 I needed to actually increase the CPU voltage even when the MHz hadn't been touched .

Click to expand...

I would, but I'm running the CPU fanless, any more volts through it, and it might not be thermally stable. I really find it odd that I would need more power through the chip, but I've experienced things just as strange.