VGA Brand : ATi Model : Radeon 4850
CPU Brand : AMD Model : AMD|PH II X2 545 3.0G AM3 R Speed : 3000 mhz
Operation System : Win 7 64-bit SP :
Memory Brand : A-DATA 4GB AX3U2000XB2G9-2X Type : ddr3
Memory Size : 2x2 Speed : 1600
Power Supply : 650 W
My first setup with this motherboard I used these two products instead of my current ram and processor.
F3-12800CL9D-4GBNQ R Gskill
CPU AMD|ATH II X2 240 AM3 2.8G
All recommended settings for voltages and frequencies were tried while setting this up. It would work perfectly fine with 1 stick of ram, but with 2 sticks, any slot combination (1,3 or 2,4 is recommended I believe) I would get frequent hangs... where everything would freeze. The only recovery was hard reset of the system. I tried updating the bios up to f5 the most current non beta bios. This made no difference in the freeze ups. I swapped parts with my friend (vid card, power supply) and reproduced the same issue. I reinstalled windows 7 on a new harddrive. Same issue. I then assumed that it was a memory or processor issue. I RMA'd those two products and grabbed the new products that I have listed above.
I received those new products about 2 months ago(adata ram and phenom X2 processor) and they worked excellent. Fast and stable. I though that I had the issue fixed. About 2 weeks ago, I started to get hangs/freezes again. My windows experience score (maybe unrelated) dropped for all of the available scores except for hard drive. The hangs began to happen more and more frequently. I checked my bios settings and made sure everything was perfect, and it was. I checked again a day later, and bios settings were totally wrong, frequencies back down, ram timing off... I set them to the correct values (for no overclocking) 1.8v for the ram, 9-9-9-24 timing. And it seemed to clear up the issue... for about an hour.
As last ditch effort I updated to the f6 bios from your website, and tried again. No change. Hangs continued very often. This was happening between 1 min after Windows loading, to 5 hours of Windows starting... The frequency of these crashes increased over the next couple days, so I tried the old fix. I pulled out one stick of ram. System runs stable. I have tried both sticks in slot one with no issues whatsoever.
I have also confirmed that there is no heat issue... processor runs at 42C under load. Ram is cool to the touch. Very well ventilated case... Do you have any other steps I can try or does this seem to be a defective MB?
I have also run memtests, prime 95 and scandisks... everything checks out (unless 2 sticks then it freezes/hangs)
Please help! Can this possibly be OS related? I dont want to RMA if it is something that may be on the software side.
Sorry, did not make that clear... I have tried slots 1 and 2 as well... with the same issues. that is also a different motherboard, but probably similar. I am going to go home and clear the CMOS and see how I do...
I, too, am so far, confused about this all. I used to just assume that EPP and XMP were different terms for the same thing - and its support seemed, generally, 'iffy'. What I've learned is the EPP in an nVidia standard, supported by (at least) three of their chipsets; XMP is an Intel standard. JEDEC seems 'open-ended', and is not very succinct about any of it (at least, that I've been able to determine - there are thousands, maybe tens of thousands of JEDEC specs ) On my list of things to do (if my poor, dead workstation ever gets working again, and I can get off this single-screened dog of a machine...) is to see if documentation for the actual register-level operation of the two exists (nVidia seems, recently, to have gotten more liberal about documentation), and coincides... I know that reading these is a BIOS-level function, which means it is likely handled on AMDs by the (top-secret, NDA-only - and oft-repaired) AGESA start-up package. Another thing I find confusing is that, somehow, the AMD boards' BIOS' "Load" functions ('Load Optimized', 'Load Fail-safe') seem to be able to set some memory sub-timings outside the BIOS' 'allowable' range - how this can be - I dunno! Another complicating factor is that, so far as I know, no &*^%ing memory manufacturer will dare to take the step of documenting sub-timings at all, which makes the task of trying to 'tweak' the stuff nigh unto impossible, barring simple trial and error... Is it that, when they make a part number, it's the 'luck of the draw' whose controller chip they use this week, and it makes the 'sub's unstable? Or, are the controllers so sensitive to individual variations in the chipsets that there's no such thing as a stable set? If that were the case, how could any of it work, ever (unless the SPD values are all really slow to play CYA?)?? All in all, I keep swearing that one of these days, I'm gonna skip a build on my list, and invest a machine's worth of funds in a good logic analyzer so I can actually look at these issues for my own self!
BTW, the voltage variations are pretty standard for everything - the issue is identified for CPUs as Vdroop, but in any 'semi-feedback controlled' DC supply, the voltage will 'settle' at a different value for varying duties - and my luck with CPU Vdroop control functions has not been all that good - my experience, just leave it - a half-tenth high won't hurt - may even help...