Does this memory/MB problem look familiar to anyone?

[East_Green]

Back in November I built a new system. After about a week's time I began to get a variety BSs, "Page Fault in non-paged area", IRQ not equal or less", "Memory Management", etc. I got new RAM but the problems persisted. I RMAed the MB to the manufacturer but they returned it having found nothing wrong. I finally convinced them to replace it and I've had the system up and running for about a week now and have had a couple "PFN List Corrupt" BSs. This time around I've been running MEMTEST86 4.2 after each BS and even after an app like Firefox crashes. Each time Memtest finds errors immediately. I then reseat all four DIMMs and run Memtest again and it all PASSES.
My question is why would the DIMM connections get flaky so that they have to be reseated every couple days? I haven't put the guts in a case yet, the components are all just sitting on my desk. The system doesn't run at all hot and all the default voltage/clock settings are in use.

Thanks,
Dan

MB: Asus P7P55D-E Pro
RAM: GSkill F3-10666CL9D-4GBRL / 4 2GB DIMMs
PSU: Antec Truepower 650

 

[jaquith]

Welcome to Tom's Forums!

ref F3-10666CL9D-4GBRL http://www.gskill.com/products.php?index=225
1333 9-9-9-24-2N @ 1.5v

A few things come to mind, first is two sets are not '4' put anywhere to create a 4X2GB matched set. Therefore, identify to 2 original sets and install:
{1} CPU: | Set 1 | Set 2 | Set 1 | Set 2 |

{2} BIOS Load Defaults and change the following:
IMC Voltage -> 1.20v~1.25
DRAM Voltage -> 1.55~1.60v
DRAM Frequency -> DDR3-1333 MHz
DRAM Timing Control / Enter
CAS -> 9
RAS_CAS -> 9
RAS_PRE -> 9
RAS_ACT -> 24
--
DRAM Timing Mode -> 2N
Save & Exit -> Yes

{3} On the MOBO verify the following:
OV_DRAM Switch verify position:
[X | _] to the left.

{4} Uninstall all ASUS OC'ing Utilities:
- ASUS AI Suite
- ASUS TurboV EVO

{5} Update your BIOS to the latest version using ASUS EZ Flash 2 method.
- version 1602 = 1. Improve memory compatibility, 2. Improve system stability

 

[East_Green]

Well, after a week or so of testing and experimenting, here's what I've found:
I updated the BIOS which and only one default clocking setting changed from the old BIOS to the new and those defaults matched the settings you listed. I tried turning on the OV DRAM but the PCs behavior didn't change and just to reiterate that, I'd been getting blue screen system failures once or twice a day and MEMTEST86 was sometimes displaying faults when I ran it and I could cure this by shutting off the machine and re-seating the memory. A reboot after passing MEMTEST has coincided with the PC NOT blue screening. I then looked at the steps I was using to fix the MEMTEST fails and tried just re-seating one one of the DIMMs at random and...MEMTEST would pass. Then I tried not re-seating any of the DIMMs and just rebooting back into MEMTEST to see if maybe some component need to warm up first but that didn't work. If MEMTEST failed the first time it would fail every time after that on just a reboot but also if I re-seated a DIMM and MEMTEST passed, it would pass every time on a reboot. The final test was this: instead of just doing the ESC key reboot from MEMTEST I flipped off the power supply switch after a MEMTEST fail, (which I had been doing each time before re-seating the memory DIMMS) flipped it back on, booted, and MEMTEST would pass and I haven't gotten any system failures since.

So it seems like if the system is off (but the PS still on) for several hours, I get the MEMTEST failing behavior when I boot the PC but if I toggle the PS switch off-on first, MEMTEST passes and the system runs fine.

The only BIOS changes I've made were to turn off the boot screen logo and enable a RAID setup.