Memory issues with Asus M4A88TD-M Evo/USB3

[Trouserdeagle]

Hey folks,

I just built a new computer this week, consisting of an Asus M4A88TD-M Evo/USB3 motherboard, AMD Phenom II x4 955 Black Edition 3.2GHz and 4GB (2x2GB) of GeIL Black Dragon 1333 RAM, with an Asus HD5770 and an Arctic power 700w PSU, and have literally had nothing but problems. Since I was suffering a great many BSODs and the like, I originally submitted my query to the guys over at sevenforums.com - if you'd like an in depth account of the goings on from a "troubleshooting Windows 7" perspective, feel free to view the thread at http://www.sevenforums.com/crashes-debugging/97507-bsod-assorted-stop-codes.html- though I'm not sure how helpful that'll be.

In a nutshell, I've been having assorted memory-related BSODs since build day, starting from when I tried to install Windows 7x64 Ultimate. I'm using memtest86+ to test the RAM and so far the results have varied so much that I'm having trouble pinning this issue down to any one thing. With both DIMMs installed in DIMM0 and DIMM2 (dual channel mode) I get 60+ thousand errors, but individually, I get no errors at all. I swap the RAM around into various configurations, including dual channel mode in DIMM1 and DIMM3, and get no errors. I then put the RAM back into its original configuration and get no errors. I run the system for several hours, memcheck again - no errors. I run memtest after the system had been off for 12 hours+ and return tens of thousands of errors again.

It seemed up until today, that the system would run fine having been on for several hours - to the extent that I can play Battlefield Bad Company 2 for 2-3 hours with not a single issue - but on a cold boot, having been off for hours, I'd get error after error, and endless BSODs. Today, however, it was running for a good couple of hours and still returned errors in memtest.

I read somewhere that socket AM3 boards don't really like to run in dual channel mode at 1333Mhz, so I underclocked the RAM to 1067MHz and ran memtest again, but of course, it showed numerous errors again. I'm now on pass 2 of a test running the RAM at 1333MHz but in single channel mode (DIMM0 and DIMM1) and have yet to recieve an error.

I can't be sure if this is a result of the system now being warm enough to be stable, the system being truly incapable of running in dual channel mode at all (memory controller fault on the CPU?) or if there is just a problem with the motherboard.

I'm rapidly running out of patience, and time to RMA individual parts for detailed testing - any ideas?

I have no access to any other compatible components to test known working parts in my system, so this isn't an option either.

Is my RAM knackered? The motherboard? The CPU? Or is it maybe something else that I'm missing completely?

Any help would be appreciated.

 

[bryanW1995]

1 stick of ram should suffice for everyday usage. try cold booting it and running stability tests on it with just one stick, then run it 24/7 for a couple days, then stab test again. if you don't get any errors do it again with the other stick.

if still no errors then it's the mobo. if you get random, difficult to reproduce errors it's probably the psu.

do some reading on your mobo, I had problems for months on my p6t before reading that legacy usb support caused havoc with them; when I turned it off all problems went away.

 

[Trouserdeagle]

I have tried some memtests with a single stick and returned no errors, but like I said, this is with a fairly warm system after several reboots and some hours of use. Moving the sticks from slot to slot and in different configs seem to return various results. I'll try a cold boot memory test on a single stick tomorrow and post the results.. I obviously can't keep this kind of testing up forever, as I'll have to RMA something by the middle of next week at the latest really.

You mention the PSU - how likely is this to be causing random RAM errors? While the voltages seem to be within normal levels, I did notice a slightly dodgy kettle lead connection when I move the system around.

 

[Seferio]

If your system is stable with 1 stick at a time then it's possible that your motherboard just doesn't like both sticks together. I would try to RMA your motherboard or look into flashing to a new BIOS. If you have a friend trying your memory out in their motherboard would also help pin down the problem.

BTW are you using Auto settings for memory or have you manually set them?

 

[Trouserdeagle]

I'm using the auto settings for RAM, but they're correct.

I just ran a memtest on the same setup that was returning no errors last night, having not moved the rig all night, and it's returning errors all over the place again. I'd be willing to bet anything that if I were to rearrange the DIMMs now, in no particular order, it'll run perfectly with no errors.

I have no intention of flashing my BIOS with the system being as unstable as it is now, so that isn't an option. Particularly if I do have a faulty PSU. Neither is trying my RAM in someone else's computer.

 

[Trouserdeagle]

M'kay, so I bit the bullet and updated the BIOS to the latest version, although the BIOS update info on the Asus website is woefully vague so I can't be sure if it even adresses any issues with RAM stability.

I'm running a test on a single stick right now, but I expect it'll return no errors, as it hasn't errored with a single stick in any slot at all since these problems began. I'll test again with 2 sticks in dual channel when this test is done.

The only trouble with this, of course, is that the computer usually runs perfectly well after being on a for a few hours. I have no way of knowing if it'll be fixed until I can run a test on a cold boot.

 

[Trouserdeagle]

The PSU is so knackered that I can't even get the damn system to boot for the dual channel memtest with the updated BIOS. I just talked to the supplier and they've agreed to pick the lot up on Monday for RMA. I'll update with details once the new kit arrives, just to let you know if the problem has been eradicated or it's a compatibility issue. The latter isn't really likely, as the mobo, RAM and CPU were sold as a bundle package.

Thanks for the help.

 

[Trouserdeagle]

My suppliers have just confirmed that it was indeed 2 faulty sticks of RAM, and one faulty PSU.

Thanks again.