New build freezes up sporadically

[SergioB]

As the title says, after my previous problems with "under-volted" memory, it now runs fine... most of the time? Unfortunately now I get sporadic system freezes. No crashes, just freezes... in XP. Sometimes after I log in, sometimes at the log-in screen. Sometimes right after I log in, sometimes after an hour. And I'm not talking the easy ones you can reset with the 3-finger salute. We're talking reset button quality - the ones "the caps-lock key doesn't do anything", "last second of sound keeps repeating" type freezes. And no blue screens either.

Other than that, I have not been able to figure out anything common between freezes. Anybody have any ideas?

System:
Motherboard: MSI 770-C45, CPU: AMD Phenom II X3, Memory: Muskin 2x2GB DDR3 1600, PSU: Rosewill 550W PSU
Video: XFX GeForce 9500GT 1GB, Case: Old Dell Poweredge 600SC case, Drives: SATA HDD & CD/DVD

 

[overshocks]

Have you overclocked your system?
Are you using the stock heatsink?

 

[SergioB]

Have you overclocked your system?
Are you using the stock heatsink?

Nope. No overclocking and using everything stock with normal settings until I know it's all stable.

 

[shortstuff_mt]

What are your idle/load CPU temps? Do you have your RAM manually set to 1600MHz with 7-7-7-20 timings at 1.95v in the BIOS? Have you run Memtest86+ overnight?

If the RAM is set correctly and passes multiple Memtest86+ passes with no errors and your temps look good then I would say your POS Rosewill PSU is causing the freezing problems by supplying unstable power to the system. There's a very good reason that nobody on these forums recommends Rosewill PSU's. They certainly don't contribute to a stable system.

 

[SergioB]

Temps were pretty good from what I remember. The RAM is set manually to 7-7-7-20 at 1.92V I think (next jump was at 2.0V). I'll run the Memtest and see what that brings up.

So as I understand it, RAM potential troublemaker #1, PSU potential troublemaker #2. Could it be anything else?

 

[overshocks]

^Well I don't think the motherboard is causing it, I suspect the PSU it's rosewill...

 

[JackNaylorPE]

Any messages in Event Viewer ?

 

[overshocks]

Any messages in Event Viewer ?

event viewer doesn't help much. it may or may not display the error messages

 

[SergioB]

Well, I finally got around to doing a Memtest. I'm currently at 23 passes with 56 errors. Failing addresses seem sporadic (from 42.3MB to 3320.3MB - that I see currently). Just for reference, here's some specs:

IMC: AMD Phenom II X3 :20 Processor (ECC : Disabled)
Settings: Ram : 800 MHz (DDR1600) / CAS : 7-7-7-20 / DDR3 (64 bits)

■WallTime: - 23:14:xx
■Cached: - 4095M
■RsvdMem: - 264k
■MemMap: - e820
■Cache: - on
■ECC: - off
■Test: - Std
■Pass: - 23
■Errors: - 56
■ECC Errs: - 0

From what I can tell, all errors occur in Tst (Test?) 5. Some passes have more errors than others. Failing Address is all over the board. Most "Good" are at 01000000, most "Bad" are 00000000, and most Err-Bits are 01000000. The "most" here means that I have noticed 1 exception, but that's about it

I guess I'll be starting the test with each module separately. Just wondering if all this error variety is indicating an unstable PSU rather than faulty memory. Any ideas?

Thanks for all your help this far.