Bad RAM? Constant BSOD.

[Tachion]

I upgraded the RAM of my 5 years old system to give it a few more months of life, and it ran perfectly for the few weeks. Now, a few times a week (especially after resuming the system from standby), it will periodically memory dump with the following screen. Also, before it happens, there are weird screen anomalies that seem to happen (a few static-like horizontal lines over the screen).

http://img262.imageshack.us/img262/9236/dsc01134xd4.jpg

I take it that the new RAM is probably bad and to blame? I would assume so, given the nature of the error, but I ran memtest86 for a cycle and it reported no errors.

Any ideas as to how I can specifically isolate if it's the memory or not, since memtest didn't pick up any problems?

 

[mpilchfamily]

Check the RAM voltages. Sometimes you need to bump the voltage up a bit when adding RAM. Otherwise tehre may not be enough voltage availible to them and it can cause errors.

 

[Tachion]

Originally posted by: mpilchfamily
Check the RAM voltages. Sometimes you need to bump the voltage up a bit when adding RAM. Otherwise tehre may not be enough voltage availible to them and it can cause errors.

Is it possible it might be the power supply and not the memory? Added a new video card at the same time, too, and it works fine without issue. I didn't think much about that until I took a look at my voltage levels, but it would make sense that the RAM is flaking out due to it not getting enough power... although even in graphics-intensive games like Crysis (which should theoretically cause it to crash if it was a power issue) it runs fine.

+12V - 11.7-12
+5V - 4.7
+3.3V - 3.2
Vcore - 1.584

Do those voltages look too far off from what they should be? I'm afraid I've been out of any serious hardware-related stuff for a number of years now, so those numbers mean little to me anymore.

 

[mpilchfamily]

The voltages look fuine but if they where taken using a software application then you can't trust those readings. You need to check voltages with a DMM.

How about listing your system components.

There could also be an issue with miss matched RAM or the new RAM requiring more voltage then the old.

 

[Tachion]

Originally posted by: mpilchfamily
The voltages look fuine but if they where taken using a software application then you can't trust those readings. You need to check voltages with a DMM.

How about listing your system components.

There could also be an issue with miss matched RAM or the new RAM requiring more voltage then the old.

Asus P4P800SE Deluxe Mobo
P4 3.0Ghz
ATI Radeon HD 3600 AGP 512mb

2 sticks of each of the following (2GB total), all in the correct channels:

PC3200 - 200mhz, CAS 2, CAS/RAS 3, RAS Precharge 3, tRAS 8
PC3500 - 217mhz, CAS 2.5, CAS/RAS 4, RAS Precharge 4, tRAS 9
Voltages for both types are 2.5V

Talked with one of my friends and changed the DRAM settings to the following since he said it could be that: 200mhz, CAS 2, RAS/CAS 4, RAS Precharge 4, tRAS 8



Debugged the memdump and got this:

MEMORY_MANAGEMENT (1a)
# Any other values for parameter 1 must be individually examined.
Arguments:
Arg1: 00041284, A PTE or the working set list is corrupt.
Arg2: 05cc9001
Arg3: 00009056
Arg4: c0503000

Debugging Details:
------------------
BUGCHECK_STR: 0x1a_41284
CUSTOMER_CRASH_COUNT: 1
DEFAULT_BUCKET_ID: DRIVER_FAULT
PROCESS_NAME: svchost.exe
LAST_CONTROL_TRANSFER: from 80529e59 to 8053767a
STACK_TEXT:
a9c7cb60 80529e59 0000001a 00041284 05cc9001 nt!KeBugCheckEx+0x1b
a9c7cb98 804f1130 00009056 05cc9000 c030005c nt!MiLocateWsle+0xc0
a9c7cbc8 804f1333 c0017324 05cc9000 00000000 nt!MiDeletePte+0x1bb
a9c7cc8c 804f14af 000000a4 05da8fff 00000000 nt!MiDeleteVirtualAddresses+0x162
a9c7cca8 805715aa 05ca0000 05da8fff a9c7cd64 nt!MiDeleteFreeVm+0x20
a9c7cd4c 804dd99f ffffffff 0411f630 0411f648 nt!NtFreeVirtualMemory+0x42e
a9c7cd4c 7c90e4f4 ffffffff 0411f630 0411f648 nt!KiFastCallEntry+0xfc
WARNING: Frame IP not in any known module. Following frames may be wrong.
0411f590 00000000 00000000 00000000 00000000 0x7c90e4f4
STACK_COMMAND: kb

FOLLOWUP_IP:
nt!MiLocateWsle+c0
80529e59 006a00 add byte ptr [edx],ch
SYMBOL_STACK_INDEX: 1
SYMBOL_NAME: nt!MiLocateWsle+c0
FOLLOWUP_NAME: MachineOwner
MODULE_NAME: nt
DEBUG_FLR_IMAGE_TIMESTAMP: 48a4044a
IMAGE_NAME: memory_corruption
FAILURE_BUCKET_ID: 0x1a_41284_nt!MiLocateWsle+c0
BUCKET_ID: 0x1a_41284_nt!MiLocateWsle+c0
Followup: MachineOwner



Also ran Memtest86+ for 15 hours with no errors reported, so I doubt the memory itself is bad.

Any thoughts? His only suggestion if it keeps doing this now that I changed the DRAM settings and also got no errors with memtest86+ was that Windows is somehow responsible.

 

[Tachion]

Anyone? It's still doing the screen thing, even on a Ubuntu (I'm dual-booting), so it's not Windows.

Only other thing I can think is it's something with the video card. I mean, it works fine in games, but I have no other idea what else it could be, especially with the identical screen staticy-lines happening in both OSes right before the crash (although I haven't been in Ubuntu long enough to witness it crash... just the static lines).

Any ideas? I'm really at a loss here.

 

[Dillybob]

I had a similar problem, which went away after replacing the RAM. Hit into the BIOS and make absolutely sure that the voltages are what they're supposed to be set at.
If they are, and you have the same problem, I'd swear it's the RAM (bad RAM might not show up right away...mine showed problems after about 2 months).

Anyhoo, hope it all works out for 'ya! Good luck to you.

 

[Tachion]

Originally posted by: Dillybob
I had a similar problem, which went away after replacing the RAM. Hit into the BIOS and make absolutely sure that the voltages are what they're supposed to be set at.
If they are, and you have the same problem, I'd swear it's the RAM (bad RAM might not show up right away...mine showed problems after about 2 months).

Anyhoo, hope it all works out for 'ya! Good luck to you.

I just removed the old modules and left the new ones in to see if it still does it. If it doesn't, I'll do the same with the old ones. Hopefully it will with one and not the other so I know which RAM it is.

Hopefully this will solve it once and for all... thanks for the suggestion!

 

[Ausm]

I would personally go into the BIOS and set the memory settings to autodetect and try again. If it doe the same thing then try running the individual matched pairs to see if you can find out which stick(s) are giving you the problems.

 

[CRXican]

I had the same error and ran Memtest one both RAM modules. One turned out to be bad.

I'm almost certian it's the RAM.

 

[Tachion]

Well, I'm almost certain it's the RAM now. I just ran an "in-Windows" RAM test that just fills up the free RAM and tests it and it's doing it again, way quicker than before (presumably because I filled it all up by running that program, even though it gave no errors).

Here's a screenshot to give a better idea: http://img376.imageshack.us/im.../2121/dsc01136axc0.jpg

Video that shows it better (yeah, it's overkill, but I really want to get this thing fixed): http://s195.photobucket.com/al...w&current=MOV01139.flv

Those colored lines just flash all over the screen and it eventually gets really bad right before BSODing. Going to test the other RAM out tonight just to be sure.

Thanks for the help everyone.

 

[Tachion]

Spoke too soon... it's doing it with each individual pair, and I highly doubt they're both bad. What else could it be? My only other guess is the video card.

I'd be tempted to just deal with it if I didn't just spend $200 on upgrades for the thing. Ugh.

 

[RadiclDreamer]

That pic and video looks like overheating/bad video card.

 

[Dillybob]

Yeah. Now that I've seen that, I'd have to agree that your RAM may still be good, but your videocard is bad/overheating. Get another fan for your case or something, or one of those spot-cool fans (Antec) to cool your videocard.

If that doesn't help, then get a new videocard.....though at this point, there's not much headroom for upgrades.

Is this your main system? Because playing Crysis on an AGP P-4 system is like.....eating a sandwich without mayonnaise.

 

[Tachion]

Originally posted by: Dillybob
Yeah. Now that I've seen that, I'd have to agree that your RAM may still be good, but your videocard is bad/overheating. Get another fan for your case or something, or one of those spot-cool fans (Antec) to cool your videocard.

If that doesn't help, then get a new videocard.....though at this point, there's not much headroom for upgrades.

Is this your main system? Because playing Crysis on an AGP P-4 system is like.....eating a sandwich without mayonnaise.

I put my old 9800 back in today and it's been running fine ever since, so it's definitely the video card. As far as the cooling goes, I have 3 case fans and it still did all that even with the side panel off. Could it really be overheating that much? I've never had a cooling problem ever before, so I'm tempted to believe the card is just bad. I also thought it might be the power supply, but I have a 430W supply, which is well above the recommended minimum.

But yeah, it is unfortunately my main system. The strange thing is, it runs Crysis, Bioshock, etc. just fine on medium settings... the glitches only appear to happen in Windows itself, so I would think if it was really an overheating issue, it would crash when running games like that much quicker.

Either way, guess I'll RMA the card and if it still does that, then it's time to work on a new system!