Power supply died...collateral damage?

[looper]

Good morning...

About 2 weeks ago I had a power supply failure (started w/ random lock-ups, BSOD's, and a burning metal smell from the PS fan). PC Power & Cooling did a warranty exchange for me in two days. (It was a Turbo-Cool 510 model).

I am, however, still getting lock-ups and BSOD's...but not as often. I downloaded and used a memory testing program called "memtest". ( HTTP://HciDesign.com/memtest ). After running "memtest", the messages I got were:
"mem error detected, copying between 3f4bc05 and 3f4bb23 did not result in accurate copy".
A later "BSOD" message was:
"memory_management: 0x0000001A ( 0x0000077, 0x00000000, 0x00000001, 0x3FFFFFFF )"

I am running a different memory test now..."memtest86+" ...will take a few hours...

Questions:
1) If there is damage from the failed power supply, is it more likely to be the RAM damaged, rather than the video card?


My system:
Asus P4C800E Deluxe MB
Intel 3.2 Northwood CPU
OCZ EL-Platinum 3200 rev2 (2 sticks 1gig total) #4001024ELDCPer2-K
BFG nVidia 6800GT video card
Creative X-Fi sound card
WD Raptor 74gig SATA HD
Maxtor backup external HD
BenQ 1640 DVD
Lite-On CD
Win XP Pro
(nothing over-clocked...everything at default settings)


Help!! Thx for any input/ideas...

Thank you...

Brian

 

[Atheus]

Could be the motherboard or the memory, you'll have to try the memory in a different system to find out.

1) If there is damage from the failed power supply, is it more likely to be the RAM damaged, rather than the video card?

You didn't get any video errors right?

 

[looper]

Here are most of the typical error messages I get:

BSOD... "IRQL_NOT_LESS_OR_EQUAL 0x0000000A (0x41006047, 0x00000002, 0x00000001, 0x804F32B9 )"

BSOD... "BAD_POOL_CALLER 0x000000C2 (0x00000007, 0x00000CD4, 0xE410E205, 0xF796260A )"

BSOD... "PAGE_FAULT_IN_NONPAGED_AREA 0xC4D3EE62, (0x00000000, 0x80589DF1, 0x00000000 )"

BSOD... "memory_management 0x0000001A (0x00000777, 0x00000000, 0x00000001, 0x3FFFFFFF )"

BSOD... "nv4_disp"

BSOD... "nv4_minisys"

These happen about once every 30-45 minutes or so...

Thx again...

 

[Atheus]

Originally posted by: looper
Here are most of the typical error messages I get:

BSOD... "IRQL_NOT_LESS_OR_EQUAL 0x0000000A (0x41006047, 0x00000002, 0x00000001, 0x804F32B9 )"

BSOD... "BAD_POOL_CALLER 0x000000C2 (0x00000007, 0x00000CD4, 0xE410E205, 0xF796260A )"

BSOD... "PAGE_FAULT_IN_NONPAGED_AREA 0xC4D3EE62, (0x00000000, 0x80589DF1, 0x00000000 )"

BSOD... "memory_management 0x0000001A (0x00000777, 0x00000000, 0x00000001, 0x3FFFFFFF )"

BSOD... "nv4_disp"

BSOD... "nv4_minisys"

Wow...

OK the page faults, memory management, and bad pool caller are almost definitely memory but could be motherboard, the IRQ error is probably motherboard, and the nv4_x smells like video card.

One thing id for sure - something has gone badly wrong. Do you have another system to test all these parts in?

 

[Operandi]

Check your motherboard for bad capacitors (badcaps.net), and other damaged components. It's possible they were damaged by the failing PSU or even the original source of the problem and the PCP&C just picked up the vented gases from the caps and blew it out it's exhaust.

If the board looks fine I would next try new RAM or verify that your RAM is bad in another system.

 

[looper]

GEEZ...

"memtest86+" is running now...about an hour in, no errors yet...

The OCZ is 2 sticks (512ea)... take out one...run comp...see how it goes, if no probs with that stick, then pull it out, and run with the other one? But, could be the video card as well? Or MB? YIKES...

Brian

 

[GalvanizedYankee]

The TC 510 has all the built in protection circuits any worthy PSU should have, so you hardware is probably OK. Because of noise complaints, pcp&c down-rated the original fan. This was OK for 60% loadings but when loaded to 80%+ the unit would run hot and service life of the Teapo capacitors suffers. The 510 is hard to cool because there is very little free air space inside this jam-packed unit.

The caps to look at on the board will be 8~10mm in dia. and located next to the RAM slots, AGP slot, between the I/O plate and CPU. As noted by Operandi, http://www.badcaps.net/Ident/


...Galvanized

 

[looper]

Thanks, everyone...

I need a "Captain & Coke"... actually, three...

So, if "memtest86+" doesn't find anything...it could be...
1) MB capacitors, and/or
2) video card

If memtest finds mem errors, then need new memory... but video card and MB capacitors could also be faulty?

GEEEEEEEEEEEEEEEEEEEEEEEZZZZZZZZ............

 

[Operandi]

Originally posted by: looper
Thanks, everyone...

I need a "Captain & Coke"... actually, three...

So, if "memtest86+" doesn't find anything...it could be...
1) MB capacitors, and/or
2) video card

If memtest finds mem errors, then need new memory... but video card and MB capacitors could also be faulty?

GEEEEEEEEEEEEEEEEEEEEEEEZZZZZZZZ............

Both the motherboard and the RAM could be the cause of those error messages. The simplest way to know for sure is to try good RAM in your board or try your questionable RAM in someone else's machine.

In my experience a faulty video card will display artifacts of some kind or crash on any 3D game or application. Download ATI Tool if you want to rule out the graphics card.

Originally posted by: GalvanizedYankee
Because of noise complaints, pcp&c down-rated the original fan. This was OK for 60% loadings but when loaded to 80%+ the unit would run hot and service life of the Teapo capacitors suffers.


...Galvanized

Is it actually a fan with less top end CFM or better fan management? I was under the impression that the Turbo and Deluxe units always had relatively poor fan control since they were based on server units where noise is not an important factor. A good fan controller should be able to keep temps inline and noise low at idle and increase CFM under load without sacrificing component life.

 

[GalvanizedYankee]

Originally posted by: Operandi
Is it actually a fan with less top end CFM or better fan management? I was under the impression that the Turbo and Deluxe units always had relatively poor fan control since they were based on server units where noise is not an important factor. A good fan controller should be able to keep temps inline and noise low at idle and increase CFM under load without sacrificing component life.

I only remember because I modded a Zippy for an external 92mm fan for lower dBA with the same cfm as the 80mm that was pulled.
The original TC 510 used the exact same NIDEC 80x25, 4.68W (.39A) fan. It was later down rated to .30~.34A? I'm not positive as it's been about two years. I looked in my scrap box and couldn't find that old fan right off or I'd give a part number.
At the time, the controller was unchanged in design or thermal setting. IIRC, it was just the fan change. That unit is 12V@34A and can be pushed hard by current high-end systems.


...Galvanized

 

[looper]

Well...

Memtest86+ has been running for @ 4 hours...on test 8...no errors so far...

HMM....

 

[Duvie]

Are you running the windows version..I believe there is a dos version of that as well...the problem with the windows version is that it doesn't test all the ram...After the oS installs many things will lock up sectors of the ram. A standard boot into windwos and 256mb of the ram is locked up....This may show no errors but the errors may still happen in those areas it will not check...

ONly the dos version will thoroughly check it all....

 

[looper]

I'm not sure which ver. of memtest86+ I DL'd and used...

Following directions, I took the DL'd files...extracted them to a formatted floppy, made it the boot drive, ran the bat file, re-booted w/ floppy and ... then memtest just started running...

"A little knowledge is a dangerous thing"... that would be me....

Based on what I said, can you tell if this is the dos or win version? Sorry for my ignorance...

 

[Duvie]

OK then you are runningthe DOS version....It didn't boot into windows so therefore the resources are not being reserved and thus not check...

Are you doing the extended test? cache on? You have to hit some keys to set that up, so most likley not. The cache will check the cpu a bit.

 

[looper]

@ 6 hours 21 minutes into the memtesting... test 7, 23 "passes", says "cache on" no errors... I can't believe there are no errors so far... should I let it continue to run?

And if no errors, then open box and check capacitors on MB...and if they look fine...

Then it's probably the video card?

 

[Duvie]

I think that is enough....

 

[looper]

One of items listed in the memtest window is this...

CAS: 2.5-2-2-5 RAM timing, right? Does this look ok?

It's EL-Platinum 3200 rev 2 4001024ELDCPer2-K OCZ (2 sticks 1gig total)

Also says: RsvdMem: 864k.... is this the part of memory it cannot check?

 

[Duvie]

I dont know....Go to OCZ's website and see if that is right...Do you have tiings in the bios set to auto?

 

[looper]

yes... I just emailed them...thx

 

[Duvie]

http://www.ocztechnology.com/products/m...dr_pc_3200_dual_channel_platinum_rev_2

says that ram should be cas 2-2-2-5

 

[looper]

So... shut off the memtest... been @ 7 hours now...no errors...

I guess now check the MB capacitors...

Shouldn't be the video card?

 

[Atheus]

Check the caps - failures might not be as obvious as you think so look at badcaps.net

If you have a digital camera you could take a (good) pic of the board and let everyone have a look at it.

 

[looper]

Good idea...hope my son didn't run down my camera batteries...

 

[looper]

I don't know how to insert photos to this post...copy/paste didn't work...sorry for my ignorance...

Examined MB quite closely, and did not see any capacitors bent, swollen, or leaking...in one photo there's lots of dust...I cleaned... that's all it was...

So, it's not the memory, nor the capacitors on the MB...

So, what's the problem? CPU? Video card? ...I've had no artifacts with my son playing several demanding games(Quake 4, Company of Heroes)...

HELP...

Brian

 

[Operandi]

Try Image Shack to host the photos.

You can try ATI Tool to further test the video card, and Prime95 to stress the CPU, RAM and motherboard.