Troubleshooting hardware errors with only one PC

[DrewChan]

Relating to my BSOD with GPU -

My previous card was a 5770 I used to often get the error "Display Driver has stopped working + recovered" blah blah this error was resolved by upping the voltage from 0.950 standard to 1.2. Never had a BSOD

New card Gigabyte got the same error "Display drver blah blah" although used to just flicker for a second then correct itself - and then just died giving a constant BSOD before startup referencing nvlddmkm.sys.

Tried every fix I could find but no luck - old drivers/new drivers you name it.

Put 5770 back in and working perfectly so far (No gaming)




Takealot has agreed to take card back - as some searching also shows its very possibly a fault on the card...

How do I start troubleshooting this if it happens with next card (only have 1 pc to work with)

Definately isn't RAM as this is swapped out and never resolved error.

Which i'm guessing leaves:

Motherboard, PSU, Processor and of course the possibility of crap luck with GPU's.

PSU is a corsair 650w TX so plenty powerful enough for these cards
CPU: AMD 950 BE quad core - the heatsink is a little dusty (buying a brush tonight to clean)
Motherboard: (Still not sure the make think asus, if its not the gpu i'd guess this to be the culprit as it was the cheapest part of my rig)


So... please someone give me some guidance here as I'm about to chuck this rig and start again.

 

[MaSeKind]

Well I would say it's either the motherboard, OS (software) or the power connector from the PSU to the GPU.

1. Motherboard could have a faulty PCI-Express slot. If you have more than one you could test that (even if the other one isn't the same speed; might help in troubleshooting). Obviously this will require you to buy a new mobo (also don't diss your motherboard just cause it was the cheapest thing, it usually is )

2. What voltage did you change? Was it on the card itself, or on the motherboard (PCI-Express slot's voltage for exampke). If you tweaked mobo settings that might have caused the Gigabyte card (what kind of card is it?; I assume some NVidia chipset by the driver error) to misbehave.

3. Is the box running very hot? This sort of goes along with the point above, if you tweaked any voltage settings the system might be running hotter and is crashing because of overheating. You could check if the BIOS shows system info (including temperatures) otherwise install something like AIDA64 (http://www.aida64.com/downloads). It includes Sensor information showing temperatures from the various components in your system. Check under the first category "Computer" for the "Sensors" section.

3. A quick search for nvlddmkm.sys brought up this: http://en.kioskea.net/faq/6210-nvlddmkm-sys-bsod
Seems there are issues some people are having with the NVidia drivers (I have a NVidia card myself but haven't experienced any such issues).

Just a few other points:

- You might want to look at Windows Event Viewer. It usually logs errors and might give you a better idea of what exactly caused the crash. Unfortunately Microsoft error messages are impossible to decypher on your own but you might be able to find out more info online if the error message is a bit cryptic

- I find it a bit weird that you had to change the voltage for the 5770 to work. What card did you have before that? Also did you ever tweak/overclock anything before you got that card?

- While the PSU wattage should be more than enough for your system, if you have lots of HDD and other components you can run into issues as well

- I assume these cards were all used on the same OS (Win7 probably). It could all just be software issues (which would be good) so I suggest after you've looked at the hardware to do a reinstall (or test with a LiveCD from a OS like Ubuntu)

Well that's enough for now I think. Hopefully something here will help.

 

[DrewChan]

Thanks for the comprehensive reply MaSeKind

Voltage was changed with ATI Tray tools and reset after each restart so not on the board

Only 1 PCIe -Slot

Box is well cooled with lots of fans- although the space IS a little small - the gtx560 is quite large


The TX series of Corsair PSU's could probably run 3 of my systems - although the "connector between the PSU and GPU" sounds like a possibility.

I can try each one seperately on my 5770 as it only needs 1 x 6 pin connector

Tried various drivers - doubt its a driver issue as card was running (mostly) fine up until the point where it bombed out completely.