I made a mistake installing an aftermarket VGA cooler...

[professorbell]

Right, i run a system with an evga x58 mobo, intel i7 920 processor, nvidia 8800 gt from some main stream maker, and 4 gigs of ddr3 memory from gskill inside of an antec 1200.

 

Generally the system runs great, but recently i had been getting frustrated with graphicsperformance, my video card is overclocked from the factory and generally ran pretty hot, 65 idle and about 110 under load. I considered that too high, but i'm waiting to upgrade to a new card so i decided to install an aftermarket vga cooler to bring things down and eliminate the annoying graphics glitches that were appearing.

 

I chose the Duorb, but decided to install a set of zalman memory heatsinks instead of the stock, and to rig the stock bits (after some cutting and modification of the aluminum folds) to the smaller chipsets. I used arctic silver thermal adhesive for that and arctic silver 5 for the gpu chip itself. also, i took off the standard adhesive tape from the zalman heatsinks and used arctic silver thermal adhesive there as well. (my assumption being that would give better hold and heat transfer etc...) Got it all put together, let it set for 2 hours, put it in my computer, plugged everything in (minus the 4 pin cable that the stock cooling had been connected to the graphics card wth originally and plus the 4 pin molex connector the Duorb requires be attached to the PS) and turned it on. my mobo startes putting out a very high pitched tone, the digital temp display on the mobo starts going nuts, showing error codes i did not know and did not memorize, and temps skyrocketing. I assumed i had done something wrong, (i've done nothing but install the vga cooler) so i assume that something went wrong there. What i am wondering is (assuming my mistake has to do with thermal adhesive, possibly bridging the pins of a memory chip or otherwise, or the arctic silver five being more conductive than the stuff the Duorb came with) does anybody have experience with ths sort of failure, and if so, does it seem that meticulous scraping of epoxy from the edges of memory chips might restore functionality, or did i incidentally fry something and should at this point be looking around for a new video card/mobo?

 

also, at this point considering the possibility that there is a short somewhere in the system, i really don't want to turn it on again until i have done something to solve the problem.

 

Any help/advice would be highly appreciated.

 

B^2

[professorbell]

also, i have been up for nearly 40 hours working on this and other projects...so that may figure into the issue.

[davidst]

Well there is nothing better than having a spare video card even a pci card for testing.

I would take the hs/fan off and look if the artic silver got some where it doesn't belong.

You can run the card without memory heatsinks and even test it for 30 sec without the

gpu cooler on it. And go sleep on it for a couple of hours, 8800's are cheap

[ccokeman]

You could have a a bit of the adhesive hit something and create a short.

[Zertz]

You could have a a bit of the adhesive hit something and create a short.

 

I doubt the tape is conductive, manufacturers carelessly apply it everywhere on the memory and VRM

[scr4wl]

How much AS5 did you use?

 

One of my friends decided that the stock thermal compound for his 8800GTS wasn't good enough. He like you replaced it with AS5, but he used way way to much AS5. I'm not exactly sure if when he shoved the heat sink back on if the AS5 just got all over the PCB, or if it got to hot and started melting on to it. He insists upon the latter, but I'm inclined to think it was the first choice.

 

Any way you might want to remove the Duo Orb and check to make sure that none of the AS5 is bridging anything. Or on the PCB for that matter.

[professorbell]

ok, i ended up taking it off and cleaning the entire board...scraping thermal epoxy cleaning off the as5. i did not find a really visible short, but when i put it back in and it worked. (40 c idle thus far, and 50 under 30 minutes worth of mass effect load)

 

thanks for all the help, it was...well...really helpful.

[Boinker]

Im glad it still works. Sometimes you cannot see where but I am willing to bet you shorted something with the TC you used. Im glad all turned out well though.

[slick2500]

Did you put the stock cooler back on or the DualOrb? Must have had something somewhere and just didn't notice.

[ccokeman]

I doubt the tape is conductive, manufacturers carelessly apply it everywhere on the memory and VRM

I was talking about the thermal adhesive he used. That is conductive.

[professorbell]

i used the duorb, just i went through all of the contacts and exposed chip pins cleaning and checking to make sure nothing was bridged...i didn't seen anything that seemed like it would have caused a short, but well at the end of the process when i plugged it back in everything functioned normally. I guess the lesson is to be much more careful whenever using thermal epoxy or a bonding agent other than the factory thermal tape. i mean...i did a lot more than the manufacture intended or provided instructions for, so it makes sense that there were some issues.