Losing monitor connection, overheating/PSU failure?

[Kine7ic]

Hi guys,

I recently upgraded my Inspiron 545s with a new case, power supply, and graphics card about 2 months ago. Here are links to the exact items I got:

Case - http://www.newegg.com/Product/Product.aspx?Item=N82E16811156226
Power Supply - http://www.newegg.com/Product/Product.aspx?Item=N82E16817171060
Graphics Card - http://www.newegg.com/Product/Product.aspx?Item=N82E16814150540

They all worked fine but there was one strange problem, whenever I bumped the computer with my chair or foot, the monitor would lose connection and go black and the fans (not exactly sure if it's the graphics card fan or CPU fan) would spin up to 100% speed. The only thing I could do was to hold the power button and reboot the computer.

Then a few weeks later this same thing started happened when gaming or watching streams (instead of just from being kicked) and began to happen more frequently, sometimes even when idle.

I assumed it was overheating so I installed a PCI expansion fan to suck out hot air from the bottom and I also installed MSI Afterburner to increase my graphics card fan while gaming. I also cleaned my CPU socket and fan and re-applied thermal paste. This all seemed to work for about a week but then it started happening again. Now more and more frequently.

Just today, I woke my computer out of sleep mode and checked my temps. The temperatures were lower than idle since the comp had just turned on (CPU was mid 30s and graphics card was around 40 degrees). Right then it happened again so I'm almost sure it's not overheating. Now I'm thinking it's probably a power supply issue. Anyone have any ideas?

Thanks

 

[El Tigre]

Make sure the PSU is properly mounted as when it is not properly mounted, even a small bump could cause it to act funny. You can also run a stress test to see if your GPU and such is to blame. Furmark is a good stress tester. Also, Kumbuster that comes with MSI Afterburner is another good alternative. I would leave the computer for about 2 hrs and then see the results. You can monitor the temps along the way to see if they are truly overheating.

 

[Kine7ic]

Thanks for the reply. The power supply seems to be connected properly. I also reseated the graphics card and power connection.

I ran Furmark for about 30 minutes. My gpu temp stabilized at 85 degrees and everything worked fine. It never got higher than 60s when gaming during the previous crashes.

One other thing I forgot to mention was that this doesn't happen when I take out the graphics card and run on-board graphics.

 

[coozie7]

You have a loose connection somewhere, mate, and it sounds like the card is not seating fully in the PCI-E slot.
I've had this problem before and the solution was to...adjust the PCI-E bracket with a pair of pliers so the card could seat fully.