I'm hopeless

Here's the thing, when I play games, my pc gets a little warm. Nothing too extreme, just warm. What I do notice, is that the NB and SB heatsink gets really hot. But CPU stays under 60°C. After a while it just shuts down and then tries to restart itself, but it's failing (continuous restart every 3 seconds or so). Until I unplug the power, it tries to reset over and over.

Here are the current specs

  • I7 920 - NOT overclocked and disabled turbo with CM Hyper 212 Evo
  • GA-x58A-UD5 Motherboard
  • Corsair xms3 1600 8-8-8-24  downclocked to 1033 10-10-10-24
  • LEPA-G 750watt PSU
  • XFX 7870 With that bigass Arctic cooler (forgot the name and no OC)
  • Fractal Design R4 with a ton of fans (some Noctua)
  • an SSD and a couple of HDD's if it's even worth mentioning.

 

What I've done so far:

  • Changed the PSU. Wasn't the problem.
  • Changed the PSU's fan to a noctua. Didn't help.
  • Upgraded GPU Cooler. Now the PC keeps shutting down, just quieter.
  • Changed the mobo from a Asus P6T. Nop.
  • Got a new i7 920. Neither.
  • Tried single RAM Modules and memtested them. Also not the problem.
  • New OS install and that whole stuff. Still nothing.

 

I consider myself pretty experienced when it comes to Computers but this shit is just beyond me.

 

The real question is, should I just give up? And throw the goddamn thing away?

FML

Try unplugging the wire for front power reset button

maybe hard drive??? maybe gpu??? how about safe mode? Check your bios settings.

at this point it sounds like you practically bought a new computer, since you replaced pretty much everything in it. If it isn't working...you may as well just build a new computer with newer parts, and save yourself the headache.

 it is almost a new pc .... the only original parts are the gpu & the case (bad power reset switch?) ... my guess it is the GPU

Sounds like a GPU problem

I really doubt its the case, unless it has zero airflow, in which case I would take a side panel off and have a fan blow into it. The GPU on the other hand...if you didn't seat the cooler on it correctly it could be causing it to overheat and turn itself off to protect itself.

try you ram timing and leave it at 1600mhz, if that does not work it will be your GPU, and try only 1 Gpu for a test

Do you get any type of error message on the POST screen when the PC boots back up? anything about over temp or under voltage anything like that?

 

I only have 1 7870. The GPU mosfet part is pretty close to SB Heatsink, which is getting really hot and has no active cooling on it. Tried to use the opening in the side of the case for 120mm fan to blow some direct air on that region and put a tiny bit of insulation between the GPU and SB. But it still fails me. I'm pretty sure it's a temp related issue. I'm just not able to catch the right moment when it happens. Might be 5 min or 40 min of gaming.

And also not all games. For example I'm able to play TeamFortress2 for as long as I want but Assetto Corsa crashes on me almost every time and other graphically intense games too. Sounds like a GPU problem, but that Arctic cooler IS doing it's job (GPU core never goes above 70°C).

Obviously when I try to overclock, the crashes will be more frequent.

Ok. A little update on things.

What I think it may be, is a GPU mosfet cooling issue. Right after the PC shut down I checked the temps with my hand, and o boy was the mosfet hot.

not sure if this is a heat mark

It has a heatsink on it, catching direct airflow from the fan

I can't think of another issue because airflow is not a problem

Any DIY's on mosfet heat dissipation? :P

You may want to look at your old GPU cooler for design differences that could affect that area ? If your repairs are for educational purposes only in the end ... you may have to drop a new card in it.           c}:-(

Buying a new part would be the least of the problems. It's just that I have overtaken some of that german mentallty when it comes to repairing things, instead of buying new. Although it seems I've already bought enough replacement parts...

I ran into a very similar problem with a friends PC (similar to your rig) recently. We replaced everything. We couldn't figure out for the life of us what the problem was. To make a long story short the problem was the hard drive power connecter the PSUs that we had used were modular and when we swapped the PSU we didn't swap the power cables and that was the problem. so if it turns out not to be the GPU you may want to check the modular power connecters if your PSU has them. 

When you said hot north/south bridge heatsinks.... how hot? So hot that you cant touch? They'll get hot but never to the point of painful to touch. You can re-apply the TIM to both easy enough or use a little thermal tape.

The back of your gpu on the other hand will get uber hot, hence the success of backplates. if you have some more of those little heatsinks, place some on those 3 hotspots with thermal tape of course between.

Also try the gpu in another machine.

There is also the matter of the board getting old and with that comes flakiness of stability. I've got a few x58 boards still kicking about - half have given up the ghost and the last 2 (ud3r and a p6t i think) still work.