HP DV7 GPU appears to overheat even after cleaning and applying compound and new copper

Modify_inc

Distinguished
Sep 6, 2011
80
0
18,640
I have a HP DV7-1130US running W7 x64.

About every 10 mins, (30 mins if it has been sitting for a while), the screen starts to flicker then the laptop freezes. Couple of seconds later the screen goes black. I have to power the laptop down by holding the power button for 5 seconds.

So I feel confident the issue is with the GPU since there are obvious signs of the display acting up every time this happens. It happens also to the external monitor when I have it plugged in to the vga port on the laptop.

I cleaned and reapplied some Artic Silver to the CPU and GPU and used a shaved down pre-1982 penny for the GPU. Yet the laptop still does the same thing, so at this point I’m lost as to what could still be contributing to this. I wanted to check the temperature of the GPU but apparently this model does not have a sensor gauge for the GPU. I have tried GPU-Z and HWMonitor but no temperature results, only clocks rates etc.

I performed a clean installation Windows 7 and it went through the whole process without issue, but within about 10 mins of using the desktop, it started again. I hadn't even installed any drivers yet. So I don't think its a driver issue. There are no errors reported when it happens or when I power it back on afterwards.

Any suggestions?

Thank you
Mike
 
Since you've already been in the laptop, I'll suggest getting in there again. Look over the motherboard for any leaky or swollen capacitors or any burnt looking areas. Most likely it is the graphics card (a common occurrence) but check the motherboard first. You could try reflowing the graphics card but I would highly recommend you look into having the graphics reballed (a much better solution IMO) if the capacitors seem to be okay.
 


I have it disassembled now, and have checked the capacitors; they all look good. Regarding reflowing the GPU chip, wouldn't that only be necessary if it wasn't working at all, like the famous black screen, when you power on the laptop, but all the LEDs are lit up? I have never had that issue and I would hate to possibly make things worse by attempting something that may not be needed. Then again after rethinking that, maybe what I'm experiencing is just that, but on a lesser degree that allows me to use the laptop for a very short duration. Damn, I hate not knowing for sure. I wish there was a way to test such a situation, I mean a practical way for a end user.



 
Most commonly the solder balls don't all fail at once, one or two will get brittle and start acting up. Much of the time it isn't realized the graphics is dying until it gives up the ghost completely which is why most guides refer to dead graphics cards. It isn't necessary for the card to fail completely as it loses one, two, even three or more connections (applies to CPU's also) - it all depends on which connectors start to fail first. I've seen a few desktop systems that seemed fully functional with bent pins, there is some redundancy built in.