Radeon R9 200 series temp causing crashes?

[whwiii]

is R9 200 series temp 94C causing PC crashes?

XFX Radeon R9 290X R9-290X ENFC 4GB 512-Bit GDDR5 PCI Express 3.0 CrossFireX
power supply is corsair RM 850 plus gold

so here is the issue, my PC (windows 7 64 bit) started crashing intermittently: complete lock up have to power down.

sometimes will not restart (goes through cycle of repeated fan/lights on then stops).

recently it just completely shut down on its own.

occurs more frequently when gaming, but not exclusively.

local computer store installed latest graphics drivers and "stress tested it for 48 hours" on high graphic settings. found nothing wrong (checked hard drives).



thinking it's bad power supply but called dealer (cyberpower), support had me run CPUID HW monitor

shows graphics card temp ~70C idle,

when gaming (world of tanks >120 fps high-ish graphics settings), the card fan turns on and temp goes up to 94C

cyberpower support said the problem is graphics card, but it appears a number of cards run this hot



HW monitor also shows mainboard temp 100C

some forums suggest this may be a spurious reading

scanned with IR camera (below) measuring temps ~80C on hottest part of graphics card

nothing over 55C on the main board, suggesting this may be spurious.



so do i buy a new power supply or get a new graphics card under warranty?


thanks for reading this far



main board hottest part:

graphics card:

 

[coozie7]

More than a few monitors will show odd readings for the motherboard, particularly the auxiliary temperatures, try using another program.
I'm not very familiar with FLIR imagery, but the bottom pic looks odd; the front of the card (to the left) seems to be running at a much higher temperature than the rear (to the right), and even the VRMS, which are notoriously hot running on these cards are quite cool.
Check all the fans are actually turning, and that they actually ramp up under load (or just set them to 100% in the drivers and see that they all spin up).

 

[whwiii]

I'm not very familiar with FLIR imagery, but the bottom pic looks odd; the front of the card (to the left) seems to be running at a much higher temperature than the rear (to the right), Check all the fans are actually turning.

thanks for your reply.
here's a better side view of the card hot part appears central, dropping off to sides:

all fans turning.
AMD Catalyst control center lists target GPU 95C and maxim fan speed 55% which certainly explains the temperature. suggestions for maximum fans speed and target temp? one post mentioned 100% fan speed was not desirable?

to the initial question, is this what's causing lock ups/crashes?

 

[azzazel_99]

I would (if you are running a reference 290x) create a custom fan profile and make it spin up to cool down faster. The reference card for the 290x was a piece if sh!t literally garbage. Either purchase a aftermarket cooler for it or take your side panel off and aim a fan at your computer. The graphics card could very well be the cause as even though amd said 95c was the stock target temp its horribly bad for it. Use something like msi afterburner or something and set a temperature target of like 85c or 80c so the card will throttle itself at that temp. I know nvidia cards have a ability to throttle themselves at a target temp but im not sure if amd cards do. Either way you need to cool that thing off. Try aiming a fan at it or getting some kinda of cooler air directly to your computer with the side panel off. If this helps and stops the locking up then you know its temp related and probably the graphics card.

 

[azzazel_99]

I would set the fan speeds to 100% at 90c if it were me. I never let my cards get to their max temp. The fans will be more noisy at that setting but your card will last longer and run cooler. Try setting fan speed to like 70% or 80% max at 90c and just tweak it

 

[whwiii]

thanks for replies.
follow up:
tweaked fans and temps to no avail. switched out graphics card, still crashed
replaced power supply with fan facing down on carpet , fixed.

 

[azzazel_99]

hmmm that's odd for psu to cause temp issues with gpu. Oh well glad its fixed