ATI Radeon HD5870 70°C Thermal Shutoff?

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540
Hey all. For the past 6 months I've been using an HD5870 as my main graphics card. I got it as an upgrade to a computer with an R7 240 I was given. I got it off of eBay for about $60, and so far it's been a pretty good card. It came in good condition with a new fan just installed. I plan on using my current rig until I get a new build, around June or July.

I know these cards are meant to run hot, around 85°c is normal for them from a bit of research. However, the problem is when mine reaches about 70°c, it stops working. It gets a bunch of artifacts on the screen, and then proceeds to freeze my whole system. At first I though this was because I had a good overclock on it, so I put it back to stock clocks at the problem still persisted. Because of this, I always have to have the fans up really high when I game, and it gets quite annoying. Also, if it hits even 71°c I have to restart my system.

I've taken the card apart and blew it out with an air duster. I also put a new bit of thermal compound on, yet nothing helped even a bit. Any suggestions would be greatly appreciated.

Specs:
CPU: A10-5800K @4.6Ghz
RAM: 14GB DDR3
GPU: Radeon HD5870
PSU: Corsair 600M
MoBo: Asus A88XM-A
 

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540


GPU-Z and MSI Afterburner don't seem to let me view my VRM temperatures, so I can't tell you how they're being cooled. To be more specific about my graphics card, It's a Sapphire Radeon HD 5870 1GB. It appears to be the model which uses the stock cooler, but with an overlay on it.
21161-00_HD5870_1GBGDDR5_C03_634532563133051346_600_600.jpg


 

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540


Oh, I thought they were simply called VRM Temputure, as I've seen them be called that on other cards. I'll put the card under load and take some readings, then report back here.
 

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540
I took some readings, and they topped out at about 60-65°c with FurMark running. While running UniEngine Heaven, they topped out at 52°c. I don't know what the optimal temps for the VRMs are, but I know once my graphics card hit 70°c, it shut off.
 

ShadyHamster

Distinguished
That's odd, the core and even the VRMs are able to run at 90c easily, not really recommend but it's possible.

Do you have another power supply to test with at all? Could be a power issue.

What happens if you crank the fan to 100%? I know it's going to be loud but what temps do you get then, does it manage to stay under 70?
 

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540
I ran the benchmarks again, however the fan was cranked up to 100%.

First, I ran UniEngine Heaven and while running it, I came across a strange discovery. I had GPU-Z open to monitor the temperatures while it was running, and I noticed there was 3 individual GPU temperatures. As you can see here, GPU 1 and 2 were about the same temperatures, GPU 2 being a bit cooler. Then there was GPU 3... it was running at 76°c even with max fan speed.

Next, I ran FurMark, which is much more GPU intensive. Looking at the temperatures, you can see that everything is HOT, with the fan still on max. GPU 1 and GPU 2 are in the 70°s, while GPU 3 is 99.5°c. Once it hit 100°c, it was game over. My computer immediately crashed. It seems that GPU 3 is causing all of the problems, though I don't know why or how to fix it.
 

ShadyHamster

Distinguished
I would avoid running furmark, it pushes graphics cards well beyond any normal use load and in some cases can even damage cards.

The 3 gpu temps i believe are:
1 - The graphics core
2 - The memory
3 - Pretty sure this is another vrm temp but not 100% positive.

Those temps for furmark are quite high but it shouldn't be causing the card/pc to crash.
The card could of been damaged from the previous owner overclocking it.
Have you tried underclocking it at all to see if that fixes the issue?
 

NascarBoy119

Reputable
Aug 24, 2014
37
0
4,540


The reason I believe it crashes is because the GPU 3 temperature is reaching 100°c under load with my current overclock. I've under-clocked it before awhile back to no avail, however I just under-clocked a bit from my over-clock a few hours ago, and it seems much more stable. I've noticed that it runs much cooler with lower voltage and clock speeds, so maybe that was affecting something? I haven't tested it on demanding games yet, only on Minecraft but I'll test other games in the morning, as I am currently exhausted lol.
 

ShadyHamster

Distinguished
It might be crashing because of an unstable overclock.
High temps shouldn't cause crashes, the VRMs are capable of 105 degrees, i don't recommend going anywhere near that temp though.
The card should start thermal throttling once temps get to high but it shouldn't crash.