5970 60c at idle, 100c at load is this normal?

[travismoore]

Hi,

I bought this card second hand and I am trying to establish if these operating temps are normal. I have put new thermal paste and tape on the card in hopes of lowering the temps. They are still about the same as before I did this.

I am using Speedfan to control the speeds, at 35% fan speed idle is anywhere from 55C-65C. Usually 60+

I tested it with Furmark 15 min stress test, It was hitting 100C the whole time with fan at 100%.

Is this normal for this card or is there something wrong?


Any help is appreciated

 

[travismoore]

I would like to use a 3rd party cooler, I have a Zalman one on my 5850 and it works flawlessly. However I can't find one that will fit the 5970 anywhere, well not a new one at least. If anyone knows where I can pick an aftermarket fan/heatsink up I would love to know.

I did take the card apart yesterday cleaned it,applied new thermal tape and thermal paste. That's why I am surprised at the high temps.

 

[travismoore]

I don't think I put too much paste on, I recently reapplied the paste on my 5850 and its running great. Used the X method.

I just did a little research though and maybe the thermal tape I used was too thick. I used 1mm thick and from looking online people suggest 0.5mm will this maybe be the problem?

 

[travismoore]

Didn't realise furmark was harmful I won't use it again in the future.

@pandem0nium
So if I used that for the GPU's and replaced the thermal pads do you think it would help the temps overall? I have already cleaned all the dust out. As far as air flow goes I have good cable management and there are 7 case fans running.

I'm not really using the card for gaming I am using it for mining that is why I am concerned about sustained high temps as it will be running at high loads for a long periods of time.

Also wanted to ask is it normal at idle the the card to utilise one GPU much more than the other. I had noticed as load increases the one core gets hotter much faster than the other. Maybe due to the wrong size thermal pads effecting the contact the GPU is making with the heatsink?

 

[travismoore]

I see, is there anyway to fix that or is there an issue with the card? Will it be problematic?

 

[travismoore]

I only have chrome open with 1 tab (this forum) and GPU-Z. Its still doing the same thing, Also now the one core is sitting at 70C and the other at 65C when I am doing essentially nothing on the PC.

 

[travismoore]

Do you have more than one screen? This makes your GPU idle higher than normal (using one screen).

ahh ok, that explains it I do run two monitors.

Yep, I use it on all my gpus. Makes a good few degrees difference to the temps.
Make sure you clean the HS and core surface properly with isopropyl alcohol.

What are your load temps while mining?

It's hard to say I haven't managed to leave it running for any amount of time because It seems no matter what settings I try within 30 sec-1 min the temps jump up into the 90's

 

[travismoore]

Then id saythe cooler isn't making proper contact with the chips, problem being thicker pads.

Ok thanks for the info, is there any particular thermal pads that would be better to use. Was looking at some on ebay that is 6W/MK would that suffice?

 

[travismoore]

Stock ones where pretty bad, when I opened it up some of them came off half on the chips and half on the heatsink. Also some of them where disintegrating. From what I've found online it should be a combination of some 0.5mm and some 1mm found this diagram I am going to follow. Hopefully all the stuff will arrive Tuesday so I can get it sorted.

Thanks again for all the help

 

[travismoore]

Is the 1st gpu downclocking at all. On boot with nothing open. Or is it stuck at max clocks?

I dont have the card in right now because I thought it was best to keep it out due to the temps, will test it once its back in.

CLU is fine as long as you're not careless.

Use masking tape around the core as you apply it.

Thanks for the tip, will be sure to do that.

I had a weird issue not long ago where one of my cards usage would rocket to 100% when I'd open GPUz. Wonder if anything similar is going on here?

Will test this out when I put the card back in.

Noticed you mentioned in the OP that you're using Speedfan to control GPU temperatures. What kind of temps do you get if you leave the card to decide it's own fan speed? At least that way you'd be able to compare to reviews to see if your card is hotter, or if you've simply set the fan to a quieter and hotter profile.

Was using Speedfan to push the fan speed higher than standard. At idle (around 20%) I have been getting 60-70C but I was pushing it to 37% which was keeping it closer to 60C and occasionally below.

 

[travismoore]

use afterburner imo make a custom fan curve
Hmm neever used speedfan to set my gpu, is it same as the others that you set it as a % of 100 and it stays there? just if its set and doesnt change might be why your hitting 100c

I had it set up with a curve so it adjusts according to the curve you set. When it was hitting 100C the fan was at 100% though as some have pointed out this was probably more due to furmark being aweful. Though still getting wayy higher temps than i should have when mining.


P.S thermal stuff just arrived, will update with how it goes. Also free Haribo? lol

 

[travismoore]

Just booted it up with the card back in and the difference is quite staggering. With only GPU-Z open it was 40-42C now with some browsing in chrome its around at 50C. This is without messing with fan speed in anyway, fan can't be heard really over all the others in my case which is fantastic compared to before. Also worth noting that the temps are the same for both GPU's I guess the thermal tape was pushing them apart. oops

Is the 1st gpu downclocking at all. On boot with nothing open. Or is it stuck at max clocks?

It does appear to be downclocking, GPU Core Clock is sitting at 400mhz with just GPU-Z open and with chrome browsing occasionally jumps to 725 Mhz

CLU is fine as long as you're not careless.

Use masking tape around the core as you apply it.

Using masking tape worked flawlessly thanks for the tip!


I am going to test it out now with some mining and see what sort of temps I get under load.

Wanted to say thanks again for all the advice and help.

 

[travismoore]

Ok have set up afterburner and idle temps are now 40C with the fan pretty quiet 30-35%. Mining at 270kh/s per core holding at 80C with 60% fan will push it higher over night but got to keep it low enough so I can still use my pc. Still really happy with that.


Many thanks for the help

 

[travismoore]

Seems I have run into a further issue, I have been running it consistently mining at around 80C and since earlier but I have started to encounter some issues.

Twice so far the screen has gone briefly black then there is a pop up in from the system tray that says the AMD driver temporarily stopped working but has recovered.

Then a few minutes ago I started seeing a little artifacting when switching tabs on chrome, I have not stopped mining to work out the issue but I am still getting artifacting.

I also have a question, would it be possible to run my 5850 for my two monitors as the primary card (pcie slot 1) and my 5970 in the secondary?
Or even better 5850 for monitors still in pcie slot 2?