• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD RX 6700 XT overheating

Associate
Joined
25 Apr 2017
Posts
44
Location
London
Thanks for the info and help, I think I got it fixed, after messing with the undervolt I reset that back to normal, then had a look at the fan sliders rpm vs heat. I done a manual setting for each and did a slow variation on each except the last one, I boosted that to come on at 80 d C and at 100%, on the game I was having most of the problems the temp shot straight up to 80 and fans cut in, didnt get much over 85 deg. I then went back to the VR game I was on when posting this, managed to nearly clock 2 hrs game play with no issues, the fans didnt get up to full speed and was operating around the 75 deg C mark where it has been over the last month. I guess I could tweek the fan settings abit to keep them more quiet but still do the job.. However I did notice switching from automatic to manual that the Zero RPM was enabled, so im guessing that was over riding the fans from properly cutting in when they should
Ah yeah, Zero RPM ftw! Seems that by the time the fans eventually kick in, the chip is already running too hot for them to ramp up quickly enough to cool anything sufficiently. I tried an undervolt/overclock on my Aorus Elite, and temps were way too high too quickly, so disabled Zero RPM, set the fans to 65% speed and the highest it goes is 77C under 100% load now. Thankfully the cooler is a good one, so noise is minimal, too.
 
Associate
OP
Joined
31 May 2018
Posts
74
Well, new settings didn't last long, card is crashing all the Time now :(. I've contacted the person I got the build from, he won't give me the recipt for the card and won't or unwilling to tell me where the card came from. Emailed the manufacturer, said I need to take it back to where it came from. So stuck with a dead brand new card. I've taken the pc to.a local shop see what they can see is wrong with it
 
Soldato
Joined
4 Feb 2006
Posts
3,202
Well, new settings didn't last long, card is crashing all the Time now :(. I've contacted the person I got the build from, he won't give me the recipt for the card and won't or unwilling to tell me where the card came from. Emailed the manufacturer, said I need to take it back to where it came from. So stuck with a dead brand new card. I've taken the pc to.a local shop see what they can see is wrong with it

Wait for the manufacturer to reply and see what they say but if they reject an RMA then you can remove the heatsink and repaste it. Almost certainly some issue with the paste. Sometimes the production workers forget to take the plastic protector off the thermal pad causing a bad interface.
 
Last edited:
Associate
OP
Joined
31 May 2018
Posts
74
I had a reply from saphire, I have to send the card back to the person that supplied the card. But as I say he's just running Mr around in circles, he's even tried to sell me a 6800 for £950 with return on my old card lol.

I have taken the pc with the gpu box to a local shop that do gpu repairs hopefully they may be able to sort something
 
Associate
Joined
2 Jun 2016
Posts
2,382
Location
UK
So you've got system that's just over a month old and has no warranty. I'd also get the shop to give the whole system the once over to see if they spot anything untoward.

TBH it sounds really dodgy. I take it was a private individual who built it?
 
Associate
OP
Joined
31 May 2018
Posts
74
Yes, he's a private individual but running a private business. And yes thats what I'm hoping, before the gpu was shutting down at 93 deg c, today this morning the gpu was shutting down at 75 deg c, from recommendations I kept the under volt to 1100, and the fan arc on a sharp angle so it spins up before it starts building temp.
I also found this morning just leaving windows 10 desktop running with no apps running except adrenaline, the gpu shuts down after 10-15 mins with the same message on the monitor no hardware detected. I turned off the power settings for the monitor, that seems to fix that.

I'm still guessing either a configuration problem somewhere either bios or within windows. Or the heatsink has become unglued, judging by the heat build up in such a fast time, I think this is most likely. I just have to wait a couple of days for them to get round to checking it
 
Associate
OP
Joined
31 May 2018
Posts
74
it was a full system build, generally speaking a custom build will only have a warranty with just the parts, probably with a larger company doing building I expect they would have some warranty with the build. I did see all the boxed parts so I didnt have a problem with the build, I became concerned when I too possession of the PC I didnt get a receipt. I did ask on many occasions but he kept ducking from the question. So far I have contacted trading standards told them whats happened and see if they can help. I have a very good build and pleased with it, just not the GPU working. I did show him the box for the GPU he said he thought he got it from Amazon, I said to check his purchase history. He then said the amazon driver "stole" his other PC gear, he went out and had a barney with him, then said his amazon account was terminated and bla bla bla. I think you can see where this is going.
 
Soldato
Joined
30 Jan 2009
Posts
17,185
Location
Aquilonem Londinensi
It's not a GPU core temperature issue. 90c is not going to trigger a shutdown, let alone 75c. Did you see any graphical corruption before the hard crashes? Any higher temperature readings from different sensors on the card or indeed?

I would still be suspicious of the power supply or even PCIe cables until ruled out with a different unit. You say "top of the range 850w PSU" - what make and model?

It's possible it's just a duff card, it happens, but when there's dodgy geezer prebuild stuff going on... It could be anything
 
Associate
OP
Joined
31 May 2018
Posts
74
PSU is corsair 850w. Big block. There was no graphic issues before the gpu shuts down. From thermal readings from the rest of the pc is all normal operating ranges, it's the center of the gpu, when I first started getting the issuse was about a week ago now. When I checked the heat levels the center gpu sky rocketed to 90 the outer temp was normal 60 deg c. The card would get to 90, but at 93 it just cost out. Black screen, had to do a hard reset, sometimes rebooting didn't show a video picture and had to hard boot again after leaving it off for a min or too, also at that time, when windows loaded up the amd software would not.load saying missing drivers, checking the hardware manager I had a code 31 on display driver. Removing the card from the hardware.and re detecting it it was fine and the software loaded up. Looking though posts on this issue, I read to undervolt the gpu by a 100, so set it to 1100, also set the fan curve and rpm to a hirer level, also to de activate the zero fan. This worked for several days, heat came back to normal working levels and under gaming load on gfx settings back to 75 deg c. It was.only till yesterday and today this new mapping stopped working, the gpu would just cut out to a black screen with no video detected on the monitor, a hard reboot was needed. Just using the adrenaline software stress test was enough to cut out the card. Also vr gaming and also Web browser gaming more of world of war ships on max graphics settings, 10 mins of play card cuts out. I turned off the power settings for the monitor to power down when not in use, this stopped the gpu powering down with just windows 10 running. But still gaming the gpu still cuts out. Again there is no graphicle glitches before it shuts down
 
Associate
Joined
29 Mar 2021
Posts
542
Do you know anyone with a pc you could test the graphics card in to see if the issue follows the card ?

also do you have 2 separate power cables from the psu to the graphics card ?
 
Associate
OP
Joined
31 May 2018
Posts
74
No I don't have any friends to test it, I tried asking the person I bought the build from, he just refused,

I think it has 2, I'll have to check. Its in a shop atm so someone is going to.have a look at it in a day or so.
 
Caporegime
Joined
18 Oct 2002
Posts
39,299
Location
Ireland
No I don't have any friends to test it, I tried asking the person I bought the build from, he just refused,

I think it has 2, I'll have to check. Its in a shop atm so someone is going to.have a look at it in a day or so.

The person you bought it from sounds like a right ****.
 
Associate
Joined
28 Aug 2014
Posts
2,214
Yea if this giybis operating at a business I'd report him to whatever authorities you can report him to. Refusing to give you a receipt would probably get him a good fine.
 
Associate
Joined
29 Mar 2021
Posts
542
No I don't have any friends to test it, I tried asking the person I bought the build from, he just refused,

I think it has 2, I'll have to check. Its in a shop atm so someone is going to.have a look at it in a day or so.

lets hope its something the shop can fix for you
 
Back
Top Bottom