Club3d RoyalAce R9 290 very high temperatures and memory errors.

Updated on 20-06-2017 in R9 290 Series
15 on 29-05-2017

So recently my RoyalAce R9 290 started to heat up significantly under gaming loads that it did not used to have issues with, like Overwatch. It would get up to about 85-88C on the reported core temperatures and textures/effects would start to show artifacts, like neon colours on textures.

I tried checking the cooler and replacing the thermal paste, since the card is now about 28 months old so I did not think I had any other recourse. Does Club3D consider replacing thermal paste to be breaking the warranty?

After replacing the thermal paste, temperatures dropped into the high 70s for a time, but they have been climbing again and the texture corruption does still appear when the card’s reported temperature gets to around 79-83C. I ran a gaming session with HWInfo in the background and it was reporting VRAM errors during this, so I think it might be the memory that’s developed a defect?

Can anything be done at this point? As I said, the card is past its retailer warranty at this point, and was working fine up until now.

  • Liked by
Reply
7 on 30-05-2017

Hi Fraevar,

Sorry ot hear this.
85 Degrees temp for teh GPU should not be a problem in the sense that it would cause errors of any kind. It does sound like a temp issue, but more likely of the VRM and or Memory.
What you could try is to cool the memory and or VRM (better). Soem users place little heatsinks on them and make sure the airflow inside your case is good. That should help.

on 30-05-2017

Hi there, 

Thanks for the reply. I did check the various chips and made sure I re-applied thermal compound to the VRM heatsinks as well. As for airflow, the card has been mounted in a Corsair Graphite 760t with 2x 140mm fans for intake, as well as a 140mm fan in the back for exhaust and 2x120mm exhausts in the roof of the case for my AIO CPU cooler. I’m honestly at a loss as to why heat would suddenly become an issue with this setup.

on 30-05-2017

I want to point out that the card was running absolutely fine with temps ranging from 67-73 for nearly the entire 28-month period that I’ve owned it, without overclocking or even adjusting the card’s default fan curves – the BIOS switch was in the default position (towards where the display cables are plugged in).

on 30-05-2017

That should indeed be enough airflow!
I always try to look at what changed, and from the info i have that would only be Windows and or Driver updates.
Please do this:
Load a System Restore Point from before the issue started.

and /or

Make a clean dirver installation

Especially when using the upgrade feature to update drivers can give issues, because old files can remain in the registry of windows and can be “in the way”.

I have heard quite a few strange behaviors with regards to HD/ R9 series cards in combination with AMD Drivers after version 16.8.3.

Please do this:

Please uninstall the current Graphics driver and all related Graphics software like Afterburner/GPU-Z or whatever graphics related software you might have installed. Best way to remove Driver Software is to boot Windows in Safe mode, so the registry can also be cleaned. When your system does not allow you to do that, you can also use a program like DDU to remove such software: http://www.wagnardmobile.com/DDU/

Reboot and manually download and install 16.8.3 graphics driver (don’t let auto installer from Windows do that). You can find that here:

http://support.amd.com/en-us/download/desktop/previous/detail?os=Windows%2010%20-%2064&rev=16.8.3

or

a newer version if you prefer … (but please do not use any driver from 16.9.1 ~ 16.12.1 for your HD/ R9 Series Card); here you can find listing of drivers …

http://support.amd.com/en-us/download/desktop/previous?os=Windows%2010%20-%2064

 

on 30-05-2017

Hi there, 

DDU is my go-to method for updating drivers, and I did have the most recent WHQL-certified driver, 17.4.4 installed due to it being certified on the W10 Creators Update.
After a clean driver install I’ve now run a high-framerate test just now with Overwatch settings adjusted to reach about 120-160fps on my 144Hz monitor to really push the card with  the 16.8.3 driver you recommended and the temperature situation is sadly the same – almost immediately under load – even just entering the game the temperatures go up to 73C and quickly climb to 84C during actual gameplay. The VRM temperatures are 88C for VRM1 and 68C for VRM2. So I still can’t really seem to pin down the issue but I do find it rather bizarre that a card that used to run 67-73C under load now runs upwards of 84C even after having had thermal paste replaced and the cooler cleaned. I also checked that all three fans were spinning properly during this session and they were, albeit at very high RPM – almost 3400RPM at 84C, actually, which also makes the noise level really bad.

on 31-05-2017

Hi fraever,
It is not uncommon that electronic equipment will use more power as the product ages. So that it will generate a bit more heat is also not all that uncommon. However that process is normally something that goes little by little over time and not from one moment to the other.
GPU’s are protected against overheating and will switch off automatically after reaching temps of around 100 degrees. So 88 Degrees celcius for a GPU is not a temp that should cause crashes or artifacts.
Did the problems start when switching to the WIN 10 Creators Update by any chance ?

on 31-05-2017

Hi there,

I know that the R9 290 isn’t supposed to be in trouble until it hits about 95C, my concern is the memory chips, which seem to not like being that hot, even if the GPU itself does, since the texture artifacts happens when the GPU core stays at 88C for longer amounts of time, like a 30-minute competitive game in Overwatch. In that game it also seems tied to how high some settings are set, like dynamic reflections or lighting. When the GPU is at around 70-75C there are no issues, but as the temperature crosses 87C the artifacts start showing up. So when the memory has to work more with at with the GPU at high temperatures and the increased load due to additonal effects, the artifacts appear.
The issue didn’t start on the W10 Creators Update, but ever since driver 17.2.1 there have been oddities, like the AMD installer always crashing when selecting the option to reboot after installation. I’ve taken to just rebooting manually now, it’s an issue I reported to AMD when it started, but it still persists. 
I’m wondering if the only viable solution here would be to try and replace the cooler with an AIO of some kind?

on 31-05-2017

Well 3rd party coolers made for reference design R9 290 should fit this board as well.
However the 3 fan CoolStream Cooler is quite a good and capable cooler, so if you replace this, it would have to be quite a good quality cooler to get better performance,

Show more replies
  • Liked by
Reply
Cancel
2 on 31-05-2017

I’ve never had any issues with the CoolStream Cooler until these issues began. In fact, since my case is the same as the one Overclock3d used in their review of this card, I had a good temperature reference and I was getting similar results to their test when I originally got the card. So replacing the cooler isn’t something I consider lightly as it’d be something like the Arctic Accelero Hybrid III 120mm AIO cooler that I’d use instead. Which is a fairly expensive cooler, so naturally it’d be better if I could figure out why the card suddenly spiked in temperature.

on 01-06-2017

Hi fraevar,
I could not tell you what caused that sudden increase in temps.
That Arctic cooler is indeed very nice, but you are right about 110 to 115 USD is a lot of money. I mean that is half the cost of a new RX 570 (which is faster than a R9 290). 
I have nothing against Arctic, they make great products, but I dont know if i would comfortable investing that amount of money in a out of warranty product that is giving me problems …

on 05-06-2017

Seems to be really sensitive to how my case’s rear exhaust fan is spinning. If that’s at 60% the card settles at around 77C. Anyway, going to try simply using a quieter rear exhaust fan set to max, and see if that does the trick. Thanks for all the suggestions!

Show more replies
  • Liked by
Reply
Cancel
3 on 03-06-2017

Your video card is good.
I encountered same problem some days ago. Spent a lot of time to understand the issue with overheating my 270x. Sure, changed the thermal paste with no positive result. After some research revealed that AMD just stopped support 2xx series for 7.xx Crimson.  Also no WattMan there and no the auto fan control. But it could burn my video card! The thermal sensor doesn’t work properly with software and hence we have overheating every time. No notification about that from AMD. If you choose the drivers on the AMD website it proposes the latest Crimson of course and a lot of  troubles. So go back to 16.8.3 go to the past.
My conclusion is never buy AMD products again.
ps People report same problem for R-3xx series too.

on 05-06-2017

It is true that AMD has had issue with Radeon HD and R7/9 2xx and 3xx series with driver after 16.8.3. That was the latest driver made with focus on those cards. After this Drivers were aimed at RX 400 and later RX 500 series which caused several isues. Especially driver 16.9.1 until 16.12.1, like i wrote before. After this most user do  not seem to have those issues anymore. So from let’s say 17.1.1 it should be good again. But by all means I do agree that 16.8.3 is one of the best drivers i have seen for R9 290.

on 19-06-2017

Thanks. I’ve got my problem with 270x next two weeks exactly after updating to 17.1.1. At first days it was back screen blinking. Very and very short so I thought that is something with my TV. But later the black screen shown while playing 3D games and finally I got it permanent after cleaning and installing the drivers again. 
After a research I figured out that all symptoms point to a short circuit of one or two memory chips. In the case the video card is abnormal hot it’s possible to check it touching the back side of the board under the memory chips. Be careful doing that! Turn off the fan. Some area(-s) will be extremely hot after 1 minute of the computer start or even faster. 
What could call the memory chip’s fail? Many factors but perhaps some problem with VRM not working properly under new driver voltage control. If somebody was lucky and got the old drivers back soon maybe the video card components couldn’t  get a critical damage. But my advise is even the video card is still working it’s better to have look around for a good discount and buy new one. Or keep in mind a budget for replacing.

on 20-06-2017

thanks for the feedback

Show more replies
  • Liked by
Reply
Cancel