Asus R9 290 high gpu and vrm temps.

[Killrose]

So I see an add on craiglist for an Asus R9 290 Direct CU "like new in box" for $75 near me. The guy even tells a little story how it works fine for awhile, but then will blue screen on him while playing games within 10 minutes to 2 hours, claims it is a pretty well known problem, but does not want to deal with it. And as deals on Craiglist usually go down, I meet the guy in a parking lot somewhere and hand him $75 cash.

Turns out, it is like new in box, because it is an RMA from Asus to replace another R290 he had which did basically the same thing this one did to him. He even gives me copies of RMA emails between him and Asus and some other copies of discussions on forums discussing the aforementioned problem. So now its my turn:

BFBC2 , I play one session for 10min and get a screen lock-up with audio noise. I hard boot it. I then try again. I get 28min of gameplay and get a dark screen which I CTRL+ALT+DEL to Win10 Blue screen with prompts and pick "Sign Out" and then Sign back in to normal Win10 desk top.

Warthunder, I play this for about 3hrs(?) with no problems. I am Alt+Tab'ing out frequently and looking at GPU/VRM temps, etc., using Gpu-z sensor utility. Temps are High on GPU and VRM compared to my other R9 290 with Zalman VF3000a cooler that I just removed in the same machine so I could evaluate the Asus. GPU is hitting 86c at its highest, VRM1 96c at its highest, VRM2 78c at its highest. Which is way higher than my Zalman'd 290, but nowhere near VRM max allowed temps of 110c I think it is.

I'm probably going to AS5 the GPU, and re-do the TIM on the VRM1 section with some high grade Fujipoly brand TIM material because I used it on my other Zalman'd 290 and it worked great. I would like to do something with the VRM2 location, but 78c is well below their max temp, so I guess they get nothing.

Other than that, if someone here has delt with one of these cards and can steer me in a better direction, i'd like to hear it. Oh ya, the guy ended up buying a 980ti and I do not think it was an Asus brand.

 

[Despoiler]

The VRMs AMD uses on those cards are good to 150c. None of the temps you report are out of the ordinary. Black screen is usually something to do with the VRAM. I would see if there is a newer BIOS you can put on to that card or try underclocking the memory.

 

[Killrose]

The card is stock clocked 1250mhz on VRAM and uses Hynix chips, I could drop it back a bit, but I am not seeing any sort of video corruption, however black screen could be considered corruption i guess, LOL! I realize too I should run some other benches and see what happens. Guess I should use a Gpu-z log file and run a real aggressive benchmark, I know from past experience that BFBC2 heats up things more than Warthunder does, but I cant Alt+Tab out to check temps durring gameplay like I can with Warthunder.

Also, this is not an AMD Ref card re-branded by Asus, this is their own PCB and Asus uses different VRM's than whats on the Reference cards, maybe they are better? am just concerened its the temp I cant see, like a spike in temps possibly.

Thank you for the info! I'll check bios vers too.

 

[Despoiler]

Hynix normally has the better VRAM, but they had a fire at one of their plants awhile back. A bunch of dodgy VRAM went into cards. I don't quite remember what line of cards, but I do know there were lots of black screen issues reported.

 

[Killrose]

Interesting take on the Hynix VRAM. I did notice that even with a lower memory clock on the Asus 1250mhz vs my AMD ref card with the Zalman on it1300mhz that I was getting an extra 3 FPS on the Warthunder built in tank benchmark with the Asus. I thought about why and figured maybe the VRAM on the Asus had more aggressive timmings.....so, if Hynix is having manufacturing issues and Asus is using aggressive VRAM timmings, then I could see a problem.

 

[fleshconsumed]

Sadly Asus R9 290 have pretty poor coolers.

www.kitguru.net/components/graphic-cards/zardon/asus-r9-290-direct-cu-ii-oc-review-1600p-ultra-hd-4k/2/
http://www.tomshardware.com/answers/id-2209489/safe-gpu-vrm-temperature.html

While it may be true that ASUS cards use mosfets rated up to 150C, you never ever want them getting that high. I would advise to crank up the video card and case fans to 100% no matter how noisy they might become and see if you get any lock up issues. If you don't, then you'll know it's due to crappy cooling.

 

[Killrose]

Good idea fleshconsumed, i'll crank up the fan manually and see what happens. If that cures it or has a measurable effect, i'll replace the TIM on the VRM's, add a heatsink to the (2) VRM's they left without a heatsink @VRM1 location, try to heatsink the VRM2's and replace the heatsink compound on the GPU. I'll also check the heatpipes for even contact on the GPU.

 

[Killrose]

So I ordered up my Fujipoly and plan to replace the stock TIM material at the VRM1 location with that, I will also use Zalman ZM-STG2 on the GPU, and fabricate some sorta heatsink for the VRM2 location. I was wrong I believe about some VRM's being un-heatsink'd at the VRM1 location, got wrong info off the web or misunderstood what I was reading, so the idea that (2)VRM's were left in-covered by the heatsink at VRM1 location was wrong in my first post.

I log filed most of a 6-8hour Wartunder session (stable the whole time) using Gpu-z and saw GPU temps as high as 86c, VRM1 as high as 92c and VRM2 as high as 78c, though safe temps all are much higher than my Zalman'd 290 hit while overclocked 1050gpu / 1300mem and the fans running off 7v of power. I'm going to log file some BFBC2 and see what goes on when it crashes, i'll also try a Firestrike run and maybe some Furmark. I'm almost wondering if the Asus GPU is getting under-volted, because the symptom that occurs is similar to a CPU under-volt crash while overclocking. I will try upping the GPU voltage and see what happens.

I did try cranking the fans up manually but it still crashed playing BFBC2.

 

[Killrose]

So I found out my problem with BFBC2, I had to force it to run in DX9 mode. It was at DxVersion=auto and I was running DX12. I tried DX11 and DX10, but got game lock-ups within 5-10 min or so. Once I forced it to run DX9 I ran it over an hour no problems. My Zalman'd 290 card had the same problems. Played BF1 for over 2hours, 2 runs of Firestrike benchmark and many 4minute runs of Furmark as well as the Warthunder gameplay previously mentioned and the Asus 290 is solid. The guy who sold it to me complained of bluescreens and other issues, but I have not had any other than the Battlefield BC2 issue which turned out to be a DX issue the AMD drivers don't like because I did not have these issues with earlier drivers using my Zalman'd 290.

Nice deal for $75

 

[Despoiler]

So I found out my problem with BFBC2, I had to force it to run in DX9 mode. It was at DxVersion=auto and I was running DX12. I tried DX11 and DX10, but got game lock-ups within 5-10 min or so. Once I forced it to run DX9 I ran it over an hour no problems. My Zalman'd 290 card had the same problems. Played BF1 for over 2hours, 2 runs of Firestrike benchmark and many 4minute runs of Furmark as well as the Warthunder gameplay previously mentioned and the Asus 290 is solid. The guy who sold it to me complained of bluescreens and other issues, but I have not had any other than the Battlefield BC2 issue which turned out to be a DX issue the AMD drivers don't like because I did not have these issues with earlier drivers using my Zalman'd 290.

Nice deal for $75

Glad to hear you fixed the issue. Good on you for sticking through it too. I feel like it's oddball situations like this that generate the most "AMD has terrible drivers" sentiments. In reality it's people that are not technically capable or willing to troubleshoot the issue properly.