Soldato
- Joined
- 14 Aug 2009
- Posts
- 3,376
Probably, which is just stupid. But hey, premium brand and all that.
Premium drivers (aka fine wine), now premium hardware.
Last edited:
-
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 Radeon MBA 7900 XTX Junction Temp issue workaround
- Thread starter Sad Panda
- Start date
Premium drivers (aka fine wine), now premium hardware.
Hi all. I have been following this overall situation quite keenly - being an owner of the Sapphire 7900 XTX (MBA). Firstly, at stock settings on the card I do count myself somewhat fortunate that I don't have the outright TJ Temp on it hitting 110*C nor power draw throttling as soon as I go in to any game or such (its traditionally orientated in my Fractal Torrent case, with the 2*240mm fans at the bottom acting as intake to supply cooler air (as the front intake has my H150i also set as intake mind you).
However I can't help but feel a bit concerned for the performance of this card in the warmer months where at the moment I am seeing TJ Temps top out at 87-95/96 degrees when under 100% load hitting boost clocks (2.5k mhz) (with the GPU at 57-64*C - so a Delta of around 30-35*C?! - Is that even normal?! - Id be interested to see what others may get in this respect).
Are we going to see potential latent issues down the line if there is an issue with the vapor chamber in accordance with Debauers findings per the video he posted today. Whilst not apparent in my cards case now, is it otherwise an issue awaiting to manifest later. Where do we then stand now with this in mind (RMA now? Will they even accept it as a pre-emptive - would OCUK even potentially entertain an exchange (my MBA card for a AIB card and pay the diff?!) - starting to wish I had perhaps waited for the AIB Custom cards and paid a bit more and avoided this 'doubt' first time around.
However I can't help but feel a bit concerned for the performance of this card in the warmer months where at the moment I am seeing TJ Temps top out at 87-95/96 degrees when under 100% load hitting boost clocks (2.5k mhz) (with the GPU at 57-64*C - so a Delta of around 30-35*C?! - Is that even normal?! - Id be interested to see what others may get in this respect).
Are we going to see potential latent issues down the line if there is an issue with the vapor chamber in accordance with Debauers findings per the video he posted today. Whilst not apparent in my cards case now, is it otherwise an issue awaiting to manifest later. Where do we then stand now with this in mind (RMA now? Will they even accept it as a pre-emptive - would OCUK even potentially entertain an exchange (my MBA card for a AIB card and pay the diff?!) - starting to wish I had perhaps waited for the AIB Custom cards and paid a bit more and avoided this 'doubt' first time around.
Last edited:
turn off the fans at the bottom of your case and see how it copes with the temp. torrent has blowthrough design afaik so card shouldnt rely on being blasted from the bottom
I run my system in a CoolerMaster NR200P MAX so my card is oriented vertically. I've had no problems at all and my junction temperature usually stays in the mid 70s or low 80s. Just out of interest I have just run Heaven benchmark in my normal orientation and also when having my case flipped on its side - with my case on its side the junction temperature was in the high 90s and still climbing. I was really surprised to see this difference.
It really is a shame because the chip itself is pretty fantastic, and this shouldn't dissuade people from buying a 7900xtx AIB (or an MBA if/when the issues are resolved). I would even had been willing to deal with the coil whine on mine if not for the heat/throttling issue. I can only imagine how the chip engineers feel having their great work stifled by bad cooler design. And yet another thing for nV and their fanboys/girls to overblow and malign... Here come the memes.AMD have already acknowledge their looking into this issue and now with Derbauer pointing out a likely culprit I wonder how AMD will react. Most of the time when AMD have had problems it's been due to faulty software and buggy drivers I don't recall a time when they've had such a widespread hardware design fault they could result in a product recall, I wonder how they will respond (I hope they respond better then they did with the 6600XT fault that was never probably addressed besides 'get 800 watt PSU for your 190 watt card'). I wonder if these were made by PC Partner Group/Zotac?
As for mine, its on its way back. I'll likely go for an AIB model, or just sit back and have a pint with my back up card until things settle down.
The mind boggles why. But doing this there has been a benefit!!!
As I mentioned above, my deltas between GPU and TJ Max was hitting between 30-35*C on average. But running the following tests after turning off my bottom two fans this is the diff (Stock Settings on the card):
Fortnight (1440p / 4k - Epic settings in both cases) - bottom case fans on:
GPU Load: 100%.
GPU Temp (edge): 64*C
Power Draw stability: Maintained 347 Watts (no throttling).
TJ Max Temp: 96*C
Delta: 32*C
Bottom case fans off:
GPU Load: 100% also
GPU Temp (edge): 61*C
Power Draw: Stable at 347Watts also.
TJ Max Temp: 80-81*C !!
Delta: 20*C
Reduction in Delta: 12*C
Control: 1440p Rendering, Max settings with RT on - moving around and going also underneath the Central Research Area where the Mushroom/Fungus is and consistently pushes the GPU I noted) - bottom case fans on:
GPU Load: 100%
GPU Temp (Edge): 62*C
Power Draw: Stable at 347Watts also.
TJ Max Temp: 97*C
Delta: 35*C
Bottom case fans off:
GPU Load: 100%
GPU Temp (Edge): 61*C
Power Draw: Stable at 347Watts again,
TJ Max Temp: 82*C !!
Delta: 21*C
Reduction in Delta: 14*C
You get the gist in the before and after. Whilst I am pleased to see some better/improved and more sensible deltas (in my opinion) - I am somewhat surprised at this positive effect the case fans being off has had (flying int he face of my logic, I have to admit).
So what is it that was occurring do we think - were my bottom fans actually 'starving' the 7900XTX of intake air (much like if you blow in someone's face, and they can't intake adequate air normally) - or perhaps the cooler air from the bottom was messing with the vapor chamber performance in such away not allowing a proper cycle of phase change???
Last edited:
After a couple of hours testing today, running Heaven benchmark for ~30 minutes each, I found that with my card:
Junction Temperature with gpu fans facing the wall - 82C
Junction Temperature with gpu fans facing the ceiling - 98C
Junction Temperature with gpu fans facing the floor - 82C
Quite strange results.
Junction Temperature with gpu fans facing the wall - 82C
Junction Temperature with gpu fans facing the ceiling - 98C
Junction Temperature with gpu fans facing the floor - 82C
Quite strange results.
I am in the same boat as Arkie
Do I just ask for a return on this and go get an AIB model when there available ?
If your card is running fine now, no one's to say if it will have an issue later in it's lifecycle. In fact, I'd assume if it's working fine in current ambient tempuratures, it's going to work fine in summer. temps A delta of 30-35c between average and tjunction is only an issue if you butt up against the thermal limit, which your card is not by a good ways. You should consider yourself one of the "lucky ones". You're in a distinctiy different situation that other (myself included) who are hitting tjunction limit (and cooresponding throttling) within a few minutes.
If it is actually faulty vapour chamer coolers then the silver lining is that the silicon is actually working as expected so all AIB cards will be fine and only a percentage of the MBA cards will be effected. My 7900xtx MBA has no coil whine so even if I did have a faulty cooler I would be keeping it to strip it off for a waterblock.
I understand this is very poor QC on AMDs part but if your temps are in the expected range do not worry and if you are hitting 110C then RMA the card if you need the cooler to work.
I understand this is very poor QC on AMDs part but if your temps are in the expected range do not worry and if you are hitting 110C then RMA the card if you need the cooler to work.
Hi guys ,
My temps when running the card in horizontal position , are going 110 junction temp with around 59-60 degrees gpu temp , but when i am flipping the case and the gpu stays vertical i get perfectly fine temps ( 78-79 junction temp - 19-20 degrees delta) . I got no problems running it with the case flipped since i dont have any fancy rgb setup or stuff like that . What should i do ? I bought it from overclockers uk , but the problem is i am living in another country (romania) so the process of sending it back would be really difficult . Could this problem affect the gpu in the long run or as long as i am keeping it in a vertical position it should be fine ?
My temps when running the card in horizontal position , are going 110 junction temp with around 59-60 degrees gpu temp , but when i am flipping the case and the gpu stays vertical i get perfectly fine temps ( 78-79 junction temp - 19-20 degrees delta) . I got no problems running it with the case flipped since i dont have any fancy rgb setup or stuff like that . What should i do ? I bought it from overclockers uk , but the problem is i am living in another country (romania) so the process of sending it back would be really difficult . Could this problem affect the gpu in the long run or as long as i am keeping it in a vertical position it should be fine ?
If the gpu temps are within the expected range , not over 90 for junction temp at standard clocks/power limit, then there is not much to worry about.
The problem we have is that we do not have all the facts to why the vapour chamber cooler is not working as intended. Is it a structural manufacturing problem ? is there not enough liquid/gas in the vapour chamber? is the design flawed and the wicking process is not working correctly? This last one is the key one , is the vapour chamber leaking meaning the problem will get worse over time?
We currently do not know why some of the coolers are faulty. The safe option is to RMA the card and upgrade to an AIB partner model. The Powercolor Hellhound is only £70 more. If you contact OCUK for RMA discuss the upgrade and ask for the date they will have stock for the card you want, or get a new MBA and roll the dice that it has a working cooler.
It doesn’t even go past 80 junction temp . I would instantly pay an extra 70£ if that would not mean i won’t have a running pc for at least 2 weeks
. I guess i ll try to talk to OCUK and see what they have to say about this . Thank you for replying anyways !
I suspect, but am not an expert, that leaving the GPU horizontal is causing the vapour chamber to dry out due to insufficient liquid in the chamber when it was built or otherwise due to some other flow restrictions. And what I do know is that a vapour chamber or heatpipes that dries it effectively becomes useless and degrades over time.
And so in theory if you are to leave the GPU horizontal you could see the temps getting worse and worse over time - that's why it's possible that everyone will need to return their MBA cards if it's part of the affected batch(batches).
Last edited:
Unfortunately a high junction temp by itself is not the only the symptom of a broken vapour chamber. I've seen some post on Reddit of people saying: my card is fine it's only 90 odd degrees; and then in the same screenshot I can see their card is pulling 270w and the fans are at 2500rpm because to compensate for the heat the GPU is dropping voltage and ramping the fans giving the illusion that all is fine.
For reference in a reasonable ventilated case, it looks like at full load pulling it's 350 TDP in watts the 7900xtx fans should operate around 1700-1800rpm. If you find that your hot spot is 90 something but your fans are 2500rpm and the GPU can't go higher than 280w then you quite possibly still have a problem.
So probably a better test for people is to make sure that you can run an intensive game or stress test for 10 minutes and at the end of the test your fans should still be under 2000rpm and your card should still be drawing 350w
Last edited:
My 7900XTX went back with days of buying and I've already received my refund. Do to stock I bought an RTX 4080 FE for the holidays and though it's a great design despite the price that's going back too as the 2nd fan isn't working. Just twiches a bit despite GPU-Z showing a 30-40% commanded fan speed!!
Just no luck so far with the latest gen GPUs!
Just no luck so far with the latest gen GPUs!
This generation is quite a downer. Colossally overpriced, basic design issues, somewhat difficult to find stock. I don't remember a generation this disappointing since maybe 9xxx-series/Tesla and 4xxx-series/Terascale. No wonder card sales are in the toilet.
My hot 7900xtx is on its way for return now. I'll likely pick up a used 3090 at this point and sit tight until the pricing situ comes back to reality and stuff works.
Soldato
- Joined
- 30 Jul 2012
- Posts
- 2,775
People who are deciding to keep when it is ok in vertical are plonkers. Guys, its broken, send it back.
That was my plan before Christmas. Get a 3090(Ti) and suck it and see but with 2 weeks off was too tempting!