• 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

its weird that when he moved it back to vertical the issue didnt fix itself, a full recall seems a bit extreme because not everyone has the issue, unless the only ones who arent reporting issues have it vertically mounted :D

I think that also, when he says two theories remain, the other theory is that the cards that are badly affected could also be RDNA3 chips that were borderline for an XTX card but were fine for an XT card. Its possible that the AIB's screened these chip batches early on, especially those with Tiered clock speed offerings and that's why the AIB custom cards are having less problems.

From the techspot 7900XT review: "The Radeon RX 7900 XT is a product AMD doesn't really want to sell, considering they suggested that they won't have that many defective 300 mm2 GCD. As noted in the 7900 XTX review, the XT version uses the same die, but with one less MCD, 5 rather than 6. This means any GCD that can't be sold as an XTX gets binned as an XT model." https://www.techspot.com/review/2589-amd-radeon-7900-xt/

The MBA cards were built and constructed by AMD and shipped to the AIB's, there would not have been any additional QA done by the AIB on these cards on whether the chip can operate at a set voltage level at a set clock speed when running in a 7900XTX configuration.

Whilst not 'defective' these borderline 7900XTX chips need more voltage and therefore, run hotter maybe, hotter than what the cooler was designed for.
 
Last edited:
People who are deciding to keep when it is ok in vertical are plonkers. Guys, its broken, send it back.
I agree and don't understand the rational here. No one's going to buy a faulty 7900xtx for resale a year from now (and everyone is aware of the issue so it's a risk/near-certain return selling one). Even if you plonk a water cooler on it, buyers normally expect the stock cooler working. You've driven a lemon off the lot and want to keep it?
 
I agree and don't understand the rational here. No one's going to buy a faulty 7900xtx for resale a year from now (and everyone is aware of the issue so it's a risk/near-certain return selling one). Even if you plonk a water cooler on it, buyers normally expect the stock cooler working. You've driven a lemon off the lot and want to keep it?

This definitely, get them exchanged. It is rubbish to be without the card for a few weeks but in the long run it is well worth it.
 
So for those of you like me who have been having this issue, I did the following things to "fix" it:

1. Undervolted to 1120mV
2. Overclock memory to 2714 MHz (Just for performance not the fix)

My Junction Temps went from 112c to 84c max

Also, my Time Spy Extreme score improved from 12,175 to 13,280

Base system for reference is Ryzen 7900X on Asus ROG X670E Hero with 32 GB DDR5 Ram at 6000MHz

Latest drivers including Catalyst 22.12.2

No tilting the case on a corner or any other BS like that :p
Looks like my comment didn't age like fine wine after Debauer's video.
 
ok so here is my little review of the 7900xtx

Since Christmas it has been great, I have not seen any degrade in performance. Only today I started checking my temps, all seemed ok I even overclocked it a bit here are the settings:

  • Clocks, 500-3050
  • Undervolt, 1065mV
  • PowerTune, +15%
  • Memory, 2700 (standard timings)
  • Fan @ 40%

My Junc temps rarely reach 80degrees and sits around mid 70s. No coil wine at all, which am happy about as my reference 6950xt was really bad for that.

I think I can say am lucky? its vertically mounted in a nr200 case, coming from a m1 ncase case my temps have been much better. Only thing I was disappointed with was the 4k performance. treated myself to 4k monitor but wasn't impressed so that is now back in the box collecting dust and am happily back on 1440p :)
 
One thing Derbauer didnt seem to do ( I might have missed it ) is try and "reset" the potential issue with vapour chambers etc on the cards that was fine but suddenly hit the 110c issue once rotated.

I wonder if he re-rotated the card in whatever direction and left it to cool , if eventually it would settle down and potentially work ok again. Obviously not a practical day to day solution.

Peeps that arent sending them back because they are currently ok, are you not worried if you have to remove the card for maintanance and move it through whatever axis, you wont somehow bring on this issue?
 
Last edited:
One thing Derbauer didnt seem to do ( I might have missed it ) is try and "reset" the potential issue with vapour chambers etc on the cards that was fine but suddenly hit the 110c issue once rotated.

I wonder if he re-rotated the card in whatever direction and left it to cool , if eventually it would settle down and potentially work ok again. Obviously not a practical day to day solution.

Peeps that arent sending them back because they are currently ok, are you not worried if you have to remove the card for maintanance and move it through whatever axis, you wont somehow bring on this issue?
none at all, worse comes to the worse id buy another one, or put my 6950xt back in that doesn't have a home. In fairness, if they did a recall and I wait for years then their was a issue, they would have to replace it. No point causing a issue with something that isn't a issue. if its not broke not point fixing it
 
So are we saying that it's just the MBA cards that have this issue and 3rd party cards (such as the Pulse) will be fine?

Another thought here, if people with MBA cards can wait long enough, once there are no more MBA cards, if you RMA it then won't you likely get a free upgrade to a card with a 3rd party cooler as they won't be able to provide an MBA card?
 
So are we saying that it's just the MBA cards that have this issue and 3rd party cards (such as the Pulse) will be fine?

Another thought here, if people with MBA cards can wait long enough, once there are no more MBA cards, if you RMA it then won't you likely get a free upgrade to a card with a 3rd party cooler as they won't be able to provide an MBA card?
they still produce MBA cards they will just be boxed as reference cards direct from AMD. I guess the suppliers will just be able to request a shipment due to RMA backorders
 
I would have liked to see a waterblock pit on a card with high temps, at least that would rule out a fault with the silcon and leave only a dodgy stock cooler
 
I would have liked to see a waterblock pit on a card with high temps, at least that would rule out a fault with the silcon and leave only a dodgy stock cooler

I mentioned this earlier in the thread. There may not be any waterblocks currently available for this PCB but Derbauer could have put a known working cooler onto pcb with the junction temp problem. If, with the new cooler , the temps were then ok it would have proved that it was faulty vapour chamber cooler.
 
Back
Top Bottom