RX 7900 XTX reference at possible vapor chamber design problem + very high hot spot (110c)

[Falck]

You're experiencing the same issues? How much does the card clock down? Do you have 3DMark? If so, can you run a Time Spy (or TS Extreme) stress test?

I'm not at home atm, but I'll download it when I get the chance.

Re troll.

Wtaf I said earlier I niether agree or disagree.

But I don't think his Research is extensive enough and absolutely not conclusive.

And as a test engineer who's scientific based proof backed testing has passed millions of pounds worth of analytical instruments.

Buying three broke ones the extrapolating to this degree is lamentable.

And yes I would tell der8aur to his face he's searching for clicks.


Also regarding your butt hurt for der8haur, I like the guy and his content, that doesn't mean I need to back every word he says despite buying a case with his name in it too.

And people actually working in statistical analyses earned they're qualification too with years but der8aur didn't, few have yet many think they are good at it well, I disagree with that thought too.

You're accusing him of buying faulty cards knowingly and crying wolf for clicks, when that's really not the case as per his own post later.

You did tell him your opinion, and when he's answering your accusations you're goalposting him with new accusations? Moving from his testing method, to conclusion, to "how many really is there?" and "Wheres NY times in this manner? It's not real until AMD goes bankrupt!" -statements (overexaggerating, I know, just for fun)

You can always question everything that's done or said.

How many faulty cards are needed for you to be happy with a conclusion?

 

[Deleted member 57642]

With a revenue of 23.5 billion $ for 2022 - AMD should be able to afford doing the right thing (recalling all the card with this issue - or at least - supplying local services with functional coolers - while supporting the costs for replacements & shipment). Still... it's a costly and time consuming process. So, are they going to do what it takes or... keep pressuring the support department - to act ignorant and bluntly lie their customers (acting as if all is fine - aka - working as intended)!?

I'm mostly a practical individual (even pragmatic to a large extent) - but this situation kinda resembles the western view on Karma (bumping into a wall - while laughing at the competition). So, hey... let's call i coincidence. Still, like most people - i do hope they'll chose to walk the ethical path. Even tho, i don't own this card - so i'm not affected " currently (in present time)". I mean, let's face it... their current attitude regarding this issue - and especially if they keep at it.... can have long term repercussions. Definitely that can be brushed under the rug - even whit the release of a better product (the fear of a possible issue with the product & get stuck with it broken - like now - with no support from AMD... is bound to hang above their heads).

Just my 2 cents on this topic - after seeing der8auer EN's YouTube video (clear proof of an actual issue with this line of products - empirical evidence - if you will).

 

[TumbleGeorge]

Does anyone work for free for public benefit by his own choice?

 

[OneMoar]

the vapor chamber manufacturing process is automated

you input a CAD design it stamps out or laser cuts some pieces they are either soldered by machine on the same line or taken to anouther line for assembly and filling

remember these need to be sealed and filled with a liquid that boils at near room temps.
nobody is going to open one on the line and go hey thats not right once they are filled
any design revisions need to come from the engineering people that designed the thing

How many dies do you think amd made?
How many cards use the same cooler p/n ? (hint its near every reference card)
How many coolers got off the assembly line and put into que to become cards ? (hint its a fuckton)

even if they caught it in meer hours those vapor chambers where already assembled boxed up and send out

the fact that you are seeing ANY amount of end user reports indicates that a significant quantity made it by all the checks
assembly line assembles logistics logisticify, shipping companies ship
the engineers job was done way before.
barring any sort of massive tooling issue which you would know right away on any modern line. this was a engineering level fuckup

mechanic: hey boss this part doesn't look right
boss: engineer says its fine
mechanic: Alright *assembles anyway*

 

[TheoneandonlyMrK]

You did tell him your opinion, and when he's answering your accusations you're goalposting him with new accusations? Moving from his testing method, to conclusion, to "how many really is there?" and "Wheres NY times in this manner? It's not real until AMD goes bankrupt!" -statements (overexaggerating, I know, just for fun)

Wtaf are you on about, put your spoon back in your pocket.

I've been consistent it's the conclusion that I at this moment disagree with, and await further proof from a third reputable person since I do count der8aur as reputable his opinion and evidence is valid just to me, not conclusive.


Or welcome to ignore because your striving again, to put words in my mouth.

You can witness my convo with him, am I dodging, no.

Oh, and more than five examples, 4 of which cherry picked, so we're clear.

 

[eidairaman1]

With a revenue of 23.5 billion $ for 2022 - AMD should be able to afford doing the right thing (recalling all the card with this issue - or at least - supplying local services with functional coolers - while supporting the costs for replacements & shipment). Still... it's a costly and time consuming process. So, are they going to do what it takes or... keep pressuring the support department - to act ignorant and bluntly lie their customers (acting as if all is fine - aka - working as intended)!?

I'm mostly a practical individual (even pragmatic to a large extent) - but this situation kinda resembles the western view on Karma (bumping into a wall - while laughing at the competition). So, hey... let's call i coincidence. Still, like most people - i do hope they'll chose to walk the ethical path. Even tho, i don't own this card - so i'm not affected " currently (in present time)". I mean, let's face it... their current attitude regarding this issue - and especially if they keep at it.... can have long term repercussions. Definitely that can be brushed under the rug - even whit the release of a better product (the fear of a possible issue with the product & get stuck with it broken - like now - with no support from AMD... is bound to hang above their heads).

Just my 2 cents on this topic - after seeing der8auer EN's YouTube video (clear proof of an actual issue with this line of products - empirical evidence - if you will).

With all corps they will go with the most cost effective measure whether its to recall batches or send out revised heatsinks with better thermal pads/pastes for competent techs to disassemble and repair.

Back when the 9700 Pro came out I didn't rush out to get it, I got it many months later. Case Is I never buy brand new hardware day 1.

I Remember XFX having to revise their Heatsink 3 times in 2019/2020 (Thicc) on their AIB cards.

Yet MSI never did (Evoke)

 

[AusWolf]

A "bad batch" would be even worse. If you compare that RTX40 gen was ~150k units for first batch. Obviously AMDs market share is lower so the amount of sold cards is also lower.
It's pretty safe to say that all cards in the current market are from the first batch as at the time shooting the video that card were out for 10 days.

But again, how is this not conclusive? If you are able to replicate a problem under the same condition every single time then it is conclusive. As an engineer you should know that. As an engineer you should also know that 10 is not a sufficient sample size to determine how many are affected.

Buying faulty cards and concluding that they're faulty actually concludes nothing. As said above, I believe your findings, but I want to see them confirmed by other sources as well before believing that this is a wide-spread issue. You cannot extrapolate from a handful of (previously known to be faulty) samples.

With a revenue of 23.5 billion $ for 2022 - AMD should be able to afford doing the right thing (recalling all the card with this issue - or at least - supplying local services with functional coolers - while supporting the costs for replacements & shipment). Still... it's a costly and time consuming process. So, are they going to do what it takes or... keep pressuring the support department - to act ignorant and bluntly lie their customers (acting as if all is fine - aka - working as intended)!?

There is no official statement whether there is or isn't a design error. Let's wait and see before jumping to conclusions, shall we?

 

[eidairaman1]

And MSI never fixed theirs yet it was faulty (Evoke)

Buying faulty cards and concluding that they're faulty actually concludes nothing. As said above, I believe your findings, but I want to see them confirmed by other sources as well before believing that this is a wide-spread issue. You cannot extrapolate from a handful of (previously known to be faulty) samples.


There is no official statement whether there is or isn't a design error. Let's wait and see before jumping to conclusions, shall we?

Thank you.

 

[Falck]

@TheoneandonlyMrK ok, so you don't take into account the other users on reddit which also are reporting temp issues that's getting resolved by turning the cards vertically? Why would a temp issue get resolved if the card is situated differently? You, as an engineer, should probably have some theories about that? Let's say the VC is working correctly, what am I and many others experiencing? My case is a ssup Meshlicious, had a Radeon 6800XT MBA with no temp issues prior in that case.

Sorry for beeing rude, got overly sarcastic there.

But you didn't answer my question: How many faulty cards are needed before you accept any conclusion? Does Igor labs findings add to der8auer's findings?

 

[AusWolf]

@TheoneandonlyMrK ok, so you don't take into account the other users on reddit which also are reporting temp issues that's getting resolved by turning the cards vertically? Why would a temp issue get resolved if the card is situated differently? You, as an engineer, should probably have some theories about that? Let's say the VC is working correctly, what am I and many others experiencing? My case is a ssup Meshlicious, had a Radeon 6800XT MBA with no temp issues prior in that case.

Sorry for beeing rude, got overly sarcastic there.

But you didn't answer my question: How many faulty cards are needed before you accept any conclusion? Does Igor labs findings add to der8auer's findings?

I still recommend this movie clip:

Once other (major) news agencies pick up the story and report on a mass fault, AMD will have no choice but to acknowledge it, and make corrections. Until then, all evidence is anecdotal, and users have no other choice but to RMA their cards.

 

[damric]

Raja's fault

 

[TumbleGeorge]

@TheoneandonlyMrK ok, so you don't take into account the other users on reddit which also are reporting temp issues that's getting resolved by turning the cards vertically? Why would a temp issue get resolved if the card is situated differently? You, as an engineer, should probably have some theories about that? Let's say the VC is working correctly, what am I and many others experiencing? My case is a ssup Meshlicious, had a Radeon 6800XT MBA with no temp issues prior in that case.

Sorry for beeing rude, got overly sarcastic there.

But you didn't answer my question: How many faulty cards are needed before you accept any conclusion? Does Igor labs findings add to der8auer's findings?

It would be curious if the design was is so bad that the effect of gravity was an obstacle to sufficient performance of the steam chamber just because of the spatial orientation of the installation.

 

[ARF]

The vapour chamber looks like at a different level - it doesn't contact the chiplets.
Can you see how the memory pads are higher than the paste in the middle?

AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Circuit Board Analysis | TechPowerUp
AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Pictures & Teardown | TechPowerUp

 

[AusWolf]

The vapour chamber looks like at a different level - it doesn't contact the chiplets.
Can you see how the memory pads are higher than the paste in the middle?

View attachment 277046
View attachment 277047
AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Circuit Board Analysis | TechPowerUp
AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Pictures & Teardown | TechPowerUp

Those are photos, what you see in them is subjective. You need schematics with accurate measurements if you want to prove a point.

 

[ARF]

Those are photos, what you see in them is subjective. You need schematics with accurate measurements if you want to prove a point.

They had to solder it with an IHS and undervolt it - win-win

 

[TheoneandonlyMrK]

@TheoneandonlyMrK ok, so you don't take into account the other users on reddit which also are reporting temp issues that's getting resolved by turning the cards vertically? Why would a temp issue get resolved if the card is situated differently? You, as an engineer, should probably have some theories about that? Let's say the VC is working correctly, what am I and many others experiencing? My case is a ssup Meshlicious, had a Radeon 6800XT MBA with no temp issues prior in that case.

Sorry for beeing rude, got overly sarcastic there.

But you didn't answer my question: How many faulty cards are needed before you accept any conclusion? Does Igor labs findings add to der8auer's findings?

Of course it would.

We like Igor no?!.

How many, that's not something I can answer since I don't know how many were made.

And I am not claiming to have statistical analyses skills so I can't just make up an arbitrary figure out my ass.

And I don't do Reddit, as a news , information resource or statistics analysis tool it's pure asss, it's that simple.

 

[LFaWolf]

AMD’s statement - https://www.tomshardware.com/news/amd-responds-to-rx-7900-xtx-hotspot-fiasco

 

[AusWolf]

AMD’s statement - https://www.tomshardware.com/news/amd-responds-to-rx-7900-xtx-hotspot-fiasco

Now, this is useful. Bickering about statistics that we know nothing about is not.

 

[blued]

Now, this is useful. Bickering about statistics that we know nothing about is not.

Was just about to post this as its 3 days old that AMD acknowledged the issue and offered affected users to contact their support.
So those reddit posts about AMD refusing RMAs were already out of date and no point for anyone to have posted them here.

 

[delshay]

The vapour chamber looks like at a different level - it doesn't contact the chiplets.
Can you see how the memory pads are higher than the paste in the middle?

View attachment 277046
View attachment 277047
AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Circuit Board Analysis | TechPowerUp
AMD Radeon RX 7900 XTX Review - Disrupting the RTX 4080 - Pictures & Teardown | TechPowerUp

Not sure if it's important but it also does not cover 100% of the main die either. ..ie the shape of the cold plate is different to that of the main center die.

EDIT: All OK seen teardown. ...I was wrong

 

[AusWolf]

Not sure if it's important but it also does not cover 100% of the main die either. ..ie the shape of the cold plate is different to that of the main die.

How exactly do you see that? Did you measure it?

 

[LFaWolf]

Was just about to post this as its 3 days old that AMD acknowledged the issue and offered affected users to contact their support.
So those reddit posts about AMD refusing RMAs were already out of date and no point for anyone to have posted them here.

Absolutely not true. I wouldn’t have known about it since I don’t use use Reddit/AMD and I don’t frequent Tom’s hardware. It is very helpful for me.

 

[delshay]

How exactly do you see that? Did you measure it?

Don't need to measure it, it's clearly visible by the thermal paste. ie the thermal paste is almost a perfect square where the main die is more elongated

EDIT: All OK seen teardown. ...I was wrong

 

[AusWolf]

Don't need to measure it, it's clearly visible by the thermal paste.

It isn't clearly visible to me. Without contact pads (the kind they use at Gamer's Nexus) and/or precise measurements, the shape of the thermal paste smear means nothing.

 

[TumbleGeorge]

Don't need to measure it, it's clearly visible by the thermal paste. ie the thermal paste is almost a perfect square where the main die is more elongated

Please zoom image and draw on it that what you see.