News EVGA Says RTX 3080 Cap Issues Caused Crashes, Confirms Stability Issues

[Ford_Prefect]

Please stop calling them "POSCAPS" - that is a name used by Panasonic to describe their own product line. Panasonic POSCAPS are not used in any GPUs.

 

[Fractalite]

It is amazing how fast ignorance travels on the Internet.

Those are not PoSCaps:

View: https://www.youtube.com/watch?v=GPFKS8jNNh0&t=88s


Buildzoid is amazing. His content is not for everyone, but for those with the willingness, he gives excellent insight into motherboard purchases. He also does the occasional PCB breakdown.

Note also; towards the end of the video, Buildzoid, who was hesitating a touch during the recording, does come out and ask for some oscilloscope readings. As in, sure, they are not PoSCaps, but to know for sure what the problem is, more data is needed.

-----------------------------------

Also, weird that EVGA called them PoSCaps when they are not? Does not exactly fill me with confidence about their cards...

There is some tepidly useful advice on this thread: never rush to purchase items, but review sites/youtubers also need to slooooooowwww down. This 30 series launch business has gotten a touch out of hand. First Nvidia comes out and lies outright about their cards and now some of these review sites/youtubers seem like they are rushing about or flailing.

 

[PillowOfWinds]

Nvidia's own RTX 3080 FE models, which only reach 1.71 GHz, aren't impacted by the crashes.

This is not true, 1.71 Ghz is the minimum guaranteed boost clock for FE models. In reality most of the cards will reach a higher frequency, depending on wattage, temperature and silicon lottery.

Many manufacters that decided to cut the number of MLCC capacitors will have to push a firmware update to limit the max frequency and avoid instability.

source:

View: https://youtu.be/x6bUUEEe-X8

 

[Phaaze88]

Not Asus, on their TUF 3080 they went with the more expesive Caps on all 6 modules.....thats why it costs more.. Love or hate Asus this they got very right.

It would appear Asus isn't safe either:

https://twitter.com/x/status/1309659834468298753

View: https://twitter.com/HardwareUnboxed/status/1309659834468298753

 

[InvalidError]

Those are not PoSCaps:

They may not be Panasonic POSCAPs but they have similar physical dimensions, similar appearance and serve the same function regardless of the specific brand being used.

They are "not POSCAPs" just as much as generic facial tissues aren't kleenexes, most people call them that out of convenience since they are functionally indistinguishable.

 

[nofanneeded]

They may not be Panasonic POSCAPs but they have similar physical dimensions, similar appearance and serve the same function regardless of the specific brand being used.

They are "not POSCAPs" just as much as generic facial tissues aren't kleenexes, most people call them that out of convenience since they are functionally indistinguishable.

does not mean similar Quality though...

 

[nofanneeded]

I wouldn't be so sure about that mate:

https://twitter.com/x/status/1309659834468298753

View: https://twitter.com/HardwareUnboxed/status/1309659834468298753

I dont know why the fuss , just keep the clocks under 2Ghz and admit thats the clock limit for this GPU and move on guys ...

When our CPU crashes after 5.3 Ghz OC we dont make such a fuss about it .

 

[InvalidError]

does not mean similar Quality though...

Possibly. Though if you are going to use polymer caps to cover the filtering range that MLCCs would cover, you don't really have any room to skimp on quality.

 

[JamesSneed]

You know none of these are using "POSCAPs ". That capacitor is a specific brand that Panasonic makes.

Edit: Here you go. Buildzoid has a great description.

View: https://www.youtube.com/watch?v=GPFKS8jNNh0

 

[JackGrat2]

Once you get above 100MHz - no tantalum cap is going to be effective - the ESR is too high - the ceramic caps in the center are the best at higher frequencies - I just hope they are a spread of values - .1uF, 0.01uF, 0.001uF, 100pF, 22pF - or something similar - see https://ds.murata.co.jp/simsurfing/mlcc.html?lcid=en-us
and browse for values and display |Z|
The PCB design is just as critical - and the layout allowed both tantalum footprint as well as ceramic devices to be used - that should have been a clue that they hadn't fully characterized the layout.

 

[Bamda]

So glad I am not swayed by the hype of new hardware and games and burned by crap like this. Thanks to those of you that beta tested these devices for me.

 

[Hadien789]

Not Asus, on their TUF 3080 they went with the more expensive Caps on all 6 modules.....thats why it costs more.. Love or hate Asus this they got very right.

From what I can tell from the article and jayztwocents explanation on the capacitors, the larger caps are actually the issue (unless you meant the smaller cap clusters are more expensive than the large ones). The smaller MLCC clusters are better at regulating voltage and that's the reason there's more issues with those using the 6 large caps vs the reference using 5 and 1 MLCC cluster (10 mini-caps in each group) and the FE card using a 4/2. From what people are surmising, expect possible firmware updates to keep the cards from going into unstable voltages...

Also, the TUF design uses 6 of the POSCAPS vs the ROG Strix and TUF both are using 6 MLCC clusters. EVGA went 4/2 for FTW and 5/1 for XC3 from what I can tell (which mirrors NVidia's FE design and reference design respectively)

Edited to reflect new info... basically strike everything out lol...

 

[InvalidError]

the ceramic caps in the center are the best at higher frequencies - I just hope they are a spread of values

If you mean different values for the same package size, that is largely pointless since MLCCs's useful range is primarily limited by their ESL which is itself a function of package height and length. A single 1uF MLCC would have too much ESL to be of much use to an ASIC with 250-300A draw from CMOS switching, that's why MLCCs get sprayed all over the place. The MLCC arrays in place of polymer caps are likely all the same value in the neighborhood of 1uF assuming they are 0603.

 

[Schlachtwolf]

Ooohhhh man Hadien789, read the articale on Igors Lab, Jayz2cents does not even have the TUF card yet, he said so himself in his video.... get the facts right please or it gets very confusing.... Asus only used the MLCCs on it's TUF 3080. Here is the articale with the picture to prove it....

The possible reason for crashes and instabilities of the NVIDIA GeForce RTX 3080 and RTX 3090 | Investigative | igor´sLAB

Not only the editors and testers were surprised by sudden instabilities of the new GeForce RTX 3080 and RTX 3090, but also the first customers who were able to get board partner cards from the first…

www.igorslab.de

 

[Hadien789]

Ooohhhh, read the articale on Igors Lab, Jayz2cents does not even have the TUF card yet, he said so in his video.... get the facts right please or it gets very confusing....

The possible reason for crashes and instabilities of the NVIDIA GeForce RTX 3080 and RTX 3090 | Investigative | igor´sLAB

Not only the editors and testers were surprised by sudden instabilities of the new GeForce RTX 3080 and RTX 3090, but also the first customers who were able to get board partner cards from the first…

www.igorslab.de

He doesn't, however you can see the layout on the ASUS webpage for both the TUF and the ROG Strix. They show the backplate and the capacitor arrangement.

Edit: I stand corrected, they have the 6 clusters of MLCC, I could've sworn they had the 6 large caps earlier... Might also explain why I couldn't open the images up later as I was shopping around.

A quick google search of TUF 3080 images confirms that I am, indeed, not crazy. There are some images with the black caps and some of them with the updated MLCCs.

 

[JarredWaltonGPU]

Dreading my 3090 GPU now incase of crashes but not heard anyone complaining about 3090 yet

What card did you get?

I have tested RTX 3080/3090 FE
Asus TUF 3080/3090
MSI 3080 Gaming X Trio
Gigabyte 3090 Eagle

Those last two had some stability issues, though the MSI seemed to get fixed with the public 456.38 drivers. The GB card couldn't finish my benchmarking in Shadow of the Tomb Raider or Forza Horizon 4 at factory stock settings, but a 15MHz downclock did the trick. But it might also become unstable during a longer play session.

Also, the FE cards definitely exceed 1710MHz -- that's just the official boost speed, but the 3080 FE hits about 2000 peak, and averages 1850-1950 depending on the game. The 3090 FE is a bit lower (30-50 MHz lower).

 

[Phaaze88]

Ooohhhh man Hadien789, read the articale on Igors Lab, Jayz2cents does not even have the TUF card yet, he said so himself in his video.... get the facts right please or it gets very confusing.... Asus only used the MLCCs on it's TUF 3080. Here is the articale with the picture to prove it....

The possible reason for crashes and instabilities of the NVIDIA GeForce RTX 3080 and RTX 3090 | Investigative | igor´sLAB

Not only the editors and testers were surprised by sudden instabilities of the new GeForce RTX 3080 and RTX 3090, but also the first customers who were able to get board partner cards from the first…

www.igorslab.de

What do you have to say about the following though? Steve from Hardware Unboxed has a 3080 TUF Gaming, and it's doing it too.
So no, Asus isn't safe - none of the AIBs are.

https://twitter.com/x/status/1309659834468298753

View: https://twitter.com/i/web/status/1309659834468298753

 

[vinay2070]

Will vendors now replace the cards? Or early adopters have to live with it? I am assuming its the later.

 

[nofanneeded]

Will vendors now replace the cards? Or early adopters have to live with it? I am assuming its the later.

if the The cards did not promise 2Ghz boost , they wont be replaced , they will just update the bios to keep the card under 2Ghz.

 

[InvalidError]

if the The cards did not promise 2Ghz boost , they wont be replaced , they will just update the bios to keep the card under 2Ghz.

I would have some concerns about the long-term stability and overall longevity of a card that needed a downgrade to be stable at launch: a design that is already marginally stable at launch assuming AiB don't down-clock their cards more than absolutely necessary to achieve stability over the warranty period has that much less headroom to accommodate normal wear.

 

[animekenji]

I hope none of you bought your cards from the ebay scalpers.

If the only solution they are offering is a firmware update that reduces performance, then return your cards for a refund. You won't be getting the boost performance that you paid extra money for. Any reduction in performance needs to be accompanied by additional value to make up for the price premium that you paid over the price of a reference card either in the form of a cash refund, a free game, or something else.

 

[spongiemaster]

I would have some concerns about the long-term stability and overall longevity of a card that needed a downgrade to be stable at launch: a design that is already marginally stable at launch assuming AiB don't down-clock their cards more than absolutely necessary to achieve stability over the warranty period has that much less headroom to accommodate normal wear.

I would too, but beyond returning the card for a refund and waiting for a solution, what are your options? Did AMD replace any of the 5600 XT's that didn't hit the last second increased memory speeds?

 

[spongiemaster]

I hope none of you bought your cards from the ebay scalpers.

Karma's a B.

 

[InvalidError]

Did AMD replace any of the 5600 XT's that didn't hit the last second increased memory speeds?

That isn't AMD's job, it is the AiB's problem. Some models got official downgrades, some got recalled, the rest are being handled via the RMA route if/when they fail should that happen in-warranty.

 

[cryoburner]

Imagine paying a big premium to a reseller, only to have the card arrive bubble wrapped, tumbling around in a generic brown cardboard box, and then it's not even stable. : P