GeForce RTX 3080 CTD Issues update #5 - we think it's fixed (and a bit of analysis)

#5835336

Loobyluggs

2020-09-26 13:25

Oh dear...

#5835341

Undying

2020-09-26 13:37

So who is to blame? Aibs, nvidia, yourself? 😀

#5835344

cryohellinc

2020-09-26 13:40

Undying:

yourself?

Yourself - for jumping the early adopter bandwagon. 😉

#5835346

Undying

2020-09-26 13:41

cryohellinc:

Yourself - for jumping the early adopter bandwagon. 😉

Another beta testers, yeah. Rtx users just cant catch a break. 😀

#5835353

cryohellinc

2020-09-26 13:51

Undying:

Another beta testers, yeah. Rtx users just cant catch a break. 😀

But it's the same nearly always - 1st gen of what-ever is always full of issues. I can recall very few successful launches. Ampere - do I want to buy it at some points? Yes. Will I buy it now? No, because AMD's response isn't out yet, and in addition to that - early adopter = high risk. Let it sit for several months and mature. On top of that prices will settle down. The same goes with PS5 - I plan to get one for myself, but I won't go for the first iteration, and instead will get a second revision.

#5835355

DannyD

2020-09-26 13:54

Here's hoping big navi early adopters are more lucky! At least nvidia's new Ampere is fast a hell, we've still to see Navi's game performance, lets hope they leak more than AOTS numbers.

#5835359

darrensimmons

2020-09-26 14:04

Asus cards seem to not be not affected by this problem due to not cutting corners. Looks like I'll be waiting for a Strix oc.

#5835362

Undying

2020-09-26 14:11

darrensimmons:

Asus cards seem to not be not affected by this problem due to not cutting corners. Looks like I'll be waiting for a Strix oc.

How do you know? Bryan from tech yes said asus tuf also crashes if overclocked but stock is fine.

#5835371

alanm

2020-09-26 14:34

Well at least AMD should not have this problem due to no AIB reviews for several weeks after launch. So should be a solid, well tested AMD reference card.

#5835372

Darren Hodgson

2020-09-26 14:40

I watched the JayZTwoCents YouTube video on this issue and it seems the Founders Edition cards do not suffer from this issue as they use above spec parts. It is some of the AIB partner cards that have this issue as they cheaped out on a particular part that reduces electric interference, something like that. However, the cards are still within specifications, it's just that when people push the cards further, beyond what they were tested for, then the cutbacks result in crashes because they are not designed to handle such high overclocks. I guess you could argue that if people are paying for £100+ extra for AIB card then it is because they want that extra performance so if they can't achieve it then the card is flawed and not fit for purpose. I would say that is grounds to return the card personally. For most people, it will never be an issue and is not grounds for a product recall in my opinion. Hopefully, the AIB partners will review this and redesign the cards to fix this properly.

#5835375

HybOj

2020-09-26 14:51

Darren Hodgson:

I watched the JayZTwoCents YouTube video on this issue and it seems the Founders Edition cards do not suffer from this issue as they use above spec parts. It is some of the AIB partner cards that have this issue as they cheaped out on a particular part that reduces electric interference, something like that. However, the cards are still within specifications, it's just that when people push the cards further, beyond what they were tested for, then the cutbacks result in crashes because they are not designed to handle such high overclocks. I guess you could argue that if people are paying for £100+ extra for AIB card then it is because they want that extra performance so if they can't achieve it then the card is flawed and not fit for purpose. I would say that is grounds to return the card personally. For most people, it will never be an issue and is not grounds for a product recall in my opinion. Hopefully, the AIB partners will review this and redesign the cards to fix this properly.

Not sure we have read the same article. There was no mention of overclocking anywhere. So... care to share the source of your info where it states its caused by overclocking? But you dont have to do that anyway. Because if it was the issue only for a overclockers, than there would not be any situation like this. The cards are malfunctioning stock, because they cant handle high clocks, and these clocks seem to appear at the low loads. Id say this is quite a serious issue, because its a HW related. Also, imagine being that guy, who buys 3080, it crashes, he buys new PSU just to see the card still crashing and than reading the article like this? Damn... my condolences. But somehow, these ppl deserve the burn. Not even waiting few months for the AMD. Sounds like a fanbois to me 😀

#5835381

DannyD

2020-09-26 15:14

It's totally nvidia's fault, they sent out the minimum spec sheet, approved their final designs and limited the AIBs in the testing process.

#5835388

Reqruiz

2020-09-26 15:27

Imagine buying GPU for 700+ bucks and experiencing such issues. Pathetic design. It should be just flawless and perfect for this kind of money.

#5835402

Kaarme

2020-09-26 15:59

DannyD:

It's totally nvidia's fault, they sent out the minimum spec sheet, approved their final designs and limited the AIBs in the testing process.

Will Nvidia accept any kind of responsibility? I think not. If some partner complains to Nvidia, Nvidia will probably tell them that they are welcome to stop producing Nvidia based video cards if working with Nvidia doesn't suit them. End of negotiations.

#5835404

TheDeeGee

2020-09-26 16:03

I may be wrong but it should be easy to verify what capacitor your GPU has? All Zotac capacitors have the number 330 on them. Then again the Founders Edition has a mix of 2x 470 and 2x 220. So i'm not sure those numbers are related to the type of capacitor used. The MSI 3080 Gaming X Trio uses 5x 470. If those numbers are related to the type of capacitor it would mean the FE uses 2 even weaker capacitors.

#5835408

CPC_RedDawn

2020-09-26 16:14

darrensimmons:

Asus cards seem to not be not affected by this problem due to not cutting corners. Looks like I'll be waiting for a Strix oc.

Judging from pictures on overclockers BOTH the STRIX OC and the TUF OC both use 6 cheap POSCAP array instead of 4pos+2mlcc like FE cards.

Only card I have seen so far that doesn't cheap out is the MSI Gaming X Trio which seems to be using the same 4pos+2mlcc setup as the FE models.

tarjeta-grafica-msi-geforce-rtx-3080-gaming-x-trio-10g-10-gb-gddr6x-galeria6.jpg

#5835419

lukas_1987_dion

2020-09-26 16:49

After the RTX 2080 Ti (not only) dying scandal, I thought they would be more careful this time. Ofc it's NVidia's fault, AIB partners were always releasing thier own versions most of the time with average/high OC and it was never a problem before. Always blaming PSU as source of such problems is so popular nowadays lol Btw, many folks with FE are also reporting the same problems.. I'm so glad I did not jump on the crazy wagon this time xD

#5835427

TheDeeGee

2020-09-26 17:10

Undying:

So who is to blame? Aibs, nvidia, yourself? 😀

Looks like Nvidia, due to a rushed launch and not giving AIBs enough time to test.

CPC_RedDawn:

Judging from pictures on overclockers BOTH the STRIX OC and the TUF OC both use 6 cheap POSCAP array instead of 4pos+2mlcc like FE cards.

Only card I have seen so far that doesn't cheap out is the MSI Gaming X Trio which seems to be using the same 4pos+2mlcc setup as the FE models.

MSI uses the higher quality SP-CAPs, not POSCAPs.

#5835429

TheDeeGee

2020-09-26 17:13

lukas_1987_dion:

After the RTX 2080 Ti (not only) dying scandal, I thought they would be more careful this time. Ofc it's NVidia's fault, AIB partners were always releasing thier own versions most of the time with average/high OC and it was never a problem before. Always blaming PSU as source of such problems is so popular nowadays lol Btw, many folks with FE are also reporting the same problems.. I'm so glad I did not jump on the crazy wagon this time xD

Probably because the FE uses 3 different caps. They have atleast 2 SP-CAPs and 2 MLCCs but they use a third type with the number 220 on it.

#5835430

Undying

2020-09-26 17:14

Amaze:

Apparently cards with the fancy caps are also crashing so who knows how this will play out.

Bryan said his Tuf is crashing if overclocked and Jay said his threadripper system become unstable and crashing with his FE card on 1200w psu. Before that 2080ti was 100% stable. We'll all see what happens next when more people get their hands on the card and do some testing.