just found this story on guru3d.com
https://www.guru3d.com/news-story/g...ely-due-to-poscap-and-mlcc-configuration.html
https://www.guru3d.com/news-story/g...ely-due-to-poscap-and-mlcc-configuration.html
-
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.
Rtx 3080 lower quality capacitor Issue
- Thread starter Shahzada
- Start date
-
- Tags
- capasitors rtx 3080
Associate
This from EVGA: https://www.tomshardware.com/uk/news/evga-nvidia-rtx-3080-capacitor-caused-crashes
"EVGA has seemingly confirmed that issues with a certain type of capacitor are the source of an increasing number of reports of crashes associated with factory-overclocked Nvidia GeForce RTX 3080 graphics cards."
"EVGA has seemingly confirmed that issues with a certain type of capacitor are the source of an increasing number of reports of crashes associated with factory-overclocked Nvidia GeForce RTX 3080 graphics cards."
I have a 3080 Gaming Trio with 1 MLCC and it crashes over 2 Ghz.
So that’s every card with every configuration of capacitors reported as crashing.
Soldato
I've just been pushing this card all the way up to +150mhz offset and it still will not crash but the boost clock will not increase over +75mhz offset.
This card hits its TDP limit and thats it with no crashing or ill affects no matter the boost clock spikes.
This card will spike MUCH less with the Prefer max performance setting over the Adaptive and Optimal ones which might tip some less stable cards over the edge?
Exactly. I'm going to be coming up from a 1080 Ti and sticking with my 1440p 144Hz monitor. I've ordered a TUF card and will give it a stable OC and call it a day regardless of what it does. There's absolutely no-point chasing numbers - folks just need to concentrate on the experience and judging from the reviews we should see a healthy bump in performance even if you're stepping up from a 2080 Ti. I don't think I'll miss a few frames here and there but reading some of the threads on here its laughable, especially considering the generational leap Nvidia has delivered this launch.
Let's be grateful - it really isn't as bad as all the C3PO's on here are alluring to.
Associate
My electronics is a bit flakey (my qualification is microelectronics is 34 years old) but the caps on the back need to be a set capacitance value for the bypass. This can be achieved using a combination of sp/caps and mlcc. The cards with 6 sp/caps differ. The gigabyte card has higher rated capacitance than that of the zotac so in theory has a better bypass delivery, but not as good as those with 5 sp/caps and 10 mlcc or 4sp/caps and 20 mlcc. The small mlcc as explain by Buildzoid has a faster frequency but a lower capacitance storage value i.e. they run out of juice faster. It would be safe to assume that a mix of the two would be the prefered for sustained power, but both should function if the card power parameters in the bios are correct.The fact the Nvidia passed all these designs means that they must meet some standard.
The crashing is more than likely a combination of this power bypass issue and the AIBs/Nvidia not having enough time to qualify their overcloacks fully in the wild to work with their choosen power bypass method. Other factors also come into play here too such as quality the main board, powertraces etc.
I would wait before making too many judgements. If too serious then they can always recall.
The crashing is more than likely a combination of this power bypass issue and the AIBs/Nvidia not having enough time to qualify their overcloacks fully in the wild to work with their choosen power bypass method. Other factors also come into play here too such as quality the main board, powertraces etc.
I would wait before making too many judgements. If too serious then they can always recall.
whilst evga post is very badly written and throws most aib under the bus.
they only state that they had issues with there own cards in testing. Every aib will use differnt power delivery. They all use differnt capacitors to 2 cards from companies are the same. They could have had issues with that setup doesn’t mean all do.
evga person who written that will be getting a good telling off by evga bosses and nvidia
look at this way if this was the only reason for crashing (hardware unboxed says even his ASU’s crashes the same) why aren’t ASUs shouting it from the roof tops that’s there’s don’t have so-caps ?That's a good find, bear in mind as well that some unrelated problems might get thrown in with this, so that "might" explain the post at HardwareUnboxed that SeeNoWeevil posted. I guess we will just have to wait and see. However all the early refrence designs I saw had 1 MLCC with 5 POSCAP as "standard" and the "beefier" ones going with 2 and 4 respectively. The Asus with all 6 MLCC, that the folk posting at HardwareUnboxed are reporting problems on might just be having some other problem.
Soldato
Nothing to stop them using cheap components after approval I bet Nvidia do not approve every single component added to the card.
Just been reading this on Facebook quite a few tech sites publishing stuff on there about it. Sounds like AIB's not following Nvidia's reference.
Could be ASUS is definitely the cards to go for as they dont seem to have scrimped on the design.
Yet there are still reports of them crashing.
Associate
- Joined
- 24 May 2015
- Posts
- 500
I'm fairly sure no MLCC is still within Nvidia's spec.
Soldato
Associate
- Joined
- 24 May 2015
- Posts
- 500
Gigabyte RTX 3080 Gaming OC 10GB
My initial overclocking results appear promising, on this card I've no option to increase core voltage, power limit is maxed out at 100% by default. Though I've set a pretty aggressive fan profile and increased my temp limit to the max of 91c. The Highest temp I've seen so far is 59c.
I have not yet clocked the memory as with this being error correcting I want to spend more time with it.
So far I have only used 3D Mark Port Royal to test but will be adding more benchmarks and games to really check for stability.
Initially I kept temp limit at stock and found I could achieve +250 to the core however this yielded a far lower final score by a few hundred points compared to +150 which I found odd. The card also wouldn't boost beyond 1950Mhz. At this point I did start to wonder whether Gigabyte had already gimped the card before shipping.
Changing the temp limit to 91c meant the card was now boosting above 2000Mhz, as high as 2095Mhz but resulted in CTD's. I eventually brought it back to +125 on the core and had my most impressive results yet. The card boosted as high as 2070Mhz and gave me a Port Royal score of 11949.
I'll keep you guys posted with my findings throughout the day.
My initial overclocking results appear promising, on this card I've no option to increase core voltage, power limit is maxed out at 100% by default. Though I've set a pretty aggressive fan profile and increased my temp limit to the max of 91c. The Highest temp I've seen so far is 59c.
I have not yet clocked the memory as with this being error correcting I want to spend more time with it.
So far I have only used 3D Mark Port Royal to test but will be adding more benchmarks and games to really check for stability.
Initially I kept temp limit at stock and found I could achieve +250 to the core however this yielded a far lower final score by a few hundred points compared to +150 which I found odd. The card also wouldn't boost beyond 1950Mhz. At this point I did start to wonder whether Gigabyte had already gimped the card before shipping.
Changing the temp limit to 91c meant the card was now boosting above 2000Mhz, as high as 2095Mhz but resulted in CTD's. I eventually brought it back to +125 on the core and had my most impressive results yet. The card boosted as high as 2070Mhz and gave me a Port Royal score of 11949.
I'll keep you guys posted with my findings throughout the day.
No clue, but gonna assume theres no way its six of the cheap ones