Gigabyte Releases Statement Regarding the SP-CAP and MLCC Capacitor on GeForce RTX 3080

Published by

Click here to post a comment for Gigabyte Releases Statement Regarding the SP-CAP and MLCC Capacitor on GeForce RTX 3080 on our message forum
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
JonasBeckman:

EDIT: Hmm that or NVIDIA or the manufacturer for cards that do have confirmed issues open up for RMA's for refreshed models with hardware fixes but that's costlier than a bios update.
Probably similar to what logitech did for the g903's double clicking issue, smooth the input a little and catch the small moments where a boost could occur.
data/avatar/default/avatar38.webp
I saw Buildzoid's video on the Gigabyte's response and I saw pictures of their card and he said if Gigabyte cut corners they did a very poor job of it because there are capacitors all over the place on the Gigabyte card that are not on the other cards even the FE cards. Also Buildzoid said pretty much the same thing as Der8auer that the caps only made a small difference in terms of overclocking. Also in my mind and Buildzoid said this as well there are too many variables out there to figure out the actual problem. In some users' case they could be using inefficient PSUs which could cause problems to the GPU AIB partners shipping out their cards that boost way too high out of the box and therefore the card is unstable because they didn't do enough testing for stability. Also this could be driver related as well because since this is a new generation of cards the drivers aren't mature yet.
https://forums.guru3d.com/data/avatars/m/209/209146.jpg
Maybe a lot of caps don't quite measure up to some higher quality or more tolerant caps when it comes to power rippling and fluctuations that might be what's happening once some of these GPU's pass the 2 Ghz clock speed threshold. I don't have the tech know-how for that though but I've seen similar for some of these power phases on GPU's and using fewer quality models over more but lower quality variants or the popularity of these doublers. I think you could put quite a bit of these bits on the GPU's PCB but if they don't measure up to handling this current or voltage rippling and the quick state switches and power load especially for how some of the GPU's binning might require a bit more for these highest peak speeds that could still be a problem. Buildzoid seems to be quite talented though so if he has the card and knows what's on these and which configurations and their tolerance and overall quality level then it's probably fine. 🙂 Might take a bit more testing and for NVIDIA to start replicating the problem themselves to identify the core issue of this problem too if there's more to it. EDIT: Suppose that's been mentioned a bit in these discussion topics by this point already too, will be interesting to read more on how this develops and an eventual full fix for the problem and what that will end up being.
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
buildzoid really only knows what the specs say, he doesn't have any formal training in how to place the parts, the matters of pcb drilling practices and hes also wrong and doesn't immediately correct himself till months later. hes good for telling you when components are crap, and good at hating on numbers he doesn't quite comprehend.
https://forums.guru3d.com/data/avatars/m/209/209146.jpg
Interesting, good to know too referenced frequently after all or cited and with a certain popularity though going by what Youtube is recommending that algorithm is...something. Igor then perhaps from Igor's Lab although the articles are primarily in German but the translation for the English material is adequate enough. Bit of a specialized field of engineering and yeah what do I know, I find it very interesting to read on and get some understanding of but it's often pretty basic much as I find the topic fascinating and try to learn what I can. Considering that though there can't be too many people with that knowledge doing things like Youtube channels or donation funded projects compared to a more stable job and the income with those types of skills and specializations. EDIT: Well and also having actual insider sources or connections or a very credible reputation over years of work in the industry itself. Can't be easy though with how secretive and protective this business can be, some of the bigger names and tech experts might be able to jump around doing contractual stuff for the major hardware businesses but those would be people at the top of their respective fields or how to call it that would always be sought after. That's a bit different from this though. 🙂 Wonder if NVIDIA will have some further statement or explanation for this later on. Or what the plan looks like for then resolving it more permanently. Well unless forced to replace hardware I assume a bios update is the cost balanced alternative. EDIT: Other than Inno3D it seems who took the incredible consumer oriented angle of "Works for me." Which is kinda amazing but not exactly a good stance compared to the others. (Some of these tech statements like this one are a bit mid-ground and the outright apologies and gifts a bit above that.)