michalvan89

MRizk

michalvan89
i dont get why there is so much complains about the SP-CAPs used instead of MLCCs. GPUs are working great with stated frequencies, anything above is just overclocking (not counting gpu boost algorithm which can be capped and is just nice bonus) where noone can guarantee stability. Ppl should just learn how to enjoy things, instead of complaining about nonsense. Fix is easy, just cap the maximum frequency bellow 2GHz.

You see the boost clock as a bonus, me and others see it as a feature. Reviews and benchmarks posted online "for marketing" where done with it enabled. Doing a VBIOS update to lower it means false advertising in my books. (I know EVGA said nothing about a VBIOS to lower the boost, in fact, Jacob denied them having any plans to do this which is why I am still sticking with EVGA)
 
Also I personally buy EVGA to overclock. I buy EVGA because they offer decent coolers. If I didn't care about overclocking I'd just go FE. I have had an EVGA 750 Ti, EVGA 1080 Ti, and EVGA 2080 Super and I have overclocked each and everyone of them. None of them had issues maintaining the factory OC. I don't expect the 3090 FTW3 to be any different. 
 
I personally don't care what capacitors a card has. I never knew a thing about capacitors and I never expected to even hear about them. I don't care what EVGA uses as long as the card works as advertised, and is not worse than others in the same price range. 

well when you buy car, and go to do some chiptuning or any modification and it will go wrong i doubt that you will complain about it not being good to manufacturer of car :) ... all i can see cards working fine, differences are like 1-2 fps when capped so why even bother. EVGA has good approach and fixed this in FTW3 series. So i see no problem here. Boost clocks are about possible headroom anyway, if there is enough power and temp to be spent. If i want to OC then i have to count on fact that there is some max spot where i cannot go over.
Not here to argue, just surprised that ppl have enough balls to even highlight such thing :)

Agree with the analogy...but any car enthusiast is going to pick a car they know produces the best results when they tinker with them (MKIV).  In this case, those choosing between an FTW3, Aorus, or Strix, are going to look for the one that produces the best results when they thinker with it.
 
Also...for those that want to slap a block on these chips, it's things like this that matter.  Is EVGA going to be limited on clocks compared to ASUS as a result of these decoupling capacitors?  When it comes to targeting the premium card of a series, you are paying for that custom PCB that is overengineered...not skimped on.  I am trusting EVGA here though and I'm sure their design is great...but I'm still waiting to see what ASUS, Gigabyte, (and AMD) put out before making my decision.
 
I think this is why people are asking these questions.  I think it's fair to ask this and challenge them.  You don't have to be an expert to ask questions...no matter how relevant. 

Vlax_Aus
Hello. I'm mechanical engineer with some electrical experience and I would like to share my opinion on this. I think that the problem is RF frequency noise at higher voltages, not the actual capacity of sp-caps (mtcc) or mlcc ones. Higher rated caps 220-330-470uf have much higher frequency noise, and due to this design (closest possible to gpu core for fast voltage discharge/flow) they have smaller or don't have at all RF filter. Because most crashes happening when clock speed is hitting 2ghz, 300+ watts. And mlcc caps are much more unreliable but faster responding than sp caps, that's why you need combination of both. I think this Evga's 4x 220 (lower noise) + 2(ftw) or 1(xc) is very clever and it will actually work fine. But due to capacitor sizes , I don't think that they will top the performance charts (but they will be reliable and higher tier).

So is there "much" difference between using 4x220 or like the fe 2x220 + 2x470? And in what way would that influence anything?
If you look up pcb back pictures on techpowerup.com it seems a lot are using 470 or 330 and not much using 220.
Maybe there can be another reason you can think if that evga uses 220's maybe they have bigger capacity somewhere else or there power delivery won't need bigger ones or something?
 
It's probably a non issue but i and some others are a bit  querious why most others seems to use bigger ones and sin I have no idea how it really works we just can do one thing and that's ask around :)
 
 

DarthSawyers
Best Buy shows the STRIX having 6 pcaps but I bet it’s not accurate.

ShaconBacon
Yeah on the Asus website the STRIX has all 6 as MLCC's...who knows

Here's a video at the 4 min mark he flips the Strix around to the back.


https://www.youtube.com/watch?v=8IDMS54jEFM

Does the 6 MLCC design mean more quality and block crashing?

Memsar
Does the 6 MLCC design mean more quality and block crashing?

I won't answer that. Some say all is better some say 2 in the middle is better. Here is a video Buildzoid did for Gamers Nexus of the breakdown FE and what he thinks. A very smart dude.
 
https://www.youtube.com/watch?v=THMukcOzB8g

Why would they need to report anything?  Its fixed on the production units.  Im sure they ran into other issues that where fixed prior to production do you expect EVGA to tells us each of those??  Love EVGA first class all the way. 

So EVGA has cheaped out on FTW3's somewhere compared to it's rival Strix, since Strix OC guarantees 1935MHz boost compared to FTW3 Ultras 1800MHz. Which kinda makes sense since later on, EVGA is going to offer those even more expensive Hydrocoppers and Kingpins.
 
The question is why should I buy more expensive FTW3 Ultra instead of the Strix OC in this specific price segment.
 
Strix PCB looks clean, would love to see FTW3 teardown and comparison.

YURIIII
So EVGA has cheaped out on FTW3's somewhere compared to it's rival Strix, since Strix OC guarantees 1935MHz boost compared to FTW3 Ultras 1800MHz. Which kinda makes sense since later on, EVGA is going to offer those even more expensive Hydrocoppers and Kingpins.
 
The question is why should I buy more expensive FTW3 Ultra instead of the Strix OC in this specific price segment.
 
Strix PCB looks clean, would love to see FTW3 teardown and comparison.

Jesus some of you are dense. Soooo many people with knowledge on the subject have said EVGA's setup is perfectly fine, if not better in some instances than the STRIX card.
 
But by all means, please don't buy one. Maybe I'll actually have a chance of checking out with one next time they're ins tock. 

R3L3NTL3SS

YURIIII
So EVGA has cheaped out on FTW3's somewhere compared to it's rival Strix, since Strix OC guarantees 1935MHz boost compared to FTW3 Ultras 1800MHz. Which kinda makes sense since later on, EVGA is going to offer those even more expensive Hydrocoppers and Kingpins.
 
The question is why should I buy more expensive FTW3 Ultra instead of the Strix OC in this specific price segment.
 
Strix PCB looks clean, would love to see FTW3 teardown and comparison.

Jesus some of you are dense. Soooo many people with knowledge on the subject have said EVGA's setup is perfectly fine, if not better in some instances than the STRIX card.
 
But by all means, please don't buy one. Maybe I'll actually have a chance of checking out with one next time they're ins tock. 

I'm not saying EVGA's cards are not fine. But if product X can guarantee you higher clockspeeds than product Y, there's probably a reasoning behind that. Those numbers printed on the box are probably not pulled out of ass.
 
Hypothetically. If I end up buying a card that promises me 1800MHz and it barely makes it, even though most people got +1900MHz with the same card, I can't do anything about it because it did what was promised. You got the idea?
 
That's why I also said that it would be nice to see both cards torn down, side by side, to find those differences.

Axejess

Vlax_Aus
Hello. I'm mechanical engineer with some electrical experience and I would like to share my opinion on this. I think that the problem is RF frequency noise at higher voltages, not the actual capacity of sp-caps (mtcc) or mlcc ones. Higher rated caps 220-330-470uf have much higher frequency noise, and due to this design (closest possible to gpu core for fast voltage discharge/flow) they have smaller or don't have at all RF filter. Because most crashes happening when clock speed is hitting 2ghz, 300+ watts. And mlcc caps are much more unreliable but faster responding than sp caps, that's why you need combination of both. I think this Evga's 4x 220 (lower noise) + 2(ftw) or 1(xc) is very clever and it will actually work fine. But due to capacitor sizes , I don't think that they will top the performance charts (but they will be reliable and higher tier).

So is there "much" difference between using 4x220 or like the fe 2x220 + 2x470? And in what way would that influence anything?
If you look up pcb back pictures on techpowerup.com it seems a lot are using 470 or 330 and not much using 220.
Maybe there can be another reason you can think if that evga uses 220's maybe they have bigger capacity somewhere else or there power delivery won't need bigger ones or something?
 
It's probably a non issue but i and some others are a bit  querious why most others seems to use bigger ones and sin I have no idea how it really works we just can do one thing and that's ask around :)
 
 

Difference is as they state 220/330/470 atc .., voltage charge. Bit you need to look at the whole card, not only this 6 caps. This 6 caps are only first responders to the immediate call for power/ followed by voltage controller to take you to the right power stage/ then vrm's kick in / then voltage is called from psu( delivery). Evga was always using best vrm capacitors, switches, and everything was tested properly. It's just my opinion that because they used 220mtcc, that tells me what they estimated to be optimum for the given chip with whole that new 8nm process and other changes. I'm sure that card will perform excellent. Bit because some other partner card's didn't implement power delivery properly, I think Nvidia will tune down nv boost feature, otherwise it would be a massive recoil on some brands.