RTX 3080 Crash to Desktop Problems Likely Connected to AIB-Designed Capacitor Choice

[moproblems99]

Excuse me but what the hell are you talking about?

Yes it consumes more but the perf/watt is the highest of any card, AMD doesn't even get close.

8% isn't close?

 

[Kaotik]

Someone probably mentioned it already, but these issues are affecting Founders Edition cards too

 

[Paganstomp]

I got this funny feeling between my toes, again! PC gaming is dead.

 

[Deleted member 24505]

How many people are glad they kept their 2080ti's which is still a damn good GPU

 

[mouacyk]

When you have EE's talking about this issue, that's not good for NVidia and especially AIB's.

 

[HisDivineOrder]

And this is why you should never buy a product at launch.

 

[tomc100]

 

[metalfiber]

Here's what happened with the 2080ti launch...It's best not to get the first run of anything be it a car, appliance, etc, etc.

NVIDIA Confirms Issues Affecting Early Production Run of GeForce RTX 2080 Ti Graphics Cards

NVIDIA, via a blog post on its forums, has confirmed widespread reports of failures affecting their flagship RTX 2080 Ti graphics card. The issues, which resulted in "crashes, black screens, blue screen of death issues, artifacts and cards that fail to work entirely," started cropping up...

www.techpowerup.com

 

[Tomgang]

So it seems the to go to cards are asus tuf/strix and evga FTW 3. Glad I dit not ordered a card yet and spared me a big disappointment.

I Will wait and see how this turns out, before ordering a card. But it does look very plausible that the problem is likely the capacotators with all these different layouts. This also truly shows who is the cheap ass manufacturer and who is more seriously. MSI surprised me while gigabyte and zotac disappointe on the capasitator layout.

It looks like as well Jay is right on asus tuf. It does have the more expensive capasitator layout for all six places.

From another side. Knowing what Jay just told. This looks really good for asus tuf card.

I Will wait and see the capacitator layout on evga FTW 3 card, before I decide. Else it looks like asus tuf card really is the go to card. Runs cool, has one the highest power target (at least for the cheaper cards at 375 watts), cooler design is great and it now also seems capasitator layout is one of the best to so far.

 

[metalfiber]

So it seems the to go to cards are asus tuf/strix and evga FTW 3. Glad I dit not ordered a card yet and spared me a big disappointment.

I Will wait and see how this turns out, before ordering a card. But it does look very plausible that the problem is likely the capacotators with all these different layouts. This also truly shows who is the cheap ass manufacturer and who is more seriously. MSI surprised me while gigabyte and zotac disappointe on the capasitator layout.

It looks like as well Jay is right on asus tuf. It does have the more expensive capasitator layout for all six places.

From another side. Knowing what Jay just told. This looks really good for asus tuf card.

I Will wait and see the capacitator layout on evga FTW 3 card, before I decide. Else it looks like asus tuf card really is the go to card. Runs cool, has one the highest power target (at least for the cheaper cards at 375 watts), cooler design is great and it now also seems capasitator layout is one of the best to so far.

On the 3090 ASUS Strix too...

 

[windwhirl]

And this is why you should never buy a product at launch.

Now, imagine that everyone followed your idea and absolutely nobody bought the card. That would be the worst launch ever

EDIT: And as someone else pointed out, eventually somebody would still get the shitty cards, so there is no escaping it.

 

[Tomgang]

On the 3090 ASUS Strix too...

View attachment 169804

Yeah I see and I also think I now know how asus can get boost clock of over 1900 mhz out of the 3080 strix oc model while most others are just below or over 1800 mhz as max. It's the capasitator layout as asus chose the most expensive but also the best for high stable factory oc cards and overclock in general. That also means thre strix card shut provide the highest Manuel overclock as long silicone and max power target allows for it. So if evga fails as well on capasitators, asus here i come. No doubt there.

ROG Strix GeForce RTX 3080 OC Edition 10GB GDDR6X | Graphics Cards

The ROG Strix GeForce RTX 3080 OC Edition 10GB GDDR6X unleash the maximum performance on NVIDIA Ampere Architecture, by featuring Axial-tech fan design, 0dB technology, 2.9-slot Design, Dual BIOS, Auto-Extreme Technology, SAP II, MaxContact Technology, and more.

www.asus.com

 

[Zealotki11er]

Asus seemed to have done their homework. They always seem to have a bad product and then overcorrect it. TUF was bad with 5700 XT but X570 TUF was one of the best as so looks like RTX 3080. Usually AIB launch their cards after Nvidia/AMD because of these reasons. They need extra time to validate their designs.

 

[KainXS]

This launch is one of the biggest GPU launch fails I can remember. Many went from wanting it, being outbid from unbeatable bots to hitting F5 for who knows how long to basically saying

 

[moproblems99]

Now, imagine that everyone followed your idea and absolutely nobody bought the card. That would be the worst launch ever

It just means someone would inevitably get the shit cards.

 

[Fluffmeister]

How many people are glad they kept their 2080ti's which is still a damn good GPU

Two years old, hated by the competiton, and still no competition.

 

[Bubster]

Nvidia Ampere-Gate...Biggest Nvidia Blunder to date

 

[mtcn77]

I did dare to make preliminary collection of RTS 3000 electrical weak points, and some one from the TPU stuff it did block my access at the topic ..... reason unproductive comments.

I couldn't agree more.
I'm not the sharpest tool in the shed, but they sometimes beat me to it and it is discouraging to see the same sort of journalistic censure somebody someone, you know who, demands from launch date ndas directed towards community forums.

 

[harm9963]

Well i will wait for sure now.

 

[dragontamer5788]

I feel like this entire thing is based on oversimplified discussion points.

POSCAP vs MLCC vs Aluminum or whatever is a very technical choice. At first you think its simple: just pick the thing with the lowest ESR and highest Capacitance, but then you start worrying about high-frequency operation (these GPUs are at 2GHz+ now). But not only that, there's temperature and more.

The article oversimplifies:

with a six POSCAP design (which are worse than MLCCs, remember)

I haven't done this stuff since I was in college. But what I do remember was pouring over capacitor spec sheets and tearing my head out trying to understand the nuanced differences between them. Just picking an MLCC capacitor series alone requires going through a giant 100+ page list, focusing on the Thermal, Frequency, Capacitance, and Resistance of your application.

"POSCAP is worse than MLCC" ?? Wut? I'm sure an 8-terminal low-inductance MLCC is superior to most others, but a cheap general-purpose MLCC may be worse. Its not like all MLCCs are made the same. Even then, "worse" in what way? POSCAPs don't lose capacitance at higher temperatures, while MLCCs are temperature-dependent and voltage-dependent (the higher the voltage and higher the temperature, the less capacitance you get).

Maybe the MLCC is better if you've found a section of the board that has superior heat-sinks / cooling, but maybe POSCAP gets better if you're in a warmer area and/or higher voltage. Like, this crap is devilishly complicated.

Heck: Even simplifying the discussion to ESR and Capacitance (ESR bad, Capacitance Good) you get utterly borked if you randomly get resonance for some stupid reason (where voltage/current "bounces" between two components, because by luck would have it... something is "ringing" at the same frequency as your capacitor). So maybe a higher ESR chip is better in those weird cases.

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

I'm sure someone out there made a mistake with capacitor selection. This is a very difficult part of high-speed electronics design. However, simplifying the discussion to "X design has 6-MLCC capacitors vs Y Design has POSCAPs" is completely useless. That level of discussion is insufficient to seriously understand the power issues going on at the sub-nanosecond scale (2GHz == 0.5 nanoseconds).

At the end of the day, you blackbox the entire decision tree and test the heck out of the electronics. If this does end up to be a capacitor issue, then it was a testing issue. Having to go under several board revisions to fix capacitor issues is like, standard EE-issues (like finding a bug in a version of the a computer program and having to issue a patch to fix it later). Your engineers are going to make that mistake, and you hope that your testing mechanisms are good enough to catch them.

----------

If anyone thinks that this job is easy, go to Murata's simsurfing site and browse around for a few minutes. That's one company's capacitor selection, mostly MLCCs. Then go to Panasonic's website, download their Capacitor tool, and search their database for POSCAPs. Then download everyone's pSpice models, and run a few simulations on LTSpice (its a free tool, you can do all of this for free).

 

[windwhirl]

It just means someone would inevitably get the shit cards.

It is inevitable. Someone has to "sacrifice" themselves for the rest of us

 

[purecain]

Well this is the first time ive been lucky with my GPU brand choice... I went with the Asus TUF OC 3090. I will update on how it performs and if i get a crash to desktop... (when it arrives) ^^

 

[rtwjunkie]

So, as an Electronics Engineer and PCB Designer I feel I have to react here.
The point that Igor makes about improper power design causing instability is a very plausible one. Especially with first production runs where it indeed could be the case that they did not have the time/equipment/driver etc to do proper design verification.


However, concluding from this that a POSCAP = bad and MLCC = good is waaay to harsh and a conclusion you cannot make.


Both POSCAPS (or any other 'solid polymer caps' and MLCC's have there own characteristics and use cases.


Some (not all) are ('+' = pos, '-' = neg):
MLCC:
+ cheap
+ small
+ high voltage rating in small package
+ high current rating
+ high temperature rating
+ high capacitance in small package
+ good at high frequencies
- prone to cracking
- prone to piezo effect
- bad temperature characteristics
- DC bias (capacitance changes a lot under different voltages)


POSCAP:
- more expensive
- bigger
- lower voltage rating
+ high current rating
+ high temperature rating
- less good at high frequencies
+ mechanically very strong (no MLCC cracking)
+ not prone to piezo effect
+ very stable over temperature
+ no DC bias (capacitance very stable at different voltages)


As you can see, both have there strengths and weaknesses and one is not particularly better or worse then the other. It all depends.
In this case, most of these 3080 and 3090 boards may use the same GPU (with its requirements) but they also have very different power circuits driving the chips on the cards.
Each power solution has its own characteristics and behavior and thus its own requirements in terms of capacitors used.
Thus, you cannot simply say: I want the card with only MLCC's because that is a good design.
It is far more likely they just could/would not have enough time and/or resources to properly verify their designs and thus where not able to do proper adjustments to their initial component choices.
This will very likely work itself out in time. For now, just buy the card that you like and if it fails, simply claim warranty. Let them fix the problem and down draw to many conclusions based on incomplete information and (educated) guess work.

It’s nice to see someone applying a little reason and common sense into this.

 

[Birdito]

Nvidia is becoming complacent

 

[windwhirl]

Nvidia is becoming complacent

That... sounds like Intel...?