EVGA responds to hot VRM area on GTX 10 series

[mbryant]

According to an EVGA rep I spoke with. Tom's Hardware ran furmark for 2 hours. If you haven't used furmark there is even a warning about prolonged usage. 

[zMeul]

2 minutes ago, mbryant said:

According to an EVGA rep I spoke with. Tom's Hardware ran furmark for 2 hours. If you haven't used furmark there is even a warning about prolonged usage. 

people that reported their cards ending up in smoke weren't anywhere close to FurMark

 

and here's some other thing: EVGA's own OC Scanner uses the same stability testing idea that FurMark uses

[zMeul]

GamersNexus has done some testing: http://www.gamersnexus.net/news-pc/2666-evga-heat-solution-thermal-imaging-of-vrm-1080-ftw

[MageTank]

43 minutes ago, zMeul said:

people that reported their cards ending up in smoke weren't anywhere close to FurMark

 

and here's some other thing: EVGA's own OC Scanner uses the same stability testing idea that FurMark uses

Another fun fact: EVGA's OC Scanner doesn't even work. Seriously, give that thing a try. Now, my 1070 SC is stable at 2050mhz, cannot exceed it. My 1070 FTW DT is stable at 2150, can't exceed that either without artifacts/driver crashing. However, OC Scanner overclocked it to 2075, said it was okay, even as it crashed during it's own test. That thing is utter garbage, and should not be used in place of a real manual overclock.

 

As for the first statement about people saying they were not running Furmark, it goes both ways. If you were running Furmark, and fried your card, would you admit to it? Or would you lie, out of fear that your warranty might be voided otherwise? Not that I am saying it's impossible for the card to fry under normal circumstances (highly unlikely, but not impossible I suppose, depending on ambient temps and airflow inside of case) but I have seen people fry different components in the past on their own merit, and lie to get them replaced. It's not like the manufacturer could prove you wrong either way, as they can't really tell what you were doing the moment it died.

 

Another side note: I'll be testing copper shims (along with the EVGA supplied thermal pads) on the VRM to see which solution is better. Since you can get them fairly cheap (and different thicknesses) https://www.amazon.com/JIUWU-Chipset-Thermal-Heatsink-Copper/dp/B00OUJQX8K/ref=sr_1_4?s=pc&ie=UTF8&qid=1478071494&sr=1-4&keywords=copper%2Bshim&th=1

 

Interested in seeing if CLU on the bare die (along with shims/pads on the VRM) will aid in my overclocking endeavors. Luckily, have had no issues out of either of my 1070's. Then again, I've used custom fan curves on both since day one.

[DrMikeNZ]

29 minutes ago, MageTank said:

As for the first statement about people saying they were not running Furmark, it goes both ways. If you were running Furmark, and fried your card, would you admit to it?

I wouldn't run Furmark myself, but if I did, and it killed the GPU I would call up the store and pretend to be as stupid and incompetent as the act of running furmark requires.

If I were to run OC Scanner (which isn't a very good program either) and it killed my GPU, I would gladly admit to EVGA that I ran it.

 

You don't need to assume the worst in everyone though, some people just get the bad luck of receiving defective products. When the card is caught on video frying during POST, it is unlikely due to be caused by furmark. The majority of the fried cards are most likely a separate issue to the VRM cooling design fault.

[Toha]

8 hours ago, zMeul said:

update Nov 02: http://www.evga.com/thermalmod/

 

cards affected:

 

cards not affected: Founders Edition, blower type, HYBRID and HYDRO COPPER

What has me confused is I have one of the 1080s listed there except my 1080's part number has a BR after it so I don't know if that means I'm not affected or something.

[zMeul]

6 hours ago, DrMikeNZ said:

I wouldn't run Furmark myself, but if I did, and it killed the GPU I would call up the store and pretend to be as stupid and incompetent as the act of running furmark requires.

If I were to run OC Scanner (which isn't a very good program either) and it killed my GPU, I would gladly admit to EVGA that I ran it.

 

You don't need to assume the worst in everyone though, some people just get the bad luck of receiving defective products. When the card is caught on video frying during POST, it is unlikely due to be caused by furmark. The majority of the fried cards are most likely a separate issue to the VRM cooling design fault.

EVGA themselves ran FurMark

and this isn't 1st time FurMark was used to show some big design fails in the power delivery system

[zMeul]

here's someone else opening it's EVGA GTX1080 FTW - the thermal pads over the VRAM do not make contact!! https://imgur.com/a/iIDjD + https://imgur.com/a/9iSGb

 

 

he bought two, same issue: https://www.reddit.com/r/nvidia/comments/5aqjvb/so_i_disassembled_my_two_evga_gtx_1080_ftws/

[Techicolors]

10 hours ago, MageTank said:

Another fun fact: EVGA's OC Scanner doesn't even work. Seriously, give that thing a try. Now, my 1070 SC is stable at 2050mhz, cannot exceed it. My 1070 FTW DT is stable at 2150, can't exceed that either without artifacts/driver crashing. However, OC Scanner overclocked it to 2075, said it was okay, even as it crashed during it's own test. That thing is utter garbage, and should not be used in place of a real manual overclock

yeah i ran OC scanner as well and didn't do much to improve or auto overclock for my 1070 SC. just the app constantly crashes 

[Arcmanov]

I've used a custom fan-profile from the day I installed my 1070 SCs as well, and I usually game with the AC on, so I'm not overly worried about temps to be bothered with applying the BIOS update, but I am concerned enough about long-term longevity, so I requested the 'thermalmod' from EVGA.

 

Question though...

Can you get your own thermal pads and apply them? 

What brand/size/thickness would be best for this?

[Kinda Bottlenecked]

20 hours ago, mbryant said:

According to an EVGA rep I spoke with. Tom's Hardware ran furmark for 2 hours. If you haven't used furmark there is even a warning about prolonged usage. 

 

Tom's card didn't die. 

 

Cards that burnt belonged to your avg gamer playing normal games. Last one i heard was shadow warrior 2.

[Techicolors]

Jayztwocents made a video about this issue

 

 

I personally didn't watch the video but his comment in the comments section is interesting, to say the least
 

Quote

 

Can't believe I have to post this... but for those people out there who live in conspiracy theory lala land...

1) I am the one who reached out to EVGA for comment. Since I am local they invited me over to discuss the issue with the CEO personally.

2) This is NOT a sponsored video. In fact I didn't even ad one of my ads on this video.

3) I stood right there and told the CEO that I am shocked that this was an issue, that they had always taken this stuff in to account and asked how they failed at this... he looked me right in eye and said "Jay, we messed up"

 

You guys think I'm taking cash and sugar coating this issue?? Y'all are crazy. EVGA deserves the praise for handing the issue with dignity as much as they deserve the outcry that this was even an issue in the first place. If they had done nothing you would be boycotting... but when the company is trying to do the right thing some of you still want to claim that they don't care about the customer which simply isn't true. I've seen companies that don't care, and this isn't one of those instances. I told the CEO personally that I am happy that they are handling this properly because advertising partners or not doesn't sway my opinion on the matter and he agreed with me. Anyway it goes without saying this is not a sponsored video. This video is for my audience to learn what the issue is and how to apply a fix. Once again ignorant comments force me to ignore the comments section once again. If you want to contact me twitter is where you will find me

 

 

[zMeul]

8 minutes ago, Technicolors said:

I personally didn't watch the video but his comment in the comments section is interesting, to say the least

the testing they're doing now, should've been done way before the cards went to market

 

there is a lot of talk on reddit and the note I'm getting is that people will avoid buying EVGA in the future

they skimped on few cents worth of thermal pads - this will hurt them, and I won't shed a single tear for them

 

and it's not over ..

new reports show that the thermal pads that should've contacted the VRAM are actually too thin

[Techicolors]

10 minutes ago, zMeul said:

the testing they're doing now, should've been done way before the cards went to market

 

there is a lot of talk on reddit and the note I'm getting is that people will avoid buying EVGA in the future

they skimped on few cents worth of thermal pads - this will hurt them, and I won't shed a single tear for them

 

and it's not over ..

new reports show that the thermal pads that should've contacted the VRAM are actually too thin

i probably won't be getting an EVGA card for the next generation cards (or next refresh). for now i kinda have to deal with it. i appreciate the fact that they're doing something about it, but they obviously should have tested this 

 

funny enough when i was looking at which 1070 to get back in the summertime, i found the 1070 SC review by guru3D and noticed how hot the VRM hot spots were, but they claimed that it was still okay, just not the best. perhaps i shouldn't have glazed over that 

[zMeul]

2 hours ago, Arcmanov said:

Question though...

Can you get your own thermal pads and apply them? 

What brand/size/thickness would be best for this?

from what I've seen, the pads EVGA sent are 2mm thick

[mbryant]

so what should I do regarding my 1070 SC? I don't have an FTW card. I know EVGA is issuing a patch, should I cross RMA my card? This card was an upgrade not only for gaming but WORK as I have a show I'll be editing soon and there will be many hours of exporting and rendering in the coming months

[Paragon_X]

I just did a visual check on mine and it seems the pads are making contact. But I will check it more when i open the card to add the pads.

And Fucking EVGA doesnt pick up calls nor answering emails, and they still haven't approved my request for the pads even though I am registered and all.

I am really pissed about this, i bought this card because supposedly EVGA is a premium manufacturer and the second reason was the supposedly good support and i have received neither of those two things. I just ordered pads my self to mod it (EK 2.0mm pads).

 

[Mira Yurizaki]

Here's what I don't get.

 

People claim MSI's Gaming X does this better. Except their chokes are naked like EVGA's. So I'm wondering if MSI's Gaming X card is really better.

 

EDIT: The chokes on the Founder's Edition are also naked

So here are my questions about the whole deal:

• What part is EVGA using for the chokes, vs. what MSI is using and what the FE cards are using?
• Under what exact conditions are applied to the affected cards when they start having chokes ramp up to 105C?
  • This is useful to know in case this is an edge or corner case, or if this has actual issue for most use cases.

[Paragon_X]

3 minutes ago, M.Yurizaki said:

Here's what I don't get.

 

People claim MSI's Gaming X does this better. Except their chokes are naked like EVGA's. So I'm wondering if MSI's Gaming X card is really better.

One of the problems of the ACX 3.0 is the fact that it does not push enough air over the VRMs hence the problem arised. You have to crank it at 30% + to have something going on

[Mira Yurizaki]

4 minutes ago, Paragon_X said:

One of the problems of the ACX 3.0 is the fact that it does not push enough air over the VRMs hence the problem arised. You have to crank it at 30% + to have something going on

Then how is MSI's any different when I'm almost certain those cards operate under a similar fan curve?

 

I think we're picking on one company without looking at the others to make sure they're not doing anything wrong either. And I don't really feel like finding if anyone else compiled data for the other cards.

 

It would be like solely singling out VW for dieselgate. It was later found out that a lot of other car manufacturers weren't exactly telling the truth either.

 

EDIT: It just feels like to me since EVGA is held in high regard that people are looking for any excuse to bring them down. You know, the nail that sticks out the most gets pounded in first.

Edited November 3, 2016 by M.Yurizaki

[zMeul]

48 minutes ago, M.Yurizaki said:

Here's what I don't get.

 

People claim MSI's Gaming X does this better. Except their chokes are naked like EVGA's. So I'm wondering if MSI's Gaming X card is really better.

 

EDIT: The chokes on the Founder's Edition are also naked

So here are my questions about the whole deal:

• What part is EVGA using for the chokes, vs. what MSI is using and what the FE cards are using?
• Under what exact conditions are applied to the affected cards when they start having chokes ramp up to 105C?
  • This is useful to know in case this is an edge or corner case, or if this has actual issue for most use cases.

I think they're using the chokes as heatsinks

[d0nsen]

1 hour ago, M.Yurizaki said:

Then how is MSI's any different when I'm almost certain those cards operate under a similar fan curve?

 

I think we're picking on one company without looking at the others to make sure they're not doing anything wrong either. And I don't really feel like finding if anyone else compiled data for the other cards.

 

It would be like solely singling out VW for dieselgate. It was later found out that a lot of other car manufacturers weren't exactly telling the truth either.

 

EDIT: It just feels like to me since EVGA is held in high regard that people are looking for any excuse to bring them down. You know, the nail that sticks out the most gets pounded in first.

you should take a job at samsung managing their pr lol you cant talk down shit going up in flames

also if you fuck up and dont get caught its bad enough but since when its that an excuse for fucking up and getting caught? 

looking for any excuse to bring them down lol this is 400$+ hardware people bought and its bursting flames

[Mira Yurizaki]

8 minutes ago, zMeul said:

I think they're using the chokes as heatsinks

Then that would be a problem for all of the graphics cards because EVGA's chokes are right next to the MOSFETs, just like everyone else's layout.

 

1 minute ago, d0nsen said:

you should take a job at samsung managing their pr lol you cant talk down shit going up in flames

also if you fuck up and dont get caught its bad enough but since when its that an excuse for fucking up and getting caught? 

looking for any excuse to bring them down lol this is 400$+ hardware people bought and its bursting flames

I'm not saying that this is an excuse for screwing up, but I'm still questioning why people are singling out EVGA when everyone's board and cooling layout is basically the same. Why haven't people investigated MSI, Gigabyte, or ASUS?

 

Besides, did everyone ditch Intel in droves when it was found out that certain versions of Prime95 causes Haswell processors to die from overvolting? No.

 

And the more I'm looking into this, the more this is an edge case that I'm finding. Tom's Hardware ran Furmark for an extended period of time. Furmark is not a realistic load to run on the GPU and is basically considered a power virus (or at least it simulates one).

[zMeul]

Just now, M.Yurizaki said:

Then that would be a problem for all of the graphics cards because EVGA's chokes are right next to the MOSFETs, just like everyone else's layout.

everyone else cooled their FETs, EVGA "forgot"

 

and there's something else ....

if you look at the FLIR images you'll see that the bottom part of the card is getting hotter, that means the VRM is not evenly distributing the load uniformly across the FETs

also, all the cards that blew a gasket are damaged in the lower area

Edited November 3, 2016 by zMeul

[Mira Yurizaki]

12 minutes ago, zMeul said:

everyone else cooled their FETs, EVGA "forgot"

Going by the pictures you posted and this PCB image:

EVGA is cooling the MOSFETs, because the pads were clearly indented.

 

For reference, this is MSI's PCB layout (the MOSFETs are right of the chokes):

http://www.guru3d.com/index.php?ct=articles&action=file&id=22264

 

Though they did at least put fins on the heatsink.

 

EDIT: Actually looking at the PCB label, those chips left of the chokes have a C in it, implying they're capacitors, not MOSFETs.