EVGA Pascal cards hotspot problem

Nelly · 30 Oct 2016 at 00:33

I may as well add this to the thread, because on the EVGA forums some mods / fan boys are trying to play down the issue, especially with the SC versions.

Guru3D tested 1070's with the VRM's using 3DMark Firestrike, using thermal imaging equipment. The EVGA 1070 SC Gaming hit 96℃ where the VRM's are located, that's between 27℃ to 39℃ hotter than three of it's competitor brand cards.

See below for results.
VRM = 64℃ - Gigabyte GeForce GTX 1070 G1 GAMING >> http://www.guru3d.com/articles_pages/gigabyte_geforce_gtx_1070_g1_gaming_review,10.html

VRM = 57℃ - ASUS GeForce GTX 1070 STRIX Gaming >> http://www.guru3d.com/articles_pages/asus_geforce_gtx_1070_strix_gaming_review,10.html

VRM = 69℃ - MSI GeForce GTX 1070 Gaming Z >> http://www.guru3d.com/articles_pages/msi_geforce_gtx_1070_gaming_z_review,9.html

VRM = 96℃ - EVGA GeForce GTX 1070 SC Gaming >> http://www.guru3d.com/articles_pages/evga_geforce_gtx_1070_sc_superclocked_gaming_review,10.html

mikeo · 30 Oct 2016 at 01:31

~>Dg<~ said:
http://wccftech.com/nvidia-gtx-1080-1070-evga-cards-dying/

I certainly like the emotive language... God help us, we are all going to be burnt to death :eek:

I particularly like the bit about "flames gushing from his graphics card". Now where did I put that fire extinguisher ?

Dicehunter · 30 Oct 2016 at 01:56

Calabi said:
I'm wondering if I've got a problem with my card now. I've been playing Witcher 3 and after a while, my monitor turns of with a signal not found message. Its happened several times. Alt tabbing seems to fix it, but maybe its drivers or something.

Do you have a FTW or Classified ?

Gerard · 30 Oct 2016 at 02:14

~>Dg<~ said:
http://wccftech.com/nvidia-gtx-1080-1070-evga-cards-dying/

Funny how this is occurring after there has been a fault with them found by a review site.

Simon Belmont · 30 Oct 2016 at 08:14

What is EVGA's UK contact number and email address. I would like to talk to them.

slade · 30 Oct 2016 at 08:44

Damn.... My PC's been randomly shutting down/ restarting while playing games in the last week and i've only just upgraded to 2 EVGA 1070 SCs.

They're not over heating as i keep an eye on them in MSI AB, fans manual at 80% getting 55 - 65°. Should i be worried?!

Tommy_Here · 30 Oct 2016 at 08:53

People shouldn't have to strip apart their card to apply this 'mod' fact is it should have been applied during manufacturer process point is it wasn't to save costs and now it's backfiring literally. So after waiting months for the evga cards to come out they do this.... so glad I cancelled my 1080 order after waiting over 2 months and now I've got a pair of 980 Tis for the same price.

Mei · 30 Oct 2016 at 08:56

evga what did u do ;_;

Calabi · 30 Oct 2016 at 09:10

Dicehunter said:
Do you have a FTW or Classified ?

Its the FTW.

Bone9 · 30 Oct 2016 at 09:14

The thing is I've got a good overclocker and wouldn't want to risk being sent a different card in return with worse performance.

Also why should we have to be without a card while this gets sorted?

JediFragger · 30 Oct 2016 at 09:18

mikeo said:
As above.

But EVGA have already stated in writing (see my post above):

" In our discussion with EVGA's Jacob Freeman on the phone, we also learned that EVGA would replace users' cards at no charge to the user. An owner would have to send the device in to EVGA in order to receive a new device with the thermal pads pre-applied. "

So I can't personally see how EVGA could back track on this. It's not like Jacob Freeman is just some office junior, he's a senior product manager for EVGA.

Raising an RMA on EVGA's WEB site is dead easy and their response time is usually pretty good (from what I've read).

I'll be interested to see how many folks do this. Could be that the flood gates are just about to open for EVGA!

Great info, thanks. I have 2 thankfully so no system downtime when I send back

Do we know a date from when the pads were being applied? Just to check the date when the cards get back.

mikeo · 30 Oct 2016 at 10:17

JediFragger said:
Great info, thanks. I have 2 thankfully so no system downtime when I send back

Do we know a date from when the pads were being applied? Just to check the date when the cards get back.

As far as I'm aware, they will start doing this to all new cars produced at some point in the future. I get the impression that they have not started doing this yet. Or if they have, it's only in the last few days.

For those that want to do it themselves.... Instructions for self fitting should be up sometime next week, with pads starting to shift sometime the week after that. Though I do appreciate that a fair few folks won't want to do this themselves (for a number of reasons). For the others, there would appear to be the RMA route. Not ideal situation, I think we can all agree on this. But sometimes **** happens and you've just got to sort it out.

Drollic · 30 Oct 2016 at 10:40

I'm in two minds as to if to rma. I don't like the idea of taking apart what is essentially now a £700 product even though the pads are easy to fit. I've seen some real horror shows with people taking cards apart, even down to being unlucky with the reapplication of thermal paste on the core resulting in failure weeks later. My core and memory clock fine and I've got no noticeable coil whine.

Toss3 · 30 Oct 2016 at 10:56

Drollic said:
I'm in two minds as to if to rma. I don't like the idea of taking apart what is essentially now a £700 product even though the pads are easy to fit. I've seen some real horror shows with people taking cards apart, even down to being unlucky with the reapplication of thermal paste on the core resulting in failure weeks later. My core and memory clock fine and I've got no noticeable coil whine.

How can you be unlucky when it comes to applying thermal paste? I constantly see people recommending a dot in the middle of the gpu, but that doesn't work for most coolers, especially with those that have heatpipes. A very thin layer, equally spread out over the die would be my recommendation, as that way you can make sure it makes contact and the layer isn't thick.

Also seeing high VRM temps on my 290x and trying to find a solution; hitting 96c, which really isn't ideal and I often see black or red screens when playing bf1. :eek:

harney · 30 Oct 2016 at 12:41

Nelly said:
I may as well add this to the thread, because on the EVGA forums some mods / fan boys are trying to play down the issue, especially with the SC versions.

Guru3D tested 1070's with the VRM's using 3DMark Firestrike, using thermal imaging equipment. The EVGA 1070 SC Gaming hit 96℃ where the VRM's are located, that's between 27℃ to 39℃ hotter than three of it's competitor brand cards.

See below for results.
VRM = 64℃ - Gigabyte GeForce GTX 1070 G1 GAMING >> http://www.guru3d.com/articles_pages/gigabyte_geforce_gtx_1070_g1_gaming_review,10.html

VRM = 57℃ - ASUS GeForce GTX 1070 STRIX Gaming >> http://www.guru3d.com/articles_pages/asus_geforce_gtx_1070_strix_gaming_review,10.html

VRM = 69℃ - MSI GeForce GTX 1070 Gaming Z >> http://www.guru3d.com/articles_pages/msi_geforce_gtx_1070_gaming_z_review,9.html

VRM = 96℃ - EVGA GeForce GTX 1070 SC Gaming >> http://www.guru3d.com/articles_pages/evga_geforce_gtx_1070_sc_superclocked_gaming_review,10.html

Agree the evidence is clear there running hotter due to poor design (the missing pads)

there does seem to be a hell of a lot of defending evga and i can understand this as the majority are most likely evga card owners themselves..
This is understandable but what is not acceptable is the fact Evga can sell cards at these prices (UK) and we have these issues its totally unacceptable for such a small oversight missing the thermal pads ...but the fact is clear they have dropped the ball... they should have used thermal pads ...

Now what i hear a lot on the forums is this "my FTW temps are fine" ect but as we know its nothing to do with gpu temps and the VRM VRAM temps can not be monitored via software has to be done physically ....this can be very misleading for peeps...And this is where the problems lie if the end user could monitor such info via software i bet there would be more of an uproar seeing there VRM temps hit +100c and VRAMS over spec +95c ...

I have had 2 clients evga cards with me this weekend 1070 ftw and a 1080 both purchased about 2 months ago...these have now been fitted with my own bought thermal pads and i can conclude that doing so helps ..

I tested both cards to the best of my ability with only being able to use a Laser heat gun yes these are not very accurate but i was able to at least get an idea of the heat around the hot spot area.

Just in normal gaming for 30+ mins the hot spot area's where high up into the 91c for the 1070 with the 1080 being hotter at 94c this could be even hotter again not sure how accurate my heat gun is . this was on a open test case amb temp 18 / 19... now imagine in a closed case with poor air flow and higher ambient temps hell even in SLI what would be the temps of the hot spots for the top card well i dread to think.never mind peeps running over-clocks and to think FTW cards are marketed as over clocking cards too.

Now i did not want to run anything like furmark ect as i would imagine the temps would go even higher and these not being my cards i do not want to cause any long term damage ....So is it ok peeps getting these pads to fit themselves which again i find unacceptable for the price paid for these cards in which the end user is then having to finish the design off...Reminds me of flat back furniture..

the other problem i have found is that after gaming the fans would slow down very quickly... with the heat around the hot spot still being high and as the vrm's and vrams are cooled via air this is not a good design and i imagine a lot of users are using the standard out of the box fan profile i think this is set too low for less noise reasons .

Back to Placing the thermal pads on...

Removing the sinks both of heat sinks fins have scorched metal where the VRM's are located same as the photo i posted in my previous post... this should not happen period...

the results after the pads fitted much better 1080 76c 1070 71c with standard out the box fan profile ..

set a more aggressive custom curve the temps drop even further ...

So i will stick with my original theory that evga have messed up here with there design all from missing some simple thermal pads ...now what i am not sure about was it designed that way or was it a mistake who knows..

either way the price paid for these cards you would think they would have got this right.....

Five.Stars · 30 Oct 2016 at 14:11

Were you able to get the ACX cooler back on again without much hassle? I found that after removing my ACX from SC to replace with Hybrid, that the length of cable on the LED and fan headers would prove a right pain for re-fitting the cooler if ever required.

Calabi · 30 Oct 2016 at 14:36

I'm really suspicious it is a heat issue now, it seems Witcher 3 is a game that pushes my gpu and general temps the highest. BF1 in comparison limited to a 100fps and gpu usage fluctates a lot and temps still stay under 70. Witcher 3 however even limited to 60fps it has high gpu usage and temps go up to 75 plus.

I'm wondering if I should test it further and risk damaging it or just not use it heavily and wait for thermal pads to come.

Nelly · 30 Oct 2016 at 14:43

Nelly said:
I may as well add this to the thread, because on the EVGA forums some mods / fan boys are trying to play down the issue, especially with the SC versions.

Guru3D tested 1070's with the VRM's using 3DMark Firestrike, using thermal imaging equipment. The EVGA 1070 SC Gaming hit 96℃ where the VRM's are located, that's between 27℃ to 39℃ hotter than three of it's competitor brand cards.

See below for results.
VRM = 64℃ - Gigabyte GeForce GTX 1070 G1 GAMING >> http://www.guru3d.com/articles_pages/gigabyte_geforce_gtx_1070_g1_gaming_review,10.html

VRM = 57℃ - ASUS GeForce GTX 1070 STRIX Gaming >> http://www.guru3d.com/articles_pages/asus_geforce_gtx_1070_strix_gaming_review,10.html

VRM = 69℃ - MSI GeForce GTX 1070 Gaming Z >> http://www.guru3d.com/articles_pages/msi_geforce_gtx_1070_gaming_z_review,9.html

VRM = 96℃ - EVGA GeForce GTX 1070 SC Gaming >> http://www.guru3d.com/articles_pages/evga_geforce_gtx_1070_sc_superclocked_gaming_review,10.html

harney said:

Agree the evidence is clear there running hotter due to poor design (the missing pads)

there does seem to be a hell of a lot of defending evga and i can understand this as the majority are most likely evga card owners themselves..
This is understandable but what is not acceptable is the fact Evga can sell cards at these prices (UK) and we have these issues its totally unacceptable for such a small oversight missing the thermal pads ...but the fact is clear they have dropped the ball... they should have used thermal pads ...

Now what i hear a lot on the forums is this "my FTW temps are fine" ect but as we know its nothing to do with gpu temps and the VRM VRAM temps can not be monitored via software has to be done physically ....this can be very misleading for peeps...And this is where the problems lie if the end user could monitor such info via software i bet there would be more of an uproar seeing there VRM temps hit +100c and VRAMS over spec +95c ...

I have had 2 clients evga cards with me this weekend 1070 ftw and a 1080 both purchased about 2 months ago...these have now been fitted with my own bought thermal pads and i can conclude that doing so helps ..

I tested both cards to the best of my ability with only being able to use a Laser heat gun yes these are not very accurate but i was able to at least get an idea of the heat around the hot spot area.

Just in normal gaming for 30+ mins the hot spot area's where high up into the 91c for the 1070 with the 1080 being hotter at 94c this could be even hotter again not sure how accurate my heat gun is . this was on a open test case amb temp 18 / 19... now imagine in a closed case with poor air flow and higher ambient temps hell even in SLI what would be the temps of the hot spots for the top card well i dread to think.never mind peeps running over-clocks and to think FTW cards are marketed as over clocking cards too.

Now i did not want to run anything like furmark ect as i would imagine the temps would go even higher and these not being my cards i do not want to cause any long term damage ....So is it ok peeps getting these pads to fit themselves which again i find unacceptable for the price paid for these cards in which the end user is then having to finish the design off...Reminds me of flat back furniture..

the other problem i have found is that after gaming the fans would slow down very quickly... with the heat around the hot spot still being high and as the vrm's and vrams are cooled via air this is not a good design and i imagine a lot of users are using the standard out of the box fan profile i think this is set too low for less noise reasons .

Back to Placing the thermal pads on...

Removing the sinks both of heat sinks fins have scorched metal where the VRM's are located same as the photo i posted in my previous post... this should not happen period...

the results after the pads fitted much better 1080 76c 1070 71c with standard out the box fan profile ..

set a more aggressive custom curve the temps drop even further ...

So i will stick with my original theory that evga have messed up here with there design all from missing some simple thermal pads ...now what i am not sure about was it designed that way or was it a mistake who knows..

either way the price paid for these cards you would think they would have got this right.....

Click to expand...

I've included the VRM temps from the 1080 reviews on Guru3D, however they have not reviewed any EVGA products on the 1080 line of cards as yet. The 1080's run slightly hotter across the brands, so it's obvious EVGA 1080 cards would also.

VRM = 67℃ - Gigabyte GeForce GTX 1080 G1 GAMING >> http://www.guru3d.com/articles_pages/gigabyte_geforce_gtx_1080_g1_gaming_review,10.html

VRM = 68℃ - Palit GeForce GTX 1080 GameRock Premium Edition >> http://www.guru3d.com/articles_pages/palit_geforce_gtx_1080_gamerock_premium_edition_g_panel_review,11.html

VRM = 76℃ - ASUS ROG Strix GeForce GTX 1080 >> http://www.guru3d.com/articles_pages/asus_rog_strix_geforce_gtx_1080_review,10.html

VRM = 76℃ - MSI GeForce GTX 1080 Gaming Z >> http://www.guru3d.com/articles_pages/msi_geforce_gtx_1080_gaming_z_8g_review,10.html

VRM = 88℃ - ZOTAC GeForce GTX 1080 AMP! >> http://www.guru3d.com/articles_pages/zotac_geforce_gtx_1080_amp_extreme_review,10.html

VRM = 96℃ - EVGA GeForce GTX 1070 SC Gaming >> http://www.guru3d.com/articles_pages/evga_geforce_gtx_1070_sc_superclocked_gaming_review,10.html

mikeo · 30 Oct 2016 at 14:46

harney said:
Agree the evidence is clear there running hotter due to poor design (the missing pads)

there does seem to be a hell of a lot of defending evga and i can understand this as the majority are most likely evga card owners themselves..
This is understandable but what is not acceptable is the fact Evga can sell cards at these prices (UK) and we have these issues its totally unacceptable for such a small oversight missing the thermal pads ...but the fact is clear they have dropped the ball... they should have used thermal pads ...

Now what i hear a lot on the forums is this "my FTW temps are fine" ect but as we know its nothing to do with gpu temps and the VRM VRAM temps can not be monitored via software has to be done physically ....this can be very misleading for peeps...And this is where the problems lie if the end user could monitor such info via software i bet there would be more of an uproar seeing there VRM temps hit +100c and VRAMS over spec +95c ...

I have had 2 clients evga cards with me this weekend 1070 ftw and a 1080 both purchased about 2 months ago...these have now been fitted with my own bought thermal pads and i can conclude that doing so helps ..

I tested both cards to the best of my ability with only being able to use a Laser heat gun yes these are not very accurate but i was able to at least get an idea of the heat around the hot spot area.

Just in normal gaming for 30+ mins the hot spot area's where high up into the 91c for the 1070 with the 1080 being hotter at 94c this could be even hotter again not sure how accurate my heat gun is . this was on a open test case amb temp 18 / 19... now imagine in a closed case with poor air flow and higher ambient temps hell even in SLI what would be the temps of the hot spots for the top card well i dread to think.never mind peeps running over-clocks and to think FTW cards are marketed as over clocking cards too.

Now i did not want to run anything like furmark ect as i would imagine the temps would go even higher and these not being my cards i do not want to cause any long term damage ....So is it ok peeps getting these pads to fit themselves which again i find unacceptable for the price paid for these cards in which the end user is then having to finish the design off...Reminds me of flat back furniture..

the other problem i have found is that after gaming the fans would slow down very quickly... with the heat around the hot spot still being high and as the vrm's and vrams are cooled via air this is not a good design and i imagine a lot of users are using the standard out of the box fan profile i think this is set too low for less noise reasons .

Back to Placing the thermal pads on...

Removing the sinks both of heat sinks fins have scorched metal where the VRM's are located same as the photo i posted in my previous post... this should not happen period...

the results after the pads fitted much better 1080 76c 1070 71c with standard out the box fan profile ..

set a more aggressive custom curve the temps drop even further ...

So i will stick with my original theory that evga have messed up here with there design all from missing some simple thermal pads ...now what i am not sure about was it designed that way or was it a mistake who knows..

either way the price paid for these cards you would think they would have got this right.....

Interesting and informative post.

I think most people, if not all, will agree with your comments \ observations. I certainly do.

The only comment I take slight objection to, is about there appearing to be a lot of people on this thread "defending" EVGA. I personally would say just the opposite. I and most others consider that EVGA have dropped the ball with this issue. And in an ideal world this should not have happened. But unluckily it has and we now as customers have to decide on what we want to do. There are going to be a number of options open to people to remedy this situation and it's really up to each individual as to what course of action to take.

Maybe the retailer should offer a replacement card, or your money back. I believe ocuk offered I think it was an upgrade to a 980 when the 970 memory issue rose it's ugly head. Or maybe a cheap step-up from EVGA. Or maybe a free extended warranty. I really think EVGA is going to have go do something to restore consumers trust.

Me... I shall just fit the "optional" (which I'm starting to think are anything but "optional") thermal pads (done similar mods a couple of times in the past) and then get on with my life.

Thanks again for the information above post☺

Dave2150 · 30 Oct 2016 at 14:47

Toss3 said:
How can you be unlucky when it comes to applying thermal paste? I constantly see people recommending a dot in the middle of the gpu, but that doesn't work for most coolers, especially with those that have heatpipes. A very thin layer, equally spread out over the die would be my recommendation, as that way you can make sure it makes contact and the layer isn't thick.

Also seeing high VRM temps on my 290x and trying to find a solution; hitting 96c, which really isn't ideal and I often see black or red screens when playing bf1.

290X is an old card, it's probably faulty and in need of RMA/replacement.

Though worth noting that AMD used very high temperature rated VRM's on the 290 series, as they knew they were such incredibly hot and power hungry GPU's.

The VRM's used on the NVIDIA 10 series are not designed to run as hot, hence why they are catching fire and burning out at high temperatures.