Nvidia GTX 1080 & 1070 EVGA Cards Reportedly Catching Fire & Dying Due To VRMs Overheating

Shintai said:
Micron is memory, not VRM. And they would simply make artifacts long before dying.
Click to expand...
Right, but the proximity of the VRAM to the MOSFETs is causing the temperature of the former to go higher from exposure. It is possible for the VRAM to exceed the 95C maximum if the VRM has a >10C delta for extended periods of time.
 
Armenius said:
Right, but the proximity of the VRAM to the MOSFETs is causing the temperature of the former to go higher from exposure. It is possible for the VRAM to exceed the 95C maximum if the VRM has a >10C delta for extended periods of time.
Click to expand...

But all you get is artifacts tho.
 
are the MOSFETs on the back of the board ? Cooling the MOSFETs from the backplate usually involves direct contact (through pads) to the rear of the PCB behind the power MOSFETs. So you are drawing heat through the PCB. Nowhere near as good as direct VRM cooling.
 
Earlier EVGA had issue with heatpipes design on 970ACX. Now they have issues with at her side of their cards. Maybe on 3rd gen of their cooler and 1170 they will make both sides working :)

It is the same "cut all costs, because there is like 0,01% chance that something goes wrong" that happened with Note 7. You can go around many laws when designing something. But you can never go around Murphy laws :)
 
TaintedSquirrel said:
This guy fell into a burning ring of fire...
Although please note, he was doing a boot which rules out overheating.

Click to expand...

I'm sure putting the sound card in the slot directly underneath the video card isn't helping matters with cooling, though.
 
Armenius said:
I'm sure putting the sound card in the slot directly underneath the video card isn't helping matters with cooling, though.
Click to expand...

Yeah that is the worst position possible for the sound card. It's like running SLI with both of those cards right below each other, one just sucks up heat from the one below.
 
Furious_Styles said:
Yeah that is the worst position possible for the sound card. It's like running SLI with both of those cards right below each other, one just sucks up heat from the one below.
Click to expand...

People have been running 4 way SLi in the past without issues with blowing up their cards. :)
 
cageymaru said:
People have been running 4 way SLi in the past without issues with blowing up their cards. :)
Click to expand...
Smart people run quad SLI with blower cards, though, which the FTW is not. Stacking blowers together is not much of an issue provided the heatsink has a vapor chamber.
 
Armenius said:
I'm sure putting the sound card in the slot directly underneath the video card isn't helping matters with cooling, though.
Click to expand...

PCI cables are not plugged in which means that card is inactive, so it's not on yet it catches fire. And is that the sound of a match or something?
 
thesmokingman said:
PCI cables are not plugged in which means that card is inactive, so it's not on yet it catches fire. And is that the sound of a match or something?
Click to expand...

Yep, its a hoax video it seems. Its certainly not related to this.
 
Terpfen said:
You seem… uneducated on this issue. The problem is that the VRMs do not make contact with the backplate. The backplate is not the ACX cooler. Don't be willfully ignorant.
Click to expand...
Rofl what?

The issue is that VRMs and cooler don't make any contact (simply because there is no contact area on ACX 3.0 cooler for VRMs to speak of). Backplate has literally nothing to do with it. A swift look at comparison of coolers and EVGA's proposition to install thermal pads would quickly lead you to same conclusion. So i gotta ask.

Who is the ignorant one?
 
The heatplate is part of ACX's design and marketing material (Read the bottom-left lol)...

EVGA's thermal pad promotion includes BOTH backplate pads and frontplate pads which go between the heatplate and fins.

D2pszLV.jpg
 
lolfail9001 said:
Rofl what?

The issue is that VRMs and cooler don't make any contact (simply because there is no contact area on ACX 3.0 cooler for VRMs to speak of). Backplate has literally nothing to do with it. A swift look at comparison of coolers and EVGA's proposition to install thermal pads would quickly lead you to same conclusion. So i gotta ask.

Who is the ignorant one?
Click to expand...

Going to be a little amusing when EVGA sends out those thermal pads and you think the instructions are telling you to put them on the wrong side of the card.
 
Can anyone confirm that EVGA actually fixed the issue on all cards shipped since August?
I assume all you have to do is look under the fins to see if a pad is there.
 
Armenius said:
Um...
Click to expand...
thesmokingman said:
Seriously? That's not the backlate wtf? That's a vrm heatsink incorporated into the cooler. That's not even the FTW cooler lmao.
Click to expand...

I suppose you guys did not follow context.

Terpfen said:
You seem… uneducated on this issue. The problem is that the VRMs do not make contact with the backplate. The backplate is not the ACX cooler. Don't be willfully ignorant.
Click to expand...

Here is context. And yes, it is not ACX , it is STRIX's cooler (as you can tell by useless heatpipe). But i think contrasting it with ACX demonstrates the reason behind VRM heat nicely.
 
lolfail9001 said:
I suppose you guys did not follow context.



Here is context. And yes, it is not ACX , it is STRIX's cooler (as you can tell by useless heatpipe). But i think contrasting it with ACX demonstrates the reason behind VRM heat nicely.
Click to expand...
I gotcha. Terpfen seems to think that because the MOSFETs sit on the heatsink side that signs of failure couldn't happen on the backside.
 
Armenius said:
I gotcha. Terpfen seems to think that because the MOSFETs sit on the heatsink side that signs of failure couldn't happen on the backside.
Click to expand...

Nah, these guys are just creating circular arguments. Look below. It's not the cooler, its that they had a fail at EVGA and a line of cards left w/o thermal pads on the vrm. The cooler itself works fine. One could say the cooler is a fail because otherwise the vrm would not have burned yea, true but that's really ignoring the crux of the issue.

lolfail9001 said:
That means it is precisely a screw up with the cooler if it fails to cool the card properly.
Click to expand...
 
thesmokingman said:
Nah, these guys are just creating circular arguments. Look below. It's not the cooler, its that they had a fail at EVGA and a line of cards left w/o thermal pads on the vrm. The cooler itself works fine. One could say the cooler is a fail because otherwise the vrm would not have burned yea, true but that's really ignoring the crux of the issue.
Click to expand...
Well, when put like that, that's perfectly just. But i think you'll agree with me that thermal pads should be considered a part of cooling solution.
 
Armenius said:
I think that just goes to show you how far simply adding thermal pads can go.
Click to expand...
Just to be clear the EVGA heatplate has thermal pads under it.
The difference shown here is that MSI has pads on the chokes, and fins on their heatplate. However the heatplate itself is a lot smaller.

EVGA's backplate does NOT have thermal pads, but I can't find any pictures of MSI's backplate. It's safe to assume they do have pads because they aren't retarded like EVGA.
 
TaintedSquirrel said:
Just to be clear the EVGA heatplate has thermal pads under it.
The difference shown here is that MSI has pads on the chokes, and fins on their heatplate. However the heatplate itself is a lot smaller.

EVGA's backplate does NOT have thermal pads, but I can't find any pictures of MSI's backplate. It's safe to assume they do have pads because they aren't retarded like EVGA.
Click to expand...
card2.jpg
I am not sure thermal pads are supposed to be here :p
 

I dunno what this means, but posting it here to ask if its crock or if maybe he is correct.
 
SighTurtle said:

I dunno what this means, but posting it here to ask if its crock or if maybe he is correct.
Click to expand...

That guy is on a soapbox. The way the acx 3.0 is made is the way a lot of gpus have been made for years. Reference cards on the other hand tend to have the vrm heatsink incorporated into the cooler as standard because that's the way they ave always been made. Reference cards are usually a cast one piece frame with an opening for the vapor chamber or cooling element for the core. Open fan coolers however eschew that for a cooling plate that goes over all the vrm/memory ic, with the actual cooler mounting over the core. The open fan cooler then stretches the length of the cooling plate and the fans push air down over the heatpipes/fins and the cooling plate, thereby cooling the vrm/memory ic.

These soapboxers are using the argument that the open coolers are not direct vrm cooling integrated into the cooler thus it is zomg, going to fry burn. It take a healthy dose of delusion to get where they are imo.
 
TaintedSquirrel said:
Can anyone confirm that EVGA actually fixed the issue on all cards shipped since August?
I assume all you have to do is look under the fins to see if a pad is there.
Click to expand...

I'd love to know this since I just got my RMA back.
 
If you have the card, check it yourself...
Get a flashlight and shine it under the heatsink where the VRM's are located. You should see THREE sets of pads:

One between the heatplate and heatsink.
One between the VRM / PCB and heatplate.
One between the rear PCB and backplate.

From top to bottom:

Heatsink
Pad
Heatplate
Pad
VRM
----------
PCB
----------
Pad
Backplate
 
EVGA claiming that Furmark damages cards makes them look really bad. Both amd and nv have dealt with the power usage / temperature aspect, and drivers no longer have any need to check for binary names wrt that old "power virus" issue.
 
xorbe said:
EVGA claiming that Furmark damages cards makes them look really bad. Both amd and nv have dealt with the power usage / temperature aspect, and drivers no longer have any need to check for binary names wrt that old "power virus" issue.
Click to expand...

Yes it makes them look bad, especially when other manufacturers connect the VRM cooling to the primary heat sink:
1070strix-8b.jpg


(Strix 1070/1080 cooler for reference, the gray thermal pad on the left cools the VRM circuits)
 
Commander Shepard said:
I contacted EVGA about doing an RMA on my 1070 due to the black screen errors I've gotten possibly related to this issue. The response follows:

"The entire outcry came from a single review, where the reviewers ran Furmark, a test known to cause damage to cards, for 2 hours, which is about 8 times longer than would be recommended for that particular software. Nvidia and AMD have both implemented limits in their drivers to throttle the performance of the cards when they detect Furmark running, due to the damage it causes and we do not recommend running it, as it does not have any bearing on the performance or durability of the cards in the real world, as it does things that no other application or game would ever do. We are offering to send these to customers who are willing to install them, so we can address customers concerns about this. We have retested the cards at higher ambient temperatures while using thermal couplers, to get a more accurate reading of the temperatures and found no issues with overheating, nor were we able to reproduce any of the reported issues under any kind of use we put the cards through.

I can assure you that any reports of burnt components and other potential issues are wholly unrelated to VRM thermal pads and could be related to issues with another component in your system. I would highly recommend testing the card in another system, to ensure that one of your other hardware components are not the cause of the issue."

Comments from my far more technically astute than I brethren?
Click to expand...
Just got another response from EVGA. I'm guessing my original inquiry got pushed up to someone with a little more seniority.

"Hello,

Thank you for taking the time to troubleshoot this on your own and we are sorry the problem did not resolve itself.

We completely understand and we wish to work with you on a solution. We can offer warranty support replace the card and install the thermal pads.

Instead of the standard RMA options, I would like to offer you a complimentary Advanced RMA ( https://www.evga.com/support/ear/ ).

Please upload a copy of your purchase invoice to your product page ( https://www.evga.com/invoice ) and verify that your address is correct on your EVGA account so that we may proceed.

Thank you,
EVGA"

Nice touch by EVGA, offering the Advanced RMA. (y)
 
Commander Shepard said:
Just got another response from EVGA. I'm guessing my original inquiry got pushed up to someone with a little more seniority.

"Hello,

Thank you for taking the time to troubleshoot this on your own and we are sorry the problem did not resolve itself.

We completely understand and we wish to work with you on a solution. We can offer warranty support replace the card and install the thermal pads.

Instead of the standard RMA options, I would like to offer you a complimentary Advanced RMA ( https://www.evga.com/support/ear/ ).

Please upload a copy of your purchase invoice to your product page ( https://www.evga.com/invoice ) and verify that your address is correct on your EVGA account so that we may proceed.

Thank you,
EVGA"

Nice touch by EVGA, offering the Advanced RMA. (y)
Click to expand...

That's great customer service! :)
 
Some are saying this has "always been an issue with the ACX cooler". Yet there were no such issues on EVGA Kepler/Maxwell cards. Could the VRMs on those cards just handle higher temps or isn't it rather an issue specific to ACX 3.0 & Pascal?

I'm not worried when it comes to my 980 ti either way, I got a lot of fresh air being blown at the VRMs with my 140mm fans.
 
TaintedSquirrel said:
This guy fell into a burning ring of fire...
Although please note, he was doing a boot which rules out overheating.

Click to expand...


What I find funny about that video besides the card having no power connected, is I have that exact motherboard (Sabertooth X79). When I'd play a Frostbite game, I'd always get these alarms from ASUS AI Tweaker that the VTTCPU temp was getting too high, indicative of that top PCIe x16 slot.
 
You must log in or register to reply here.
Back
Top