EVGA GTX 1060/1070/1080 overheating issues

SalemSaberhagen said:

As well as always using a far more aggressive fan curve than stock.

That area of the card still got hot enough to discolor the cooling fins of the card.

Click to expand...

Im very curious to know how long did it take to get that nice tan ?

justinNWI said:

Except it's the VRM overheating, not anything you can monitor via software. Thought I'd chip in here so that people are inspired to make the mod, as most cards are probably suffering from this.

Click to expand...

Wait you can't monitor the vrm using HWinfo or something? Don't those cards have the sensors?

Some cards, though not many, have sensors so you can read vrm through software.it is Usually reserved for high end models, bUT again, is not common.

Kenrou said:

Im very curious to know how long did it take to get that nice tan ?

Click to expand...

I've only owned the card maybe three months. The fans were barely dusty.

SalemSaberhagen said:

I never "had any issues" but there's a flaw in the design.

Thankfully prior to watercooling my card I had used a case with a 140 side fan blowing on it.

As well as always using a far more aggressive fan curve than stock.

That area of the card still got hot enough to discolor the cooling fins of the card.

View attachment 184557

Click to expand...

By looking at the design of the cooler:
- Air is drawn in from outside of the card
- Pushed through the fins on the cooler
- Exhausted over the heatplate on the card and out of the top and bottom of the card

How did this miraculous heat from the VRM manage to fight the upstream airflow and only deposit it's "heat" in only the spot above the VRMs?

How did this vrm heat make it to the other side of the card (left on your picture?)

How did this vrm heat make it to the GPU side of the fins (middle right on your picture?)

Have you completed your metallurgical analysis to actually verify that this discoloration is caused by heat from the VRMs?

Is it more probable that since the area you circled is connected directly to a large heat pipe coming off the GPU cooling plate, that you could have a higher temperature due to GPU heating?

JrClocker said:

By looking at the design of the cooler:
- Air is drawn in from outside of the card
- Pushed through the fins on the cooler
- Exhausted over the heatplate on the card and out of the top and bottom of the card

How did this miraculous heat from the VRM manage to fight the upstream airflow and only deposit it's "heat" in only the spot above the VRMs?

How did this vrm heat make it to the other side of the card (left on your picture?)

How did this vrm heat make it to the GPU side of the fins (middle right on your picture?)

Have you completed your metallurgical analysis to actually verify that this discoloration is caused by heat from the VRMs?

Is it more probable that since the area you circled is connected directly to a large heat pipe coming off the GPU cooling plate, that you could have a higher temperature due to GPU heating?

Click to expand...

There's a heat spreader between the card and the cooler that makes contact with the vrm section and the memory.

EarthDog said:

Some cards, though not many, have sensors so you can read vrm through software.it is Usually reserved for high end models, bUT again, is not common.

Click to expand...

You'd think these days a GPU in the 1070/1080 region would have vrm temperature monitoring, even my cheapo nitro 4GB RX 480 has it. Seems EVGA dropped the ball on these for the sake of some inexpensive thermal pads.

Saul Goodman said:

You'd think these days a GPU in the 1070/1080 region would have vrm temperature monitoring, even my cheapo nitro 4GB RX 480 has it. Seems EVGA dropped the ball on these for the sake of some inexpensive thermal pads.

Click to expand...

The Nitro is an overclocked model on a custom PCB, IIRC. Not the cheapest 480 out there by any means.

ATMINSIDE said:

The Nitro is an overclocked model on a custom PCB, IIRC. Not the cheapest 480 out there by any means.

Click to expand...

All the RX 480 cards have vrm temperature monitoring even reference. Maybe it's because they run hot though who knows

They do?

I wouldn't bet my life on it, but I don't think the msi I reviewed did? As was said, it's not really common to see that value monitored in many gpus.

EarthDog said:

They do?

I wouldn't bet my life on it, but I don't think the msi I reviewed did? As was said, it's not really common to see that value monitored in many gpus.

Click to expand...

Maybe the MSI card is an exception not sure, I know that the reference, Powercolor, XFX, Sapphire and HIS all have VRM temperature sensors.

That leaves MSI, Asus and Gigabyte as unknown and seeing as they are more expensive cards I'd assume they have a vrm sensor, then again I've heard really bad things about the Gigabyte card so not sure on that one.

It's a really useful temperature to monitor for obvious reasons not sure why it's neglected so often.

And apologies for going off topic.........

Can you link a source to that info? I would like to know for sure.

For some reason this annoyed me this evening and I opened the box and aimed my potato best I could.



If you cannot see some heat cause issue there?

Unfortunately, I seriously took like 40 pictures, this is the best capture of the effect, it's blued metal, not dust!



But so out of focus, but the COLOUS ARE TRUE

Could be heat from the 150 W GPU heating up the fins on the heat sink (meaning...it's doing it's job).

Could also be a "not so good job" with the welding of the heat sink to the fins.

Could be "junk" in the air (when you had an external fan blowing into it) that discolored the fins (the GPU is 70-ish C after all).

Without a picture of the fins with the card being brand new...it's really hard to say what is the cause.

If the temperature of your VRMs was so hot that it would cause aluminum to discolor...the thermal pads would be melted and the card would not be functioning. Why is it just that one area and not all VRMs? The cooling mechanism is the same...the VRM current is balanced.

You can see the indentation of the VRMs into the thermal pads on the black heat plate. The fans suck air in, blow over the cooling fins, blow air onto the black cooling plate, and then the air leaves by the top and bottom of the card (i.e. the direction of the heat sink fins.)

Again - for heat from the black heat plate to fight "upstream" against the airflow from the fans is very unlikely.

EVGA will honor their warranty if the card poops out.

JrClocker said:

Could be heat from the 150 W GPU heating up the fins on the heat sink (meaning...it's doing it's job).

Could also be a "not so good job" with the welding of the heat sink to the fins.

Could be "junk" in the air (when you had an external fan blowing into it) that discolored the fins (the GPU is 70-ish C after all).

Without a picture of the fins with the card being brand new...it's really hard to say what is the cause.

If the temperature of your VRMs was so hot that it would cause aluminum to discolor...the thermal pads would be melted and the card would not be functioning. Why is it just that one area and not all VRMs? The cooling mechanism is the same...the VRM current is balanced.

You can see the indentation of the VRMs into the thermal pads on the black heat plate. The fans suck air in, blow over the cooling fins, blow air onto the black cooling plate, and then the air leaves by the top and bottom of the card (i.e. the direction of the heat sink fins.)

Again - for heat from the black heat plate to fight "upstream" against the airflow from the fans is very unlikely.

EVGA will honor their warranty if the card poops out.

Click to expand...

WTF?

There is no welding.

No junk causes blueing shy of random Thermite ignition amd ooh, 70ish C, so that's hot, we'll get back to that.

If the temperature of your VRMs was so hot that it would cause aluminum to discolor...the thermal pads would be melted and the card would not be functioning. Why is it just that one area and not all VRMs? The cooling mechanism is the same...the VRM current is balanced.

Click to expand...

A second ago 70C was hot, tomshardware.de recorded temps over 105C in that area.

Fight a fan upstream, we're talking RADIANT HEAT of significant levels, no fan is going to push that anywhere.

I'm posted about this issue on several sites and this is the only one INCLUDING EVGA's I get fanboyish denial?

No need to get upset...sell the card if you are worried.

Not denial...physics.

I assumed welding. If there is no welding, then how is the aluminum heat fin attached to the heat pipe?

The primary heat transfer mechanism is convection, not radiation.

At standard atmospheric pressures (i.e. not in a vacuum) radiant heat will not dominate the convection heating cause by the AIR FLOW of the high CFM output of the fans...not at 100 C anyway (if that's the actual temperature.)

Don't believe me?

1. Heat up a piece of metal (can use a match) until it's red hot
2. Place this red-hot piece of metal in front of a standard PC fan (that is running of course)
3. You won't feel the heat from the red-hot metal behind the fan

Your own picture posted in post #16 of this thread...the discoloration goes pretty far down...a "long" ways from the VRM location.

Another approach...Let's do the math...assuming the card is drawing 180 W of power:

- Let's assume that this power is consumed in only the GPU and VRMs (to keep the analysis simple).
- If the VRMs are 90% efficient (which would be a crappy VRM), then they dissipate 18 W.
- This means that the GPU is dissipating 180 - 18 = 162 W.
- If it's a FTW card, there are 10 power phases.
- With 90% efficiency, each VRM phase is dissipating 1.8 W

SOOOOOO:
- How does the heat generated by a 1.8 W VRM cause spot discoloration when there is 162 W of heat being generated by the GPU?

I'm not denying the discoloration. I am doubtful that the root cause is a super-hot VRM (or VRMs)...and just trying to get you to think of other possible causes.

Again - I have a 1080 FTW and I am not worried.

Friday math?

If the card is using 162W across 10 phases, isn't that 16.2W per phase assuming it's divided evenly?

In the above example, it's true that each VRM is supplying 16.2 W of power to the GPU.

However, the amount of power being dissipated in the VRM is directly related to their efficiency...not the total power draw of the board. 90% efficiency means that they will dissipate 10% of the power they are delivering. So, 180 W in, 162 W out...18 W total in the VRMs...1.8 W per VRM for 10 phases.

It's the same concept as the efficiency of a PC power supply. If the PC power supply is 90% efficient, it delivers 90 W, pulls 100 W from the wall, and dissipates 10 W during the conversion process.

Make sense?

SalemSaberhagen said:

There's a heat spreader between the card and the cooler that makes contact with the vrm section and the memory.

Click to expand...

The thermal pad mod applies on the back plate to cool the VRM. The heat spreader is in contact with the VRM with thermal pads stock. So the VRM is not heating the cooler fins directly, the VRM heat is conducted by the heat spreader.

The test used in the referenced review from Toms Hardware (Germany) is running under Furmark, an extreme usage case, as most overclockers know. We believe this is a good approach to have some idea about the graphics card limit, and the thermal performance under the worst case scenario. EVGA has performed a similar qualification test during the design process, at a higher ambient temperature (30C in chamber) with a thermal coupler probe directly contacting the key components and after the Toms Hardware (Germany) review, we have retested this again. The results in both tests show the temperature of PWM and memory is within the spec tolerance under the same stress test, and is working as originally designed with no issues.

With this being said, EVGA understands that lower temperatures are preferred by reviewers and customers.

During our recent testing, we have applied additional thermal pads between the backplate and the PCB and between the baseplate and the heatsink fins, with the results shown below. We will offer these optional thermal pads free of charge to EVGA owners who want to have a lower temperature. These thermal pads will be ready soon; and customers can request them starting Monday, October 24th, 2016. Also, we will work with Toms Hardware to do a retest. http://www.evga.com/thermalmod/

Click to expand...

I contacted EVGA by phone and they say there not having any problems with RMA caused by heat and there is a lot of misinformation on the web. He also said cards are running well within specifications and the maximum temperature for the VRM 120c

EVGA SUPPORT PHONE NUMBERS

Americas
888.881.3842
714.528.4500 Europe 49.89.189.049 11

JrClocker said:

In the above example, it's true that each VRM is supplying 16.2 W of power to the GPU.

However, the amount of power being dissipated in the VRM is directly related to their efficiency...not the total power draw of the board. 90% efficiency means that they will dissipate 10% of the power they are delivering. So, 180 W in, 162 W out...18 W total in the VRMs...1.8 W per VRM for 10 phases.

It's the same concept as the efficiency of a PC power supply. If the PC power supply is 90% efficient, it delivers 90 W, pulls 100 W from the wall, and dissipates 10 W during the conversion process.

Make sense?

Click to expand...

Completely.

Isn't the other energy that IS doing something (16.2W) also generating heat?

Please don't misunderstand me, I agree with you, just making sure I have my head wrapped around the fundamentals of why.