Let's begin this properly.. This is the guide on how to install the thermal pads for the 1070/1080 FTW.  These thermal pads were not sent to me from EVGA.  These are my spare thermal pads from my watercooling kits, and they are .5mm each.
 
What you need:
 
Phillips screwdriver
Thermal Material (NONCONDUCTIVE/NONCAPACITIVE:  Thermal Grizzly Kryonaut, thermal grizzly hydronaut, Artic MX-4, and many others.)
2 pieces of Thermal Tape.  should be 1mm for the backplate and .5mm for the front, or slightly larger
Rubbing alcohol or Articlean 2 part TIM remover
lint free clothes or coffee filters
about 45 minutes worth of time
music in the background.  I suggest classical for those that are susceptible to relaxing with music.  
A workspace the size of a GPU.  
 
 
Step one, Remove all screws from the backplate.  (they have silver rings around them, go for those):

 
Remove the 4 spring screws.  
(Do NOT remove the four screws in the PCB unless you are taking the heat spreader off to check the thermal tape on the VRM and Vram from the factory. If you wish to check the thermal pad contact on the mosfets and VRam, remove the three on the PCb and one on the SLI finger side of the I/O bracket)

 
Turn the card over:

 
Gently twist the ACX slightly to free the TIM on the core, and lift up slightly, then slide it back from the I/O bracket slightly or it will not clear the bracket.

 
Lift the front of the cooler to expose the Fan and LED plugs, and disconnect them (I suggest fan first, since it is easier to reach at this time):

 
Clean the TIM off of the GPU core using Rubbing alcohol or Articlean 2 part TIM cleaner, and apply the thermal pad just behind the chokes, directly above the mosfets: (do NOT reapply thermal paste to the core just yet)

 
Place the cooler on the core, and make sure it fully seats.  Do not connect the fans or LEDs just yet. 

 
Remove the ACX once again, and check for Fin impressions on the thermal pad over the mosfets: (Hard to see in this picture, sorry about that)

 
Reapply Thermal paste with your desired method of application.  I prefer spreading because i KNOW it is on the entire GPU Core.  if you do the pea method or rice method, it may form an air bubble.  Air bubble will cause the GPU to shut down even when temps are low.  the thermal sensor on the core is only in one spot.. If that spot has thermal paste, it will read low temps.  If it read low temps and you get hard locks or black screens that are unrecoverable, simply repeat this step and try a new TIM method.  Again, I prefer spreading. 

 
Reattach the LED plug, then the Fan plug.  

 
Lower the front of the cooler slightly past the lip of the I/O bracket and then into place over the core making sure it seats evenly and properly.

 
Turn the card over, and reinstall the 4 spring screws.  Install them using a cross tightening pattern until they all bottom out properly.  Do not tighten one all the way down, and then the next.

 
Apply the thermal strip (I used two to make a 1mm strip, as I did not have a single 1mm strip) directly over the back of the PCB where the Mosfets are. You can see the resistors or capacitors in a perfect line.. cover them all: 

 
Should look like this:

 
Install the backplate and retighten all of the screws making sure none are missing. You should see a screw in every silver hole and 4 spring screws exposed:

 
*Update 1: Something that every user should keep in mind.  The 700 and 900 series FTW and classifieds utilized the same heat spreader and ACX style design.  They also used louder and faster fans that moved more air through the cooler and across the heat spreader under the cooler.  To get optimal temperatures, you will need to turn the fans up a little more than stock, or use a more aggressive fan profile.  A fan running at 30% is not going to dissipate heat from an air cooler like a fan running at 50%.  On the FTW, Master bios starts the fans at 60c around 35% and on the slave bios, my card runs at max 60c with 50% fan speeds.  My max observed temperature to date is 65c on the slave bios.  Turn your fans up a little and let the air cool you card off.  These thermal pads are NOT going to remove heat if air isn't moving across them.
 
**Update 2: http://www.evga.com/thermalmod/ EVGA just announced they are posting a vBIOS update for the fan speed soon. The thermal pads that EVGA is going to send out are much large than then ones I had available, and will cover the VRM as well as the chokes on the front side of the PCB, and a much larger ares on the back of the card. the EVGA guide is available here: http://www.evga.com/therm...installation_guide.pdf

Nice instruction and very useful.
 
I feel some will be apprehensive to undertake such a task and could also be put off somewhat from buying said cards, especially as they may see it as an aftermarket fix, rather than an upgrade to a functionary device?

lebel
Nice instruction and very useful.
 
I feel some will be apprehensive to undertake such a task and could also be put off somewhat from buying said cards, especially as they may see it as an aftermarket fix, rather than an upgrade to a functionary device?

It is super easy to do. People need to understand, these cards are not made of very thin glass.. Just don't drop them and everything will be fine.

If there is anything I can do to make the guide better, please let me know.

Scarlet-Tech

lebel
Nice instruction and very useful.
 
I feel some will be apprehensive to undertake such a task and could also be put off somewhat from buying said cards, especially as they may see it as an aftermarket fix, rather than an upgrade to a functionary device?

It is super easy to do. People need to understand, these cards are not made of very thin glass.. Just don't drop them and everything will be fine.

If there is anything I can do to make the guide better, please let me know.

I think your guide is fine & for someone like me, I did delid my 6700k and applied liquid pro, so it would be no bother for the likes of myself. But some will not and could not do what others find straightforward especially when it comes down to a hands on or a somewhat enthusiast level.

Nice writeup but I have a couple questions.  You say it's for the FTW cards but shouldn't this apply to all the custom PCB 1070/1080's?  Or just any card with the ACX cooler?  
 
And did you by any chance obtain any temperature measurements?  I'm still trying to figure out where the main source of the hot spotting is from.  I know it's the VRM's but is it a result of the back plate holding in too much heat without the pads on the back of the board or is the main fix from the pads on the front of the card under the cooler?  
 
I ask because I have a 1070 SC Black Edition with no back plate and I'm trying to see if this pertains to me and/or if I would benefit from the pads on the front.  Review sites keep saying any card with an ACX cooler is effected but EVGA has yet to clarify that.  Thanks.

Simple yet thorough instructions. Great job.
 
Although I have to say spreading is more likely to create air bubbles. I've always been more of a pea in the center guy. Although a line is more effective in GPUs.

FscuderiaX
Nice writeup but I have a couple questions.  You say it's for the FTW cards but shouldn't this apply to all the custom PCB 1070/1080's?  Or just any card with the ACX cooler?  
 
And did you by any chance obtain any temperature measurements?  I'm still trying to figure out where the main source of the hot spotting is from.  I know it's the VRM's but is it a result of the back plate holding in too much heat without the pads on the back of the board or is the main fix from the pads on the front of the card under the cooler?  
 
I ask because I have a 1070 SC Black Edition with no back plate and I'm trying to see if this pertains to me and/or if I would benefit from the pads on the front.  Review sites keep saying any card with an ACX cooler is effected but EVGA has yet to clarify that.  Thanks.

 
The FTW and the Classified PCB are closely "related" to each other and have nearly the same setup as it pertains to the are of the mosfets that need cooling.  This would work exactly the same for the classified card.  Since the FTW and Classified are the only two cards that are custom PCB's, then this would apply for both, but it would extend the header if I added "and Classified".  
 
I took temp measurements before and after:
 
Before:
Backplate was 38.8c after folding for 30 days.
Backplate showing 54c after an hour of running furmark before the thermal pads were placed, while the room read 21c.
 
After the thermal pads were placed, the backplate now reads 59c at its hottest spot directly over the back of the mosfets, and the room reads 24c.  The GPU core read 64c in both tests.  I apologize, but I do not have pictures of videos for that part :-(  I also do not have any really nice thermal monitoring equipment either. 

Would applying the pad to the spot under the back plate be sufficient? I have no problem with that.
However, I am very hesitant to remove the heatsinck and have to reapply the thermal paste again.
Thanks for your guide

adamchamney@yahoo.com
Would applying the pad to the spot under the back plate be sufficient? I have no problem with that.
However, I am very hesitant to remove the heatsinck and have to reapply the thermal paste again.
Thanks for your guide

It honestly wouldn't change anything at all.  Just turn your fan speed up from stock speeds.  I am using the second bios, and when I turn the fan up, the entire GPU cools off a LOT. 
 
I think people are not taking into account that the fans have to push air through the cooler, which is already dissipating heat, and then that hot air has to move across the heat spreader for the VRM.  The 780, 780ti,, 980, and 980ti classified and FTW all used this exact same type of heat spreader, and never had issues, but the fans also ran constantly and much louder.  The only change was the fan speed from those cards to these.  Move more air across the card, and it will cool off much more efficiently.
 
 

brokencross
Simple yet thorough instructions. Great job.
 
Although I have to say spreading is more likely to create air bubbles. I've always been more of a pea in the center guy. Although a line is more effective in GPUs.

 
I have used the line, the pea, and the spread method.  the times that I used the line and the pea method, there was sections of the GPU core that had absolutely no TIM covering them, so I started using the spread method and making sure it was all covered. Four 780's, four 780tis, four 980, four 980tis and now one 1070, and I have never had a GPU fail... except the time I used dry ice cooling on a 780 and killed it, but that was because of condensation lol.

Scarlet-Tech
 
I have used the line, the pea, and the spread method.  the times that I used the line and the pea method, there was sections of the GPU core that had absolutely no TIM covering them, so I started using the spread method and making sure it was all covered. Four 780's, four 780tis, four 980, four 980tis and now one 1070, and I have never had a GPU fail... except the time I used dry ice cooling on a 780 and killed it, but that was because of condensation lol.

Agree, I would also recommend the spread method as the whole chip needs to make contact, a little bubble is better than no tim around the edges. I cringe when I see the pea method used on youtube vids.