• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

PowerColor 6700XT Fighter Teardown

Let the fun commence. :)

GPUltimate.jpg

Permafrost2.jpg
 
Bah! Hit a snag, the trials of being first I guess...

So, 4mm on the rear is too thick. Who would have thought 1mm would be too much. I've ditched the x2 2mm, and replaced them with x2 1.5mm. 3mm is a good fit, there is indentation into the pad material. 4mm causes the pcb to twist and bulge and yeah.... Just not good.

Now my snag. 1.5mm on the front is no good. It stops the heatsink from making full contact with the GPU core. I thought it would compress down on the memory modules but I think the spring loaded screws simply means they can't tighten any further than they are designed for. And 1.5mm means it has a small gap. When I took the heatsink off, only a small corner of paste had transferred across. (the corner furthest from the memory modules)

I've had to order the 1mm pads. Should have them for Monday, perhaps Saturday hopefully.

I'll then pop some pictures up to help explain my ramblings above!
 
Firegod said:
Bah! Hit a snag, the trials of being first I guess...

So, 4mm on the rear is too thick. Who would have thought 1mm would be too much. I've ditched the x2 2mm, and replaced them with x2 1.5mm. 3mm is a good fit, there is indentation into the pad material. 4mm causes the pcb to twist and bulge and yeah.... Just not good.

Now my snag. 1.5mm on the front is no good. It stops the heatsink from making full contact with the GPU core. I thought it would compress down on the memory modules but I think the spring loaded screws simply means they can't tighten any further than they are designed for. And 1.5mm means it has a small gap. When I took the heatsink off, only a small corner of paste had transferred across. (the corner furthest from the memory modules)

I've had to order the 1mm pads. Should have them for Monday, perhaps Saturday hopefully.

I'll then pop some pictures up to help explain my ramblings above!
Click to expand...

Bah, reading this post felt like watching an episode of something you are really enjoying only to realise it's the end of the current season, ended on a cliffhanger and now have to wait to find out what happened . Roll on Monday
 
Rejoice and re-heat your popcorn @Th0nt. My 1.0mm pads arrived today. Build on! :) So it has taken a little while to pull the figures together, hopefully the graphs are easy to understand, but, if not, I have summarised my findings. But first some pictures.

1. 4.0mm pad issues (rear)

Here we can see what happens to the PCB when using 4.0mm pads on the rear. Remember, I measured this at pretty much 3.0mm and so figured 4.0mm would work fine. Guess not! I had to remove these and I have replaced them with 3.0mm.

Pads-on-rear.jpg


4-0mm-bending.jpg


2. 1.5mm pad issues (front)

And here is the result of using a 0.5mm thicker pad on the front - notice that the paste only transferred on the corner furthest away from the pads on the memory modules.

Paste-Problem.jpg


I have replaced these with the 1.0mm pads today.

1-0mm-pads-applied.jpg


3. Fresh Paste

And some fresh Permafrost 2 paste on the GPU. I had a little go over the edge of the GPU but it isn't conductive so I wasn't too worried. It is runny out of the tube compared to the NT-H1 but yet, harder to spread.

Permafrost-Paste-applied.jpg


4. Temperature Results

Temp-Graphs.jpg


Just annoyingly my GPU hot spot temps are slightly higher than before - yet overall GPU temps are down (especially the maximum temperatures recorded) - but I wasn't expecting much given that the GPU thermal paste I used previosuly was NT-H1. I guess for the GPU hot spot temps, the heatsink must still not be seated quite right. But, the memory junction temps are down across the board which is really what we were wanting to achieve here by swapping out the thermal pads..

To summarise the temps.

The Average temperatures saw the following:
  • GPU temperature is 1.3 degrees cooler in Days Gone.
  • GPU temperature is 9.6 degrees cooler in The Ascent.
  • GPU temperature is 1.2 degrees warmer in No Man's Sky.
  • GPU temparture is 0.1 degrees cooler in FurMark.
  • Memory Junction temperature is 6.0 degrees cooler in Days Gone.
  • Memory Junction temperature is 10.7 degrees cooler in The Ascent.
  • Memory Junction temperature is 1.5 degrees cooler in No Man's Sky.
  • Memory Junction temperature is 1.5 degrees cooler in FurMark.
And the Max temperatures saw the following:
  • GPU temperature is 1.0 degrees cooler in Days Gone.
  • GPU temperature is 3.0 degrees cooler in The Ascent.
  • GPU temperature is 5.0 degrees warmer in No Man's Sky.
  • GPU temparture is 1.0 degrees cooler in FurMark.
  • Memory Junction temperature is 5.8 degrees cooler in Days Gone.
  • Memory Junction temperature is 6.1 degrees cooler in The Ascent.
  • Memory Junction temperature is 6.3 degrees cooler in No Man's Sky.
  • Memory Junction temperature is 2.0 degrees cooler in FurMark.
So I am pleased overall. :) The Memory temperatures for sure have responded really well here with the better pads front and back (I'm no official benchmark tester I just wanted to get a quick and dirty feel for how the change went). GPU temps are down slightly as well (but GPU hot spot temps are similar if not increased slightly).

I will certainly next focus on the fan profile for the card and see if we can't get a little more airflow going at higher temperatures.

Overall cost wise, once I had the right pads, I think it's cost me just under £50. Worth it - I'd say so.
 
Last edited:
Well done, decent results. As for the Hotspot temps being worse, did you go in an X pattern, half turn when tightening the retention mechanism? Failure to do this can result in a less than perfect mount.
 
LtMatt said:
Well done, decent results. As for the Hotspot temps being worse, did you go in an X pattern, half turn when tightening the retention mechanism? Failure to do this can result in a less than perfect mount.
Click to expand...

I am still seeing a very minute gap on 3 of the 4 standoffs (one each corner of the GPU area). I wonder if I can tighten it down a smidge more. Failing that, I may crack it open again and see how the paste has transferred and see if that sheds any light. But yeah, I did the X pattern but must admit I may have rushed the tightening. Could explain it. Cheers
 
You must log in or register to reply here.
Back
Top Bottom