Thermal pads, is there really any difference?

I need some new 0.5mm thermal pads to go on the used water block I got. I see that there are variations with better performance at (of course) a higher price. For the VRMs I'm sure it makes a difference, for the memory likely not. I only plan to buy one kind. Is there really any tangible difference between the different quality pads available? They all appear to be Fujipoly products, but there's G-M, XRe, XRm, X-E ...

Thermal conductivity numbers range from 6 to 17, so it seems like there's a substantial difference there. I just don't know if that difference makes a difference.

I've read that the 1mm-1.5mm are usually good to get. .5mm is too thin. I grabbed this for my M.2 PCI-E SSD. You can find similar ones from Fuji which are the best pads in the globe on either PPCS or maybe Ebay. I know amazon had them but they were from FCPU. Really sucks that they are gone as that was the last few things that came to me before my next order got lost.

The thickness needed depends on the application. Although I am sure their thermal properties are slightly different, one size is not better than another, it's just fitment.

That said there is a better than most thermal pads people use but I can't for the life of me think of what it is.

Edit: yes, those fujiopoly's above are what I was as thinking of, lol!

GTXJackBauer said:

I've read that the 1mm-1.5mm are usually good to get.

Click to expand...

I can't change the thickness. That would change how the block sits against the GPU. It's a spec'd part. EK lists it as: "Replacement thermal pads: Thermal Pad A – 0.5mm (100x16mm), Thermal Pad B – 0.5mm (75x50mm)"

Source: http://www.ekwb.com/shop/EK-IM/EK-IM-3831109857106.pdf

EarthDog said:

Edit: yes, those fujiopoly's above are what I was as thinking of, lol!

Click to expand...

Fujipoly makes a whole range worth of stuff though.

See: http://www.insulfab.net/products/em...ep549-8ox79c&gclid=CI2qkcGX98UCFRSPfgodaq0Aww

The stuff we use is in the "Hi-Performance Filler" variety. XR-m is the highest performing variety. I can find it, it's available at PPC: http://www.performance-pcs.com/fuji...ze-sheet-100-x-15-x-0-5mm.html#Specifications

But so is this: http://www.performance-pcs.com/fuji...ze-sheet-100-x-15-x-0-5mm.html#Specifications

It's XR-e. Exact same size, but thermal conductivity rated at 11 instead of 17, for a little more than half the price. It's one item down from XR-m according the Fujipoly info linked above.

So the question stands, is getting the "good stuff" really worth the extra cost? Or will the second tier product perform just as well?

I've never seen any testing on that, sorry.

I wouldn't imagine it's worth it... at least it wouldn't be to me. I can't imagine it's 2x better (cost).

Yeah, I couldn't tell you but only say what I've heard and read that Fuji the king when it comes to thermal pads. I personally like Fuji and only Fuji. Its up to you really.

JeremyCT said:

So the question stands, is getting the "good stuff" really worth the extra cost? Or will the second tier product perform just as well?

Click to expand...

Ensure that the thickness stays as specified ie. Do not go thicker etc. as the thicness is specified to fill the correct spacing left between the pcb and block.

Thicker pads are actualy worse for transferring heat, the thinner the pad the better the heat transfer however you need to keep in mind the gap spacing as mentioned above. So in some cases you have to opt for thicker pads to fill a larger void.

You will not be able to notice any difference if using the cheaper thermal pads as their thermal specs are not really diiferent enough to be significantly measureable in our PC watercooling type applications.

If you want to improve a thermal pad's performance, the best thing you can do is to select the correct thickness and then smear some quality non conductive thermal grease on the actual electrical component(s) first. Now apply the pads over the components and then mount your block.

Thermal grease has way better thermal conductivity specs than thermal pads and adding thermal grease between a pad and the mating surface actually improves the overall thermal transfer performance quite substantially.

This is well known in electronic circles....and I did note that even EK mentioned this as an optional step in their instructions when I fitted an EK block to my 780Ti a while back. The VRM section on my EVGA Classified 780Ti dissipates quite a bit of heat especially when you overvolt the GPU and overdrive the VRM components.

I did see the mention of using "non-conductive TIM" in the instructions mentioned above and meant to ask about it. Does it need to go on both metal surfaces, ie the component side and block side? Or just the component side? Seems like they're saying just the component side, but I want to be sure.

I have no particular reason to push a 780ti with my current monitor (1200p), but that doesn't mean I won't try anyway.

Thermal grease has way better thermal conductivity specs than thermal pads and adding thermal grease between a pad and the mating surface actually improves the overall thermal transfer performance quite substantially.

This is well known in electronic circles....and I did note that even EK mentioned this as an optional step in their instructions when I fitted an EK block to my 780Ti a while back. The VRM section on my EVGA Classified 780Ti dissipates quite a bit of heat especially when you overvolt the GPU and overdrive the VRM components.

Click to expand...

great post (what I cut out), but about the above...

Even though the thermal properties of paste is better than pads, doesn't adding additional interface material just decrease the amount of heat that can be removed from the IC? Think of it this way... if a pad is 80% efficenct but paste is 95% efficient, wouldn't you still be dissipating less because of the two interfaces before the heatsink due to the losses of the paste? So the pads are not getting as much heat as it would with direct contact?

ED - My thinking is that the components in question are relatively high on the RA "surface roughness" scale. The tops of VRMs, memory modules, and other whatnot don't hold a candle to a CPU or GPU heatspreader or bare die surface smoothness. Liquid TIM would fill those rough voids way better than a silicone rubber TIM pad would, thus leading to greater total contact area to transfer the heat through. That may not be correct, but it's the way it makes sense in my head.

One can't argue that EK recommends it though:
"EKWB recommends using small drops of electrically non-conductive (for example: Arctic Cooling MX-2 ™, MX-4 ™ or GELID GC-Extreme™) thermal grease on each phase regulator (that is being covered with thermal pad) in order to even further improve the thermal performance of the EK-FC Titan series water block."

Source. Step 3, Installing the Water Block: http://www.ekwb.com/shop/EK-IM/EK-IM-3831109857106.pdf

I also found the specs on the stock pads. EKWB lists them at "3.0-5.0 W/mK" for both Pad A and Pad B. So the performance of the "second best" pad from Fujipoly at 11 W/mK is already over twice as conductive as the stock pads.

ALL THAT SAID, I just realized that PPC doesn't sell the XR-e pads in the larger size 60x50mm size EKWB says I need to replace the "Pad B" spots. Those sheets either come as XR-m (top stuff) or G-M which is bottom of the ladder. I'm tempted to buy 3 of the 15x100mm strips of XR-e. One would replace Pad A, the other two Pad B. It looks like it could be cut to fit just fine based on the pictures and what I have in front of me. Any particular thoughts on that last bit? Also, what do you cut these pads with? X-acto? Utility knife (razor)? Scissors?

Thanks makes sense...

JeremyCT said:

I also found the specs on the stock pads. EKWB lists them at "3.0-5.0 W/mK" for both Pad A and Pad B. So the performance of the "second best" pad from Fujipoly at 11 W/mK is already over twice as conductive as the stock pads.

Click to expand...

That's why I went with the best. You know it will at least do a better job than the stock ones.

Nonetheless, its just a thermal pad so you should be fine either way anyways.

JeremyCT said:

Liquid TIM would fill those rough voids way better than a silicone rubber TIM pad would, thus leading to greater total contact area to transfer the heat through.

Click to expand...

Exactly... Standard thermal paste/grease is meant to go on very thin and its only job is to fill in the small voids and air gaps. Thermal pads are no where near as effective in this department.

By applying thermal paste/grease to the sandwidch of materials you actually improve the overall thermal transfer performance between the pad and whatever it is touching.

Ideally you should apply it to both sides of the pad but sometimes this can be tricky as the pad wants to slide all over the place and you can risk having the pad slide out away from its correct poition.

As a suggestion, I would only bother with applying the TIM between the component and pad as this is where the majority of thermal transfer advantage will be gained. As you lower your block to the pcb, the pads will have fairly good contact to the blocks flat surface and the fact that the pads will stick to the block will prevent them sliding about at any future stage.

Pads are used primarily to fill relatively large gaps where your typical thermal paste will not stay in place (and is less efficient for this purpose) or to ease assembly/manufacture where ultimate thermal transfer performance is not required.

You don't see thermal pads being utilised on high performance CPU or GPU mount applications as they are not all that great in transferring heat as compared to thermal greases....

From what I have heard the guys at EK did some (unpublished) testing of thermal pads and did not find a significant enough difference from the high cost Fuji pads. If there was I would assume EK would use them as standard just like they include Gelid GC extreme as thermal paste.

Lots of users tested these and saw a big difference (though I would tend to trust EKs over most users). I am wondering if they chose not to use them because of additional cost?

EK has never been inexpensive. It's possible (probable?) that they saw similar or better gains with the dab of liquid TIM than by using a better pad and that's why it's included in the directions.

EK is already overpriced as they dominate the GPU/MB waterblock market so imagine adding premium priced Fujipoly pads, I guess they feel it would just be crossing the line as if they haven't already in pricing.

I started dressing down the card yesterday and found something interesting. There was a greasy substance on the vRAM and MOSFET components that had thermal pads on them from the factory. It would seem that EVGA knows the same thing EK does and puts some type of TIM on their chips before applying the pads, at least on their more expensive offerings.

Personally I prefer using K5-PRO. It is the only thermal paste on the market (that I know) that can be used instead of thermal pads. You won't have to consider the measures cause you just put it and it takes the shape your pc needs.

There are also some videos that show how to apply it:

Hope this helped

The amounts that guy gooped on there made me cringe.