Thermal Paste Roundup - Performance, Price, Recommendations (Update: Thermal Pads!)

I see then ill go with the arctic as recommended by most. Thank you!

 

For those interested in the application of thermal pads to NVIDIA boards, MXM format, this thread is very informative and contains very nice pictures:

http://forum.notebookreview.com/thr...71dm-p775dm-thermal-paste-replacement.789825/

Cheers

 

Arctic?
Don't.
I suggest you use GeLid GC Extreme instead (far superior).

 

Are you talking towards the thermal paste or pads???

 

Thermal paste

 

He was referring to thermal pads.

Enviado de meu Pixel 2 usando Tapatalk

 

but gc extreme is thermal paste?

 

We know. But @Deks missed the previous posts, hence g c extreme was put in the discussion

GP-EXTREME is different product!!

 

Oops xD

 

And likely to be horrible overpriced too.
Why pay for a tiny thin rectangle strip when you can get a 145x145mm sized pad for $15 dollars? Plus if you cut it wrong, you wasted that part and have to do it again. Better to have more to work with, than less!!

 

I unfortunately went to get the gelid thermal pads in the end. The arctic ones were out of stock in my area.

 

As a modder you end up having these sorts of pads on hand but they can be quite tricky to source quickly.

 

Thanks a lot for the TIM shootout!

 

Tuniq's - TX-4 is great paste too, I'd put it just under Gelid Extreme.

grizzly Minus 8 pads are a must IMO. They beat the others in pricing and performance is perfect, and more re-usable then the super expensive pads like Fujipoly which are not even that great to be honest and extremely difficult to use.

 

thx for your input!

the thermal grizzly pads are included in the OP. to be honest, im not a big fan of them because they perform worse than the arctic thermal pads but youre paying the TG price premium. so theres basically no use for them, even though in "theory" they sport a higher thermal conductivity at 8 vs the arctic's 6 W/mK. and im not just getting pads for a name brand, id rather stick to hard performance data


Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

Is Intel’s Thermal Paste ACTUALLY That Bad (vs Noctua NT-H1)?
Published on Mar 24, 2018
By harvesting TIM from old Intel CPUs we can finally find out just how bad it is.

 

Linus tech tips are a friggin joke, more so then Intel's Paste.

 

I know I should have commented here, but I commented on youtube instead, as I didn't think there was any point.
tl;dr: if Linus were competent, he would have taken those Ivy bridge (or haswell?) chips, tested them as is with temps, recorded the results, then delidded them and applied NT-H1 inside the IHS, relidded them back (now here's another sticking point: you would want to relid without having the IHS with even more pressure on the core, as a more tightly mounted IHS=lower temps also) and then compared temps. That's a true applies to applies comparison. Or at least the best realistic test you could get. of course it would require the chips to be working, but is that a problem considering how much free hardware he gets?

 

That was the point he was trying to make, you are never going to get the lid back on in the same place so moving the compound to a repeatable setting was a better option.

It's a flawed test either way to be fair.

 

@Meaker@Sager I don't follow your logic.
It's not hard to 1) buy a delid tool. 2) take off the thermal goop. 3) apply that NT-H1 paste on the bare die CPU and a very light coating under the IHS. 4) buy some RTV silicone and use small dabs in the corner, and relid the IHS (there are relid tools incuding Der8aeur's own tool, however you spell his name, 5) test. And yes, you can relid the IHS BACK in the exact same spot. it's been done countless times already.

Then you record temps BEFORE the delid, and after.

You don't anything except a little common sense and some hard, cold logic, to do that.
People have done this already on OCN (But usually with poor longevity results, but with improvement over the stock material; most pastes are not made for direct die cooling over extreme heat stresses like that)

 

the point was probably that paste swap under the ihs has been done to death / countless times already. so LTT wanted to do something "new" and get lots of views. ikr pretty shocking for a YT channel

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

Ah okay. Publicity rules :/

 

If not putting lid in the same palce would result in better temps, then the lid should've been put there first place.

 

Most likely worse temps as the sealant will be thicker. The Intel stuff is fairly thin.

 

Is there a possilbility to use CPU without IHS? Just like BGAs in laptops do?

 

Just use a small amount of super glue on each corner when re-lidding. It doesn't increase z-height and make temps worse like the thick RTV silicone does, and is easy to dissolve using acetone (nail polish remover). Rockit Cool recommends the Loctite Gel Control super glue.

 

With a specially designed shim soldered to the heatsink and a retention plate round the die to hold it down properly you could. It's a fair amount of engineering needed.

 

Yes you can but it's not worth it. Your temps won't improve very much at all. Just delid the CPU, put Conductonaut between CPU and IHS (coat the underside of the IHS , right above the shape of the CPU slug, with conductonaut also for best heat transfer, tests on overclock net proved that doing this gives lower temps than not coating the underside of the IHS), then relid. RTV silicone gasket sealer works well for relids; just put 1 tiny drop on each corner of the IHS edge for relidding.

As long as your heatsink base/plate is copper or nickel plated (NOT ALUMINUM, PERIOD), you can also apply conductonaut between IHS and heatsink. Putting LM here requires extra prep work to protect against LM runoff away from the IHS area; I would suggest a cutout foam dam made of VERY VERY compressible, and very lightly dense foam (it must be compressible enough to the width of a human hair in your fingers with no effort at all--anything denser will cause resistance which is bad). Here's an example of foam used like this.

(This foam dam is NOT needed if you are using traditional thermal paste between IHS and heatsink and only liquid metal for the delid and relid on the CPU core).

 

I was actually thinking about custom heatsink to cover the die directly. It would be actually a nice to see option in desktop CPU laptops.

 

@Falkentyne - Wouldn't some silicone gasket work just as well (or better)?

 

Work for what? A dam? or for sealing the IHS? I've never seen or heard of gasket being used as a 'dam' barrier, and seems a bit dangerous for LGA chips for sure. K5 Pro has been used as an effective barrier for BGA CPU's, but anything extremely thick will work for that, even good old classic radio shack thermal paste or Arctic Ceramique (that super white thick stuff!); just paint it around the BGA chip housing and you're good. Can be used in combination with nail polish insulation and Super 33+ tape.

For LGA chips? I would not use anything like that because it has to go around the actual socket now, and thus gets exposed to the PCB. Would only use foam dams here. Other stuff would only work if the housing is completely closed (like bare die CPU slug or GPU slug around its housing).

For Delids, you should only use nail polish for insulation (for the exposed contact traces on the CPU package) then RTV dabs to re-lid.

 

A solid with no real give demands precision shaping otherwise it would impact contact pressure/fit.

 

I really want to see some test results with that new thermal thin pad made of carbon that can be used now.

 

Read... http://forum.notebookreview.com/thr...mal-pad-available-for-test-and-review.815439/

 

You rock brother!

 

difficult to determine clearly at this point. there are lots of types, thicknesses and heat conductance specs on those things. but all in all, they seem to be on the lower to mid-tier performance level of regular thermal pastes.

 

Main advantage in my mind would be that you don't need to worry about spread, and they can be cut into shapes if needed.

 

yep. so for fire and forget applications at stock clocks its a good option. although ud still have to worry about the electrical conductance. in the end probably the same outcome in terms of longevity and temps like a cheap(er) viscous thermal paste that doesnt dry out

 

Yeah, not going to touch them while they're new and a novelty. But I see a lot of potential in them too.

 

I think these are more industrial focused, as the testing data rises it sounds like a great solution for embedded devices.

 

Might get me some of this

 

I've been looking at the IC Graphite pads and it looks like they have 2 major benefits to them:

• Thermal and time stable. This would be great for OEMs since you wouldn't have to worry about the thermal performance changing over time and "re-pasting" is not needed.
• High Z-Axis heat transfer. This could potentially help with mobile CPUs which have a huge problem with generating all of their heat in small localised spots over the cores and lack an IHS (unlike GPUs that are very uniform and thus tend to run cooler despite actually generating more heat in most cases). The pad could theoretically act like a heat-spreader which may allow the cooler to function more effectively, at least until the entire system reaches heat-soak.

Can't seem to get it in Australia yet (and Amazon won't ship it which is odd).

 

haha a bit late to the party there but yes, GC extreme is among the top end pastes just a tiny tad less performant than kryonaut.

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

The results have been rather disappointing in laptop usage - http://forum.notebookreview.com/thr...mal-pad-available-for-test-and-review.815439/

There's been mention of a Panasonic 'spongier' type of pad ( http://forum.notebookreview.com/thr...-test-and-review.815439/page-14#post-10720210 ). However, I don't know if anyone's published results just yet.

 

I second this...

FTW - Thermal Grizzly Kryonaut!

 

I wonder if any of you will cut them up and use them as thermal pad replacements...

 

They work rather well as thermal pads. But need intense amount of care so that they dont slip.
One thing that should be noted though is that i used the 0.5mm and 1.5mm variants.

 

I see. But the idea of replacing TIM between heatsink and CPU or CPU die and IHS is pretty much out of the question?

 

Umm in laptops yeah. Desktops not so much.

Even in laptops it can work, but will need mods to increase the mounting pressure.

 

Soft PGS (spongier) pad works well with a desktop 2600k + Thermalright heatsink with decent mounting pressure. On par with decent thermal paste. No complaints, and it's a mount and forget system (won't ever pump out). Not anywhere near close to liquid metal.

ONE user claimed he tried the Soft PGS on his laptop and had LM type results--or so he claimed (Mr Fox didn't have the soft pgs version; he had the regular version which was atrocious), but NO ONE else has given any feedback at all with the Soft PGS on a laptop. No one. Not one person. And I'm not going to remove my LM job after all the work it takes to get 1-2C core temp differences in Prime95 (AVX disabled), you guys have no freaking idea how hard it is to bend over when you're 6 foot 6, have a metal bar in your spine and no suitable work area for your height, to clean LM and unscrew bolts and stuff. Otherwise I woudn't mind at all being the guinea pig.