Hard to believe it won't migrate over heat/time. Not impressed with the video as it didn't really show anything. Best of luck.

 

I've been using TG-PP-10 for a while, mostly on video cards. Not entirely convinced the 10W/mK rating is accurate (though it's probably not less accurate than anyone else's thermal conductivity ratings), but it easily matches or beats the thermal performance of quality mid-range (7-8W/mK) pads while being much more conformable than even crappy (~1M/mK) pads. It also won't crack, bleed silicone oil, or migrate, and is largely reusable, as long as it's clean.

It's kinda messy, being tackier than Play-doh or light modeling clays and of similar stiffness. It sticks to nitrile gloves and getting the last bits out out out of the smaller tubs (50g or less) can be a bit of a chore, but it's still generally more convenient than cutting thermal pads to size, if you have more than a few different gap thicknesses to fill. Some people freeze it, or work with it with bear hands, but I'm a stickler for not contaminating TIMs I'm about to use with either condensation or skin cells/oils.

It's also dense, and you don't get a lot of volume for your money unless you buy it in bulk. It took sixty dollars of the stuff to replace and supplement all the TIM on my 6800 XT Red Dragon (except the core, obviously), mostly because I was using it on the backplate as well. Admittedly, trying to put quality pads everywhere I used the putty would probably have cost even more.

Ultimately, any pads that are more conformable perform significantly worse and any pads that perform better are stiff and hard to work with. I wouldn't replace good pads with it, but on the balance, if the choice is between buying a stack of different thicknesses of quality pads and cutting them to size, or buying high-end putty (like TG-PP-10), the putty will usually get my vote.



It's thermal putty, and one with a pretty high filler ratio at that. It hasn't migrated in the slightest six months and thousands of thermal cycles since I've had it on my 6800 XT. There are also other, softer, putties that are part of the stock cooling on some of my components that haven't moved in more than five years of use.

 

Interesting idea. Am interested in seeing your before and after test results. I'm sure you know this, but please monitor and record air temp entering cooler when you record GPU temp. With all the heated exhaust air moving around GPU it's quite hard to keep air temp entering GPU at a fixed temp. Recording air temp entering cooler at same time as GPU temp is recorded and using that delta temp makes results much more accurated.

 

I bought a 500g tub of the stuff, it worked well under most of my 3090 backplate and I have most of it left. It is pretty cheap per use once you get that much at once. 500g could cover a lot of GPUs. It was way faster to apply than thermal pads and I see similar temperatures, if not slightly better.

It wasn't a huge change compared to thermal pads though, it is easier and more reusable then premium thermal pads, not significantly better. You can also put it all over everything because it squishes out if it needs to, as longs as you don't go too crazy.

 

Nice. I know the various pad heights necessary for re padding my EVGA 3080 ti card ... so just going to estimate the putty height as a bit over the regular pad height and let it compress in to place. I also have an Alienware Area51m R2 laptop that I'm going to re pad but it takes a lot more pads and pad heights than a GPU. Not sure if the putty will work or not but will give it a shot. Main hurdle is figuring how much putty to put on each point/location. Here's a diagram of what the Area51m R2 requires in regards to pads:

 

I just roll the putty into small balls for individual chips and use a string/snake for rows of very small ICs (like powerstages or mosfets). Takes a bit of practice to estimate the ideal diameter, but it's generally faster, easier, and more economical than trying to mold it to flatter shapes.

You don't need to be super precise as it is highly conformable, but it may be a good idea to practice with a smaller/less critical component before trying to do an entire laptop.

 

Putty should work as well as pads. Pads HAVE to be compressed to get their full conductivity rating. Putty only needs 10% compression.

The likely hood of compressing a thermal pad to what it needs to get its maximum thermal conductivity rating is rare since gaps are random. But at least putty you can form into place.

That being said, you maybe see literally no improvement over the putty that's already there if it is well placed. If your junction temperatures are around 80-90C, I wouldn't even bother.

 

If your junction temperatures are around 80-90C, I wouldn't even bother.

Tjunc temps usually don't go beyond 82C .. this is during the 10th run of the Metro Exodus Enhanced Benchmark ... 1440p ... "Extreme" settings DLSS off. It's odd as I occasionally see core speeds boost upwards of 2080mhz but this benchmark hardly ever employs my max GPU overclock ...

 

Observe the shore scale of the thermal pads.
Highest isn't always the best. The Gelid Extreme is what you should be looking at. I had the Thermalright ones and my hotspots increased over time. I since switched to Gelid Extreme's and it made a tremendous difference by almost shaving 8 degrees off hotspot temps and 10 degrees of GPU temps. Memory temps were down 6 degrees as well. Fact is the Gelid Extreme pads were just more compressible and had better GPU Die to block contact.

And to add, I had to use the average EK-TIM when repasting my card with Gelid Extreme pads cause the Thermalright TFX ones were a ***** to apply and I only had EK-TIM ones left.

Take note that if you're on 3090, the Gelid Extreme's will melt on the back as some have reported but they don't affect performance just hard to clean up. So I stuck back to Thermalright Odyssey pads just for the back but front was all Gelid Extreme's. If you have a 3080 or 3080 Ti, then this doesn't concern you as the back has no memory nands.

Here's my results. Hope this helps...

 

Observe the shore scale of the thermal pads.
Highest isn't always the best. The Gelid Extreme is what you should be looking at. I had the Thermalright ones and my hotspots increased over time. I since switched to Gelid Extreme's and it made a tremendous difference by almost shaving 8 degrees off hotspot temps and 10 degrees of GPU temps. Memory temps were down 6 degrees as well. Fact is the Gelid Extreme pads were just more compressible and had better GPU Die to block contact.

And to add, I had to use the average EK-TIM when repasting my card with Gelid Extreme pads cause the Thermalright TFX ones were a * to apply and I only had EK-TIM ones left.

Take note that if you're on 3090, the Gelid Extreme's will melt on the back as some have reported but they don't affect performance just hard to clean up. So I stuck back to Thermalright Odyssey pads just for the back but front was all Gelid Extreme's. If you have a 3080 or 3080 Ti, then this doesn't concern you as the back has no memory nands.

Here's my results. Hope this helps...

His memory junction temperatures are at 80C. Why fix something if it isn't broken nor will it add any performance to his memory?

I say, leave it all alone.

 

I love thermal putty. I don't even want to know how much money Ive wasted trying to find the perfect thermal pad thickness without compromising GPU contact or limiting memory contact. My first attempt with the 3090 with "recommended" thickness resulted in no GPU contact at all (17 w/mk pads), second attempt resulted in some memory not getting full contact and mem temps reaching over 100C. The putty removes all guess work but I would agree that in a perfect scenario, there are better pads (I never tried the Gelid Extremes)

Some lessons learned:
-Nitrile gloves get sticky but are recommended
-ifixit has a prying tool that has a hook on one end and a spatula on the other. The spatula is perfect for getting all the putty out of the bottle and cleaning up excess putty as necessary.
-Freezing/cooling the putty before hand helps a lot but being a thermal conductor, it warms up fast so its kind of a useless step unless you have a desk refrigerator
-Its hard to get the putty to stick to most surfaces; when attempting to apply, it generally it just sticks to your glove. I found starting with a small ball of putty, I can rub that into the surface in order to "prep" it so when I apply a larger glob, it sticks better. I then slap it with my finger to flatten it and shape it (this also helps clean off the putty from your gloves)
-I always apply to the heatsync and I mound it to be thicker in the middle to allow a cleaner conforming to the memory chips without getting air pockets
-There's no such thing as too much but its a bit $$$.

Watercooling an RTX 3090 with a simple alu backplate, temps were in the high 70's and low 80's when Eth mining. I mounted a RAM waterblock onto the backplate and now mining temps are in the mid 60's.

 

Is this stuff like super sticky like K5 Pro?

Because it looks like K5 Pro, which is very comparable performance wise. And K5 Pro is a mess lol. It will ruin a PCB of any electronic board.

You could have a tub of alcohol and a million brushes. it won’t come off!!

You would need an Ultrasonic cleaner to dump the GPU board in and let it run for a few hours to clean this thing.


Just a head up for anyone using this stuff.

 

Is this stuff like super sticky like K5 Pro?

Nope

 

I see in the video that he is using both thermal pads and the putty, would you advise that ? or should use only one of them ?
what thermal pads size should I get for my EVGA 3090 FTW3 ?

 

I would recommend using only the putty. You could also use high quality pads on the memory and putty everywhere else, but I wouldn't put pads on top of putty or putty on top of pads.

A thinner thermal paste with pads might make sense, but this putty and pads at the same time does not.

 

any idea why he use both on the video ?

 

I tried to source TG-PP10-50G THERMAL PUTTY but it's nowhere to be seen on stock and I think only Digikey is selling it and they are out of stock. Does anyone know where I can buy it and ship it to Europe?

 

Any thoughts on this putty for use a CPU? I'm wondering due to the viscosity if it will be less prone to pump out and need replacing less often.

 

It might not need replacing as often, but it would perform poorly. A proper high grade thermal paste is much better (thinner) when filling large gaps is not a requirement.

 

For GPU you want a thick thermal paste, from

Thick thermal paste for GPU?

Hi, I need to repaste my gpu (5700XT) because of pump out. Noctua NT-H1 lasts a week and half before needing to repaste, MX4 lasts about a month or so (let's say month and a half) Is there any good high viscosity (like OEM) thermal paste? Where I live there is no ICD so that's not an option. Are ...

linustechtips.com

"Thermalright TFX is the best choice for GPU's but it's very hard to spread.
It's suggested that you put the syringe in very hot water (although not boiling) and leave it in
about 15 minutes before application. This makes it easier to spread. Another trick is to put it
in an X pattern on the chip and then use a hairdryer to heat the paste and then spread it fully.
But not everyone has a hairdryer so dunking the syringe in very hot water is an option.
But it's literally almost the most expensive to buy on the market, only being topped out by
Kryonaut Extreme, which wins the gibberling award at a horrendous $23 for two grams.
(cheaper if you import the 6.2g from Aliexpress).

SYY-157 is good also, and unless you got a very dry batch, it's much easier to spread than TFX
and it's a lot cheaper to buy. Almost on MX-5 level of cheapness if you buy the 8 gram one.

A third option is FuzeIce Plus. "

 

Gelid GC Extreme is also very thick and has high longevity. The new Thermal Grizzly Kryonaut Extreme is thick as well (and should have the fast pump-out slowed (the original Kryonaut especially gpus had to be reapplied too often.

 

Like i asked in many topics but not a good answer:
The putty has a shelf life,i mean only 6 months and the u need to replace it?

 

Not an easy question to answer.
"Shelf life" varies dramatically with how it's stored.

How well container seals / isolates material from outside influences​

Plastic containers allow way more osmosis than metal or glass containers.​

Temperature of storage area​

Freezer storage usually dramatically increases shelf life.​

Light in storage area​

Dark storage areas usually work much better than bright storage areas.​

Then there is the fact some things react differently than others to temperature and light.

My guess is this thermal putty will keep best in colder darker places than in warm light places .. but that is just a guess.

 

thanks for the reply, my main question is about the time it last when you used it.
I mean once i used it, when i need to change it?
Sorry my english

 

Your English is fine.

Hard to say. Depends on how hot the component it's cooling runs. The more heat it has to transfer from chip to cooler the faster it will dry out and start loosing it's cooling ability.

I would guess at somewhere between 6 months and several years.

Sorry I can't be more precise. It's kinda like asking someone how long tires will last. Depends on how many miles person drives per year, if they drive like a little old lady or like a road racer.

 

ok, very thanks man!

 

I just replaced the cooling on my 6800 XT and removed my TG-PP-10 from it in the process. Was on there for ~10 months of heavy use (gaming and benching with up to a 450w power limit, on air, and mining when it wasn't being used for something else). The putty was like new, and if it wasn't for the dog hair now embedded in it, I'd reuse it.

From my experiences with it so far, I'd expect it to last several years, and outlast most other TIMs.

 

thanks again!

 

Hi again, why my gpu hotspot delta to gpu core inscreare in like 8-9 days?
It's startted like 11°c, and now it's 13,5 and something.
I used putty like i said and i mounted the card and not touched... i start to think this degradation depends on the mountings sistem of the FE.
Any raccomendation? I used TFX thermalright paste, but i have mx-5 and many others too, what you raccomand?

 

Try applying several layers of 3M high temp polyimide (kapton) tape directly on the PCB where the X-bracket leaf spring "black" section contacts the PCB (not the screw holes, but the central square area). Apply it on the PCB, not on the x-bracket. Do a few (or maybe more than a few) layers in a full square right where the X bracket's inner area contacts the PCB. I don't know how many strips you would need and I just thought of this idea a few days ago and am too busy to even bother, but I suspect this will help your problem. By increasing the amount of material that the X bracket's middle ring section has to compress down, this should in theory help increase contact pressure. This is the same idea as what the washer mod does on some Radeon video cards to increase the pressure their X bracket applies, by compressing the springs down more.

I would guess that about 0.2mm of 'extra' material thickness would do the job. The only question is how many layers of stacked Kapton tape would you need to create enough thickness resistance for this.