Overclock.net banner
1 - 20 of 32 Posts

·
Premium Member
Joined
·
6,359 Posts
Discussion Starter · #1 ·
I have a new EVGA 3080ti coming in this monday and I'm def going to do a re paste/re pad on it once I burn it in/test it. I understand that the EVGA 3080ti's are using a sort of thermal putty on certain areas of the PCB as opposed to the typical thermal pads (which they still use on the main memory chips). I found this T-Global TG-PP-10 thermal putty online ... got two 30 gram bottles for roughly $50. It's rated at a very decent 10 W/mk thermal conductivity, which is nearly as good as the Thermalright 12 W/mk pads and it seems super easy to work with as it's literally like a "playdo" consistency (see video below), not running or melting .. super easy to shape and mold and compresses very easily. According to some banter online, it's better than pads in applications with varying heights among components that need to be covered.

Here's what I got...

2518541


My plan is to replace the EVGA stock putty with this TG-PP10-30G putty as I'm sure it's thermal conductivity is much better than the EVGA stock putty. I'll use either Fujipoly or Thermalright pads, most likely the latter, for the main memory chips where EVGA has their stock thermal pads.

This dude claims it dropped his chip temps by 20C on his RTX 30 series cards and his experience, real or imagined, is pretty much what sold me on this stuff. Here's the vid:


I'm also considering re "padding" my Alienware Area51-m R2 laptop with this putty as well but not sure how that would work. The Area51-m has a whole bunch of pads on it, ranging from 1mm to 1.5mm to 3mm and what I've read is good about this T-Global putty is that you just sort of shape it a bit thicker than the pad would have to be and the cooling assembly sort of presses it to just the right thickness with the excess squeezing out the sides but just hanging there and not running all over the place like a typical compound would.

So do any of you have experience with this stuff? Any comments?

Thanks in advance,
~s1rrah
 

·
Registered
Joined
·
1,241 Posts
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.
 

·
Premium Member
Joined
·
27,803 Posts
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.
 

·
Iconoclast
Joined
·
31,435 Posts
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.

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.
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.
 
  • Rep+
Reactions: J7SC

·
Robotic Chemist
Joined
·
4,084 Posts
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.
 

·
Premium Member
Joined
·
6,359 Posts
Discussion Starter · #6 ·
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:

2518699
 

·
Iconoclast
Joined
·
31,435 Posts
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.
 
  • Rep+
Reactions: s1rrah and Asmodian

·
Facepalm
Joined
·
10,200 Posts
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.
Where did you even find 500g in stock?
It's been out of stock for (skeleton gif) forever on Digikey.
The 50g one is out of stock too and it doesn't seem 50g will go very far.

$117....yikes.... $117 in Gelid Extreme / Ultimate pads can do a LOT of video cards...
 

·
Registered
Joined
·
1,683 Posts
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.
 
  • Rep+
Reactions: s1rrah

·
Robotic Chemist
Joined
·
4,084 Posts
$117....yikes.... $117 in Gelid Extreme / Ultimate pads can do a LOT of video cards...
Maybe two 3090s?

I did get it from Digikey, during May this year.
 

·
Premium Member
Joined
·
6,359 Posts
Discussion Starter · #11 ·
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 ...

2518946
 

·
Registered
Joined
·
1,683 Posts
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 ...
That's good, I wouldn't bother changing them out. You will get little to no improvement and maybe even worse.
 

·
Iconoclast
Joined
·
31,435 Posts
$117....yikes.... $117 in Gelid Extreme / Ultimate pads can do a LOT of video cards...
So will 500g of this stuff.

I paid that much for 150g cause I had to buy 50g at a time.
 

·
Registered
Joined
·
1,683 Posts
So will 500g of this stuff.

I paid that much for 150g cause I had to buy 50g at a time.
I'm certain you can split it up and sell it over at reddit hardwareswap relatively fast if you're interested.
 

·
Registered
Joined
·
283 Posts
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...
2519165
 

·
Registered
Joined
·
1,683 Posts
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.
 
  • Rep+
Reactions: satinghostrider

·
Registered
Joined
·
771 Posts
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.
 

·
Registered
Joined
·
1,254 Posts
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.
 

·
Registered
Joined
·
283 Posts
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 mean if you're mining, it is worth it.

If you're chasing benchmark scores, lowering temps is always a good thing.

If all you do is game and not OCD, then I'd say leave it alone.

I also would say if you're stock cooler and temps are bad, it's worth. Mostly I'd do this only if moving from air to water as it makes sense.
 
1 - 20 of 32 Posts
Top