Phase Change Material (PCM)

the thermal conductivity numbers are nothing special. it sounds like the old school "pink gum" FCM/TIM.
 
jlbenedict said:
Is the hype warranted regarding PTM7950?

https://thermalmanagement.honeywell...nical/PTM7900-ThermalManagement-Datasheet.pdf


What is the life of this type of thermal conductive material once it is put into application? (as from what I have gathered... with the temperature cycles.. it will change from solid to "less solid" and back and forth..at 45* C )
Click to expand...
Honeywell certified the material through 1,000 thermal cycles, with a pretty wild temperature sweep (really cold to really hot)----and only found the stuff to get better and better. Its designed to not "pump out" and be essentially permanent. As it was made for industrial use.

I recently put some on the CPU and GPU of my girlfriend's Asus gaming laptop. RTX 2060 and an 8 core 10th Gen intel. The CPU had liquid metal on it, stock from Asus. The liquid metal had a huge dry spot in the middle of the CPU and had also leaked way off to the side, only held back by the foam barricade they put in there.

The GPU had some kind of thermal grease. Which looked fine, albeit a bit pumped out.

I don't have the screenshots with me right now. But, IIRC:

The CPU's worst core had a 13 degree improvement. And the smallest improvement was I believe 9 degrees. And now none of the cores are more than 3 degrees apart from eachother. Whereas before, two fo the cores were like 8 degrees worse than all of the others.

The GPU I think had an 8 degree improvement---while boosting higher and using more power.

I ran a game for 10 minutes, and then shutdown the computer and let it rest----4 different times over a couple of days, before tracking the temps. This stuff needs a couple of thermal cycles to melt and fully seal, before it performs the way it should.

I have re-pasted laptops in the past, with stuff like Gelid Extreme. And it works for a bit, but usually pumps out after a month or two. And then the temps get worse.
 
Last edited:
Heatpipes work on the principle of phase change. The "working fluid" inside is what carries the heat away from the heat source, loses energy and "falls back" onto the heat source. So, yeah, I'd assume it might be good.
 
chameleoneel said:
Honeywell certified the material through 1,000 thermal cycles, with a pretty wild temperature sweep (really cold to really hot)----and only found the stuff to get better and better. Its designed to not "pump out" and be essentially permanent. As it was made for industrial use.

I recently put some on the CPU and GPU of my girlfriend's Asus gaming laptop. RTX 2060 and an 8 core 10th Gen intel. The CPU had liquid metal on it, stock from Asus. The liquid metal had a huge dry spot in the middle of the CPU and had also leaked way off to the side, only held back by the foam barricade they put in there.

The GPU had some kind of thermal grease. Which looked fine, albeit a bit pumped out.

I don't have the screenshots with me right now. But, IIRC:

The CPU's worst core had a 13 degree improvement. And the smallest improvement was I believe 9 degrees. And now none of the cores are more than 3 degrees apart from eachother. Whereas before, two fo the cores were like 8 degrees worse than all of the others.

The GPU I think had an 8 degree improvement---while boosting higher and using more power.

I ran a game for 10 minutes, and then shutdown the computer and let it rest----4 different times over a couple of days, before tracking the temps. This stuff needs a couple of thermal cycles to melt and fully seal, before it performs the way it should.

I have re-pasted laptops in the past, with stuff like Gelid Extreme. And it works for a bit, but usually pumps out after a month or two. And then the temps get worse.
Click to expand...

Damn... I posted the PDF that stated the 1,000 thermal cycle certification.. smacking me right in the face... RTFM lmao
May have to give this stuff a go, and gauge the performance in my own setting. Appears you can get this stuff in paste form, and solid form...
Thanks so much for your input! (y)
 
jlbenedict said:
Damn... I posted the PDF that stated the 1,000 thermal cycle certification.. smacking me right in the face... RTFM lmao
May have to give this stuff a go, and gauge the performance in my own setting. Appears you can get this stuff in paste form, and solid form...
Thanks so much for your input! (y)
Click to expand...
Yeah the 7958 only comes in paste form. I'm not sure how much better it actually is. The specs are very similar.

For me, personally: the pad form of 7950 is preferrable. They are delicate. But, It really is so easy to apply. and you don't waste any from spreading or applying too much, in general.

I do recommend some fine tipped tweezers, to peel off the plastic from the top of the pad, when you are ready to mount your cooler.
 
All the evga 3000 series gpus did use it.
Tough cookie if you try to remove the cooler from the die cold.

fyi
In all the documentation i have found they talk about two applications:

first -
slap it on an give it a few cycles

second -
this one mentions that you have to have a
certain amount of pressue - If i understood correctly this one is not doable/cheatable with home equipment
certain amount of heat
certain amount of time

all that to simply bond it correctly to the die and cooler surface the first time before it can work as efficient as advertised and keep its endurance promises.

No US or EU news outlet mentions this one and i find it odd.

Dunno why since it is mentioned in a lot of spec and docu sheets that google can find...
 
Last edited:
Carlyle2020 said:
Alle the evga 3000 series gpus did use it.
Tough cookie if you try to remove the cooler from the die cold.

fyi
It all the documentation i have found they talk about two applications:

first -
slap it on an give it a few cycles

second -
this one mentions that you have to have a
certain amount of pressue - If i understood correctly this one is not doable/cheatable with home equipment
certain amount of heat
certain amount of time

all that to simply bond it correctly to the die and cooler surface the first time before it can work as efficient as advertised and keep its endurance promises.

No US or EU news outlet mentions this one and i find it odd.

Dunno why since it is mentioned in a lot of spec and docu sheets that google can find...
Click to expand...
Ah, that looks to be the main benefit of 7958. Lower thermal impedence. So, less pressure required for optimum performance. Or performance similar to 7950. And that's going to be better for laptops, where mounting pressure is low. So that must be why Lenovo uses 7958.

**Also, The ideal pressure is a little north of 40psi. And I think that's a lower pressure figure for a typical heatsink on a CPU.

Direct to die, like with a GPU or Laptop, would of course be lower pressure. I wonder how much pressure a typical dedicated GPU heatsink applies to a GPU die?
 
Last edited:
I use it in my MSI GL65 Leopard on both the CPU and GPU along with K5-Pro for the rest of the chips that needs cooling, they both work great but I think it's time to blowout the HS/Fs with air.
 
Installed Kryonaut paste on my Framework and my Dell lattitude 7480. Worked okay.

I then switched to PTM7950. Got a 10 degree C drop on the max temps for the 7480 and a 14 degree C drop on the Framework. Little to no throttling. Used cut to measure areas on the CPU and iGPU.

I won't bother using pastes on my laptops and stuff going forward.
 
I just put PTM7950 on my Sapphire AMD 6800xt which replaced Gelid Extreme. 2c drop in temp and 12c drop in Hotspot temp. It hasn't even temp cycled yet and there is a benefit. Well worth it.
 
Last edited:
Yeah I would probably put this on my 6700XT next year. Mine runs fine as it is but I'm curious to see the result after the how well it worked on my laptops. I'm just wary of taking GPUs apart nowadays as there are so many screws, clips. pads and cables to deal with. Gone are the days of 4 screws and a gentle twist.
 
daglesj said:
Yeah I would probably put this on my 6700XT next year. Mine runs fine as it is but I'm curious to see the result after the how well it worked on my laptops. I'm just wary of taking GPUs apart nowadays as there are so many screws, clips. pads and cables to deal with. Gone are the days of 4 screws and a gentle twist.
Click to expand...
If you do then I would scrap the pads and use K5-Pro for them.
 
I'm running this on my 5950x and love it. I plan to use it on my GPU next time I do maintenance. I got it from ModDIY and also got some from amazon to compare. as far as I can tell the $40 piece performs the same as the $16 piece. I'm going to be putting this on everything from now on. Its just so easy to use and the results are amazing.
 
lobstar said:
I'm running this on my 5950x and love it. I plan to use it on my GPU next time I do maintenance. I got it from ModDIY and also got some from amazon to compare. as far as I can tell the $40 piece performs the same as the $16 piece. I'm going to be putting this on everything from now on. Its just so easy to use and the results are amazing.
Click to expand...
Got mine from what LTT got theirs, 7Buy or Buy7, whatever it was called. NVM, it was EBUY7. IDK if it was the real thing but it does the job, the same with K65-Pro instead of dealing with pad heights.
 
EBUY7 for the source... ptm7950 Dropped hotspot temps on my 7900 xtx by 15C... Well worth the money.
 
Will be applying to my 7900 as well. Tested 7950 an older cpu awhile back that was clocked to the moon and over volted the same way RIP 6300 vishera. It performed the same as really good pastes but took several heat cycles to really get bonded to the chip and cooler. Best thing about it is no pump out so one and done.
 
lobstar said:
I'm running this on my 5950x and love it. I plan to use it on my GPU next time I do maintenance. I got it from ModDIY and also got some from amazon to compare. as far as I can tell the $40 piece performs the same as the $16 piece. I'm going to be putting this on everything from now on. Its just so easy to use and the results are amazing.
Click to expand...

Link to Amazon supplier/seller?
 
chameleoneel said:
40mmx80mm
https://www.amazon.com/gp/product/B0BRJB8JNX/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

80mmx80mm
https://www.amazon.com/gp/product/B0BX42N9SZ/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Click to expand...

I saw that source and that is why I was asking for the specific seller. That seller, and a lot others that are most likely not actually selling the real the Honeywell PTM7950, have 0.20mm as the thickness. The PTM7950 does not come in 0.20 thickness...
https://thermalmanagement.honeywell...-ltm6300-pcm45f-ptm5000-series-data-sheet.pdf

See Note 2 under the Performance Chart...
"PTM7950 is available only in 0.25mm thickness
Thickness tolerance: ±0.075mm"
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
you must be careful there is alot of counterfeit ptm7950 out there.... I buy mine from EBUY7 always has come in factory honeywell packaging when I receive the order.
 
Last edited:
Enigma said:
you must be careful there is alot of counterfeit ptn7950 out there.... I buy mine from EBUY7 always has come in factory honeywell packaging when I receive the order.
Click to expand...
I bought mine there also, I think it was the 40mm, enough to do my i7-10750 and 2060RTX 4 times in my laptop, I think it came in a repackaged no-name sleeve.
 
bluestang said:
I saw that source and that is why I was asking for the specific seller. That seller, and a lot others that are most likely not actually selling the real the Honeywell PTM7950, have 0.20mm as the thickness. The PTM7950 does not come in 0.20 thickness...
https://thermalmanagement.honeywell...-ltm6300-pcm45f-ptm5000-series-data-sheet.pdf

See Note 2 under the Performance Chart...
"PTM7950 is available only in 0.25mm thickness
Thickness tolerance: ±0.075mm"
Click to expand...
could be that its not the real thing. or it could be that its the real thing, with mis-stated thickness spec (EBUY7 seems to mis-state the thickness, on supposedly legit listings).

I've used it on two laptops, 1 GPU, and on 1 CPU instead of a typical thermal paste. The handling characteristics, first use behavior, and initial performance results mirror youtube tests of legit PTM7950 purchased from MODDIY, etc.

Even if it isn't actual PTM7950, it seems to be a solid PTM. Maybe it won't last "forever". We'll see. But, so far, I think its at least great, for the price.
 
Last edited:
Whether real or not, I put mine in and it made a massive difference...so...

Quite easy situation, slap down the $15, what's that to anyone on this site? We are all H here arent we??? (as I've been told many times on this site) and give it a go. If you get a 5+ drop job done.
 
lobstar said:
Sorry for being so late to reply. PTM7950 40x80x0.2mm Phase Change... https://www.amazon.com/dp/B0BRJB8JNX
Click to expand...
This stuff is fake.
I put it on my 13900k and ran Cinebench R23 and got 100C in *seconds* on my AIO.
Imprint pressure was perfect. Kryonaut Extreme was 85C.

It was fine however on my radeon r9 290X (with a broken screwhole too).
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Falkentyne said:
This stuff is fake.
I put it on my 13900k and ran Cinebench R23 and got 100C in *seconds* on my AIO.
Imprint pressure was perfect. Kryonaut Extreme was 85C.

It was fine however on my radeon r9 290X (with a broken screwhole too).
Click to expand...
1. Is your 13900k delidded? If its not, you likely won't experience much improvement over normal thermal paste, in the short term. However, normal paste can still pump out in a conventional CPU application. So, long term, PTM can be beneficial.

2. As I already said:

I've used it on two laptops, 1 GPU, and on 1 CPU instead of a typical thermal paste. The handling characteristics, first use behavior, and initial performance results mirror youtube tests of legit PTM7950 purchased from MODDIY, etc.

Even if it isn't actual PTM7950, it seems to be a solid PTM. Maybe it won't last "forever". We'll see. But, so far, I think its at least great, for the price.

and to further elaborate, the Laptop CPUs improved around 13c. and the Laptop GPUs improved around 9c.
 
chameleoneel said:
1. Is your 13900k delidded? If its not, you likely won't experience much improvement over normal thermal paste, in the short term. However, normal paste can still pump out in a conventional CPU application. So, long term, PTM can be beneficial.

2. As I already said:

I've used it on two laptops, 1 GPU, and on 1 CPU instead of a typical thermal paste. The handling characteristics, first use behavior, and initial performance results mirror youtube tests of legit PTM7950 purchased from MODDIY, etc.

Even if it isn't actual PTM7950, it seems to be a solid PTM. Maybe it won't last "forever". We'll see. But, so far, I think its at least great, for the price.

and to further elaborate, the Laptop CPUs improved around 13c. and the Laptop GPUs improved around 9c.
Click to expand...

No it's not delidded, and it was 15C worse than Kryonaut Extreme.
If that isn't complete counterfeit, well, I just do not know what to tell you.
I already have Thermalright Helios (which is also PTM). I know what PTM is supposed to perform like.
 
Just to be clear when using PTM7950 it does take multiple thermal cycles for the material to "settle" into place... You will actually see a temperature drop over time.... For Example my 7900XTX started with a hot spot of 90C and over time it has settled to a hot spot average of 80C, all in all I think this took about 100 Thermal cycles to acheive.
Another interesting thing to know about PTM7950 is that the temperature actually drops after the PCM changes state during use. See Below graph :
1707663013948.png


This product really does not have a benefit on the CPU over a modern thermal paste. The main reason that it is used on a GPU is because of its unique ability to resist pump out. When using Kryonaut Extreme on my 7900XTX the grease was pumping out after 2 weeks of use and the temps were starting to climb, by the third week the card was thermal throttling.

Long story short stick with a traditional thermal paste/liquid metal on your CPU, and use PTM7950 on the GPU, in conjunction with a good thermal pad/putty.

On another note I think that Thermalright Helios is just a rebranded PTM7950. As I have stated before I would get my PTM7950 from EBUY7, or MODDIY. I have never tried the Aliexpress or Amazon versions. I however would not be scared to try the Thermalright Product as the reviews put it in line with Honeywell PTM7950.

**Edit** Source for Graph:
View: https://www.youtube.com/watch?v=2BhKx0iQ4K8
 
Last edited:
I have rerun some of the testing myself and I can confirm the same experience on my own machine when running PTM7950 on my 7900XTX:

This one shows a run of Timespy Extreme notice the temperature drop on the second test as the hotspot temps climb, They Gradually start to fall as the PCM changes state:

1707665025417.png


Here is a second one that shows the gpu clocks increase as the temp changes with the PCM change:

1707665121299.png


This is just a short snapshot I will do a longer run with a stress test to confirm the same results

Here are the long run results:
Notice that the temperature for the first 13.5 minutes of the test show that the temperature is steadily increasing to the point that there are hot spot spikes to about 82C, Then after that point the temperature spikes decrease to the point that they are only hitting the 81C mark after that point.
The difference in the clock speed spikes is also further minimized at the same point. The main increase in temps up to the 10 minute mark is because of the water loop temperature equalizing in my system. At the 13.5 mark on the graph the water in the loop has reached equilibrium and is now staying consistent.
I do not believe the change in temperature to be drastic upon the PCM changing states but the change is there none the less. The biggest benefit to the PCM is that there is no pump out. I have been running it consistent on this GPU for a year now and there has been no change in temps or "pump out effect".
1707667331044.png
 
Last edited:
Falkentyne said:
This stuff is fake.
I put it on my 13900k and ran Cinebench R23 and got 100C in *seconds* on my AIO.
Imprint pressure was perfect. Kryonaut Extreme was 85C.

It was fine however on my radeon r9 290X (with a broken screwhole too).
Click to expand...
Huh, haven't had that experience on my Steam Deck and 5950x. Sorry about that! I never intend to recommend bad stuff.
 
You must log in or register to reply here.
Back
Top