Cool GTX

donta1979
Those wondering and wanting to see how good Honeywell PTM7950 pads are if you do not believe those of us who are using it, maybe you will believe Igor.
Not over-hyped: Honeywell PTM7950 in a lab test and a real game changer for graphics cards | igor´sLAB (igorslab.de)

nice find, great details
 
(excerpt)
 
"Where I see problems are with water coolers and water blocks, where there might not be a complete phase change, at least on the cooler side. But you need at least 55 to 60 degrees throughout the pad for a complete change of state. “Half-defrosted” pads are counterproductive and are even more prone to cracking. At least that’s what one of the OEMs has now confirmed to me. Everyone has to work out for themselves how to solve this properly. I wouldn’t use a pad on a water block, at least not a PCM,"

well I was chatting with Igor about it on FB once he posted it, and I do warn people you really do not want to do this with a block or super crazy low temps like Igor mentioned, stock cooler, or a loop that runs really warm or an AIO cooled card? yup.

donta1979

Cool GTX

donta1979
Those wondering and wanting to see how good Honeywell PTM7950 pads are if you do not believe those of us who are using it, maybe you will believe Igor.
Not over-hyped: Honeywell PTM7950 in a lab test and a real game changer for graphics cards | igor´sLAB (igorslab.de)

nice find, great details
 
(excerpt)
 
"Where I see problems are with water coolers and water blocks, where there might not be a complete phase change, at least on the cooler side. But you need at least 55 to 60 degrees throughout the pad for a complete change of state. “Half-defrosted” pads are counterproductive and are even more prone to cracking. At least that’s what one of the OEMs has now confirmed to me. Everyone has to work out for themselves how to solve this properly. I wouldn’t use a pad on a water block, at least not a PCM,"

well I was chatting with Igor about it on FB once he posted it, and I do warn people you really do not want to do this with a block or super crazy low temps like Igor mentioned, stock cooler, or a loop that runs really warm or an AIO cooled card? yup.

I've ran an EK block on the old 4090 /W EK suck-pads and KPx,  since Dec '22 and I don't have the Temp needed for PCM , even at max OC 

I was replying to a 30 series thread, and somehow got here. Weird..

Ignore -. Original message deleted.

kraade
 
 
I've ran an EK block on the old 4090 /W EK suck-pads and KPx,  since Dec '22 and I don't have the Temp needed for PCM , even at max OC 

Yeah that would be a bit painful to try and cure you can do it yet you will have to reduce fan and pump speeds to make it really heat up enough to cure properly.
 

donta1979

Cool GTX
Where is everyone ordering PTM7950 from?
 
Seems to be some quality issues or knockoff/counterfeit product out there, that some people have reported on other forum's posts
 
The results look very good for the PTM7950, I just want to get it from a reliable vendor

MODDIY if you want an authentic one that performs better than some of the pre packaged knockoffs of Amazon. There are legit ones on Amazon but I wanted to ensure I got the real thing.

Got My ---> Honeywell PTM7950 on hand, got it from MODDIY - (2 weeks for delivery & most of that time was inside of the USA ) No one else seemed to have genuine stock or it was expired, based on Honeywell stamped date
 
My Gigabyte 4090 Gaming OC has been going over 106C Hotspot @ full power.  Ive had to tune it down to 85% Voltage & limit OC to 110% to stop thermal throttling issues.
 
Now I just need to pull the GPU & do the surgery
 

Edit: The image links posted and previewed fine but no don't work so I am going to link the ocnet thread mentioned later in this post: https://www.overclock.net/posts/29233230/
 
Been a while since I have been around but this thread caught my eye.  I used liquid PTM 7950 on both of my 4090s (Gigabyte Gaming OC and Asus Tuf OC) with their respective waterblocks and used some thermal putty for good measure (the 2080 TI FTW3 and Fujipoly 11w pads left a bad taste in my mouth so far as fitment and GPU die pressure).  I did a little write up on it on overclock dot net and i am going to do a copy and paste job from over there.  I would point you toward the much cheaper Upsiren UTP-8 thermal putty for thermal putty from the Middle999 Store on ALiExpress.  I sourced the PTM 7950 paste from Moddiy.  A year after, my temps and deltas are the same and I like the EK waterblocks a little less since the 2080 TI Hydro COpper left me with such great expectations on GPU die temp to coolant delat while under load.  Originally posted Sept. 9 2023:

I just wanted to throw in on this conversation. I was not happy with my temperatures on my Gigabyte Gaming OC 4090 after installing the EKWB waterblock to it in February 2023. Particularly the GPU Hotspot number bothered me as seen in here
https://www.overclock.net...kryonautx-jpg.2626967/" />
 
I had no idea I would be posting anything about this so apologies about the variability in details documented going forward.

I used the EKWB supplied pads and Thermal Grizzly Kryonaut Extreme. In that preceding screen shot (taken from a video I recorded of the testing/capture) I chickened out and stopped the testing 15 minutes in. The GPU temp w/ a 41c delta vs the coolant is bad enough but that 19c GPU hotspot delta for 98c is what forced the issue of repasting and trying some Upsiren UX Pro thermal "putty" (I do think clay is just as accurate a descriptor). The test conducted above was with 133% PL. The following picture is of the waterblock separated from the GPU immediately after the temp test above.
 
https://www.overclock.net...ompressed-jpg.2626974/" />
 
Note the weird coloration of the TG at the 12 o'clock position on the GPU die and the corresponding 6 o'clock poisition on the waterblock... I've been at this since the 90ies so trust me when I say this was surgical bay grade clean when applied and assembled.

TL;DR paragraph: I read great things about Honeywell PTM 7950 and had some on hand when I waterblocked my 4090 but I had tried and true TG KX also and opted to use the familiar option to start off with. Always interested in thermal putties since encountering it on my launch 2080 TI FTW3 Ultra, eventually and recently I came across posting on Upsiren UX Pro and U6 pro putties. I liked what I saw from Snark's Domain testing but had some serious doubts (still do) given the small size of the channel and not really finding anything reputable about Upsiren otherwise I decided to test it out on my recently retired 2080 Ti. That card always had a sky high bank 3 memory temp on air using the stock Laird pads or anything short of massaged-to-perfect-thickness FujiPoly 11w pads. Since it was a card I had assembled and disassembled about 2 dozen times in pursuit of best thermals possible I ordered up 150g of UX Pro and decided to test it and PTM 7950 at the same time w/ the factory air cooler. Long story short the UX Pro outperformed my steno pad notes of best bank 3 temps by about 5c while giving top GPU & hotspot numbers. I was pretty impressed with PTM 7950 so I decided to use it on my 4090 G-OC. I ordered up some more UX Pro End TL;dr paragraph

The following image is of my post PTM 7950 (paste, not pad) and Upsiren UX Pro putty job at the end of a one hour and fifteen minute 133% PL, 570W test where you will note that vs the 15 minute test I posted up my GPU Core temp improved 8c, hotspot improved 14c, and memory junction improved 6c.
 
https://www.overclock.net...133pl_end-png.2626979/" />
 
Since the OCCT test used barely hits the memory controller I look at the memory junction number as informal indication of how effective thermal conductivity is from heatsoak of the PCB from the GPU die. In Furmark testing hitting the memory controller (30% or so) as hard as my config can while keeping the GPU @ 500w (the most I can get it to even w/ 133% PL) the memory actually runs cooler. If anyone is interested I would be happy to capture that data or test upon request, but the memory actually runs right around the same number of 42c with my coolant generally topping out at 36c in from heat sources, 34c out from radiator pair.

Lastly I have some screen shots showing the beginning and end of 100% PL OCCT GPU testing with a 1 hour 15 minute duration. I captured a bit more information this time around as I generated it for this post.

Start of 100% PL OCCT run: Coolant in/out 30c/29c, GPU 54c, hotspot 67c, memory junction 32c
 
https://www.overclock.net...0pl_begin-png.2626980/" />
 

End of 100% PL OCCT run: Coolant in/out 34c/32c, GPU 58c, Hotspot 70c, memory junction 38c
 
https://www.overclock.net...100pl_end-png.2626981/" />
 
ten seconds post run info
 
https://www.overclock.net...10secpost-png.2626982/" />
 
So I am of the opinion that the hype on PTM 7950 is legit. I am estatic about the core temp and pretty confident this stuff will never pump out no matter how silly I get with pushing it or long I run it. I also think that Upsiren UX Pro putty is pretty damned good considering the results vs effort required of harder pads. I think a 12c delta of GPU core to hotspot at 100% PL & 445w TBP average over an hour plus in OCCT GPU is good. I also think that the 16c delta of the GPU core to hotspot @ 571w average TBP over an hour is quite solid. I attribute the overall reduction in delta vs coolant (from 41c to 26c) in the 133% PL tests is owed in part to the increased pressure of the waterblock to the GPU die that the Upsiren thermal putty allowed.

Now this is an important note. This stuff is like clay, not putty to be sure as Cobra26 validly pointed out. Kind of reminds me of kinetic sand with a fair bit of silicone oil in it. When doing the initial assembly of my EKWB waterblock to the Gigabyte Gaming OC 4090 there was enough "lift off" from the putty to prevent the GPU die screws from engaging with the block threads through the PCB. I pressed and very gently "wiggled" the card focusing on making the majority of the pressure come from the permeter of the GPU die. The thought process here was to get the squeeze out of putty to push out away from the die lest it contaminate the PTM 7950. Yes. I did use that much putty. No, my job wasn't anywhere near as clean, precise, or neat as Cobra26's (truly, nice work there). Years from now (hopefully) when I go to disassemble its going to be one ugly mutha in there but Cobra26's concerns with compressibility is particularly valid for the uninitiated.

After I got the GPU mount screws in and initially torqued the PCB was obviously bowed from the application of UX Pro. So I worked my way from the AD 102 die as a centerpoint and pressed the PCB and the waterblock towards each other until the PCB just barely floated above the mounting posts common to the waterblock. I final torqued the 4 GPU die mount screws and installed the backplate. After doing so I gave it one last press and the results are what I have laid out here in that post.

Upsiren UX Pro requires some careful motions while installing but given that vs FujiPoly 11w or 17w, or Thermal grizzly Minus 8 pads, I'll take the UX Pro putty everytime. I have concerns about it and advisories:

   *I am curious: if this stuff is going to prove corrosive to PCBs after time.
   *Not sure how it is going to perform after the silicone dries out of it.
   *Testing on the 2080 Ti I performed, unlike thermal pads that more or less stay in contact and spring back to a certain point if you overcompress (by twisting a card or sag or even lopsided assembly), Ux Pro does not pop back to make contact with the items it was placed in between. If you make an assembly error or otherwise manage to queeze it out too much you'll now have an air gap. I flex tested this on the 2080 Ti to the point I was not sure the card would still be functional. I grabbed either end of the card after initial thermal testing and twisted it like I was wringing a wet towel and reinstalled and tested. GIven the reinforcements on the that particular model I am not surprised the only points that suffered were the VRM temps at reading points 2 and 3 but I was able to introduce an air gap between VRM components and the heatspreader plate. Shining a light between the card and the dissipation plate i could see the gap I introduced by squeezing out putty through twisting the PCB, Since it is putty I just disassembled, reformed the putty, slapped the card back together, and all was well again but this characteristic of this putty can be very problematic. Caution is advised in this regard.
    *Stupidly expensive in UX Pro form. I have U6 Pro on order and in the next few weeks plan on testing it out to see if Snark's Domain testing bears out that it is not worth the difference in price while still being well worth using over traditional thermal pads


To sum it all up I am very happy with the result of going from Thermal Grizzly Kryonaut Extreme and EK supplied thermal pads to Honeywell PTM 7950 & Upsiren UX Pro Thermal putty on my Gigabye 4090 Gaming OC and EKWB waterblock combo. I definitely recommend the PTM 7950 paste from the ModDIY and am happy with the Upsiren UX Pro putty from the middle999 store on AliExpress but only recommend that with the cautions/advisories i listed above. If you're cool with rolling the dice like I know I am, it's definitely nice upon initial installation.

✌
 
 
edit by Cool GTX changed all that giant font to 12pt - to improve readability
 
 

removed the oversized Font
 
@ rchiwawa
 
Did you try preheating the Upsiren UX Pro putty, GPU card & water block - with an electric heating pad to see if the product was slightly easier to use
 
I've never tried Upsiren UX Pro putty, so I've no idea how temp sensitive it is - responds with better pliability as the Temp Delt above ambient increases
 
I always preheat my parts, as I've found it has helpful advantages with viscosity issues
 
100% agree, that if the Pads are too thick or too hard the impact to GPU Temps is unmanageable
 
So, for clarity which putty did you test; because, my web search calls several versions "Ultra" Grey or White putty ?
 
UPSIREN UX PRO Ultra 16.8w/mk Thermal Putty (newegg = grey & Amz = white)
 
there is also 12.8w/mk Thermal Putty = pink

Cool GTX
removed the oversized Font
 
@ rchiwawa
 
Did you try preheating the Upsiren UX Pro putty, GPU card & water block - with an electric heating pad to see if the product was slightly easier to use
 
I've never tried Upsiren UX Pro putty, so I've no idea how temp sensitive it is - responds with better pliability as the Temp Delt above ambient increases
 
I always preheat my parts, as I've found it has helpful advantages with viscosity issues
 
100% agree, that if the Pads are too thick or too hard the impact to GPU Temps is unmanageable
 
So, for clarity which putty did you test; because, my web search calls several versions "Ultra" Grey or White putty ?
 
UPSIREN UX PRO Ultra 16.8w/mk Thermal Putty (newegg = grey & Amz = white) there is also 12.8w/mk Thermal Putty = pink

The pad will melt and it will not be too thick once its cures and you slightly tighten your block whatever it is down a tiny bit more, or if you use a direct die method with the block with a limiter. Even the gpu temps will be phenomenal if you want to use water cooling you have to let it get warm enough and while its in the curing process tighten your block just a tiny tiny bit. I am using the pad on my 4090 but with its stock vapor chamber and on my direct die aio cooled 14900k. If you look a the information igor posted once its cured and setup the ptm7950 paste does not compare nor comes close to the pad.

I guess you could try turning the pump off and start that fuzzy Vulcan donut until you hit target temp then bring up pump speed to maintain it .
I am curios once you have it cured and your normal temp under water is 43-45 , would it help?