The Official Thermal Interface Material thread

[Borc]

I finally found a review for the 17 W/mK rated LK-17, here the link:

Slightly below Kryonaut Extreme, not bad. Even more promising to me seems AMeCh SGT-4 which is on par with Kryonaut Extreme in all tests on youtube. This is from BullsLab:

Differences are not so big on a desktop CPU though. I'm curious how it performs on my laptop compared to Kooling Monster KOLD-01 or AeroCool Fuzion.

 

[Deleted member 234997]

PK3 by a landslide on a GPU...

 

[jayjr1105]

PK3 PTM7950 by a landslide on a GPU...

ftfy

 

[Deleted member 234997]

ftfy

Check clock speed and watts hit.

 

[freeagent]

141w PPT from a 5800X3D

 

[:D:D]

What is this? And why?

Sorry, I don't understand the question.

 

[lexluthermiester]

Sorry, I don't understand the question.

You posted a photo of a CPU IHS with a deep channel cut out of it. I was asking what purpose it serves and why that was done.

 

[:D:D]

@lexluthermiester to measure case temperature.

A good example of Intel method.

https://www.intel.sg/content/dam/doc/design-guide/xeon-3000-thermal-design-guide.pdf

This isn't new, for instance

P4 IHS Heat Distribution - Overclockers

Temp analysis of the P4's copper cap - Joe

www.overclockers.com

In my case I have used an encapsulated thermistor with insulated wires epoxied into the larger groove. This was done with a 4 inch cutting disk so would have to go past center to get the depth at the the center of the IHS so I decided might as well go all the way. A thermistor should be more responsive than a thermocouple and can easily (relatively) be connected to the mainboard hardware so can be read directly with software. From this I could see (IIRC) the Tjc thermal resistance of my 6800k is around 0.22C/W which means with a 200W load core temperature would be around 44C above case temperature and case temperature would be ?C above cooler connection due to TIM thermal resistance. So if for example the cooler base is at 40C and TIM loses 10C at 200W then core temp would be at 94C. I did want to also place another sensor in a copper block above the TIM on the CPU IHS and below the cooler to actually see the difference across the TIM but due to my poor soldering need to redo it sometime. Could hope for higher thermal conductivity epoxy aswell but it will have to do for now. I wonder if copper putty would be better but maybe too hard to remove?

 

[lexluthermiester]

@lexluthermiester to measure case temperature.

A good example of Intel method.

https://www.intel.sg/content/dam/doc/design-guide/xeon-3000-thermal-design-guide.pdf

This isn't new, for instance

P4 IHS Heat Distribution - Overclockers

Temp analysis of the P4's copper cap - Joe

www.overclockers.com

In my case I have used an encapsulated thermistor with insulated wires epoxied into the larger groove. This was done with a 4 inch cutting disk so would have to go past center to get the depth at the the center of the IHS so I decided might as well go all the way. A thermistor should be more responsive than a thermocouple and can easily (relatively) be connected to the mainboard hardware so can be read directly with software. From this I could see (IIRC) the Tjc thermal resistance of my 6800k is around 0.22C/W which means with a 200W load core temperature would be around 44C above case temperature and case temperature would be ?C above cooler connection due to TIM thermal resistance. So if for example the cooler base is at 40C and TIM loses 10C at 200W then core temp would be at 94C. I did want to also place another sensor in a copper block above the TIM on the CPU IHS and below the cooler to actually see the difference across the TIM but due to my poor soldering need to redo it sometime. Could hope for higher thermal conductivity epoxy aswell but it will have to do for now. I wonder if copper putty would be better but maybe too hard to remove?

Ah! I've seen this, but never done it myself. What threw me off was the channel cut fully from side to side and very deep. Very interesting.

 

[oiawshboiwsahnoidaws]

@Borc So i tried the paste you mentioned, kooling monster kold-01
Comparing WX-130 and KOLD-01 with Aircooler, 150W load, load duration 30 minutes, fixed fans, no internet (didn't want windows to invalidate results)
Figured also a line chart could be interesting...

... the line chart wasn't really helpful

Max temp with WX-130 was 75°C and KOLD-01 also 75°C
Now for the really funny part over the same number of 2839 data points/seconds, the average temp for WX-130 was 54.42°C and for KOLD-01 54.32°C
... Yeah im done comparing thermal pastes performs way too similiar whats the point lol
Guess i could also test the mx-6 i have and syy-157 with better methodology again but too much of a hassle (for now at least)

 

[jayjr1105]

This thread is quickly becoming "the obscure and unconventional thermal interface material thread"

 

[oiawshboiwsahnoidaws]

Lemme know if you know any other weird / obscure thermal pastes haha

 

[Deleted member 234997]

This thread is quickly becoming "the obscure and unconventional thermal interface material thread"

Your posts contain obscure stuff too, most only know of some brans like TG, Arctic etc.

 

[lexluthermiester]

This thread is quickly becoming "the obscure and unconventional thermal interface material thread"

This thread is titled:

The Official Thermal Interface Material thread​

As such it can contain discussion and comparisons for any material that can used in such a way, regardless of whether or not it is well known or popular.

 

[jayjr1105]

I didn't mean for it to come off as complaining, it's just interesting how it goes down certain rabbit holes for periods at a time.

 

[lexluthermiester]

I didn't mean for it to come off as complaining, it's just interesting how it goes down certain rabbit holes for periods at a time.

No worries. Was just offering that as clarification.

 

[Borc]

@Borc So i tried the paste you mentioned, kooling monster kold-01
Comparing WX-130 and KOLD-01 with Aircooler, 150W load, load duration 30 minutes, fixed fans, no internet (didn't want windows to invalidate results)
Figured also a line chart could be interesting...View attachment 318072

... the line chart wasn't really helpful

Max temp with WX-130 was 75°C and KOLD-01 also 75°C
Now for the really funny part over the same number of 2839 data points/seconds, the average temp for WX-130 was 54.42°C and for KOLD-01 54.32°C
... Yeah im done comparing thermal pastes performs way too similiar whats the point lol
Guess i could also test the mx-6 i have and syy-157 with better methodology again but too much of a hassle (for now at least)

On a desktop (heatspreader) CPU similar temps are expected for the better pastes and max 75°C isn't really high to make a big difference. Nice to see they are similar for your CPU, actually it's a good result for the WX-130. Did you spread Koolingmonster, how was it compared to WX-130?

I tried AMeCh SGT-4 today on my laptop, unfortunately the results are not what I hoped for. PTM7950 and AeroCool Fuzion same temps as KOLD-01, these three are the best for me. Longevity performance could differ though.

 

[sLowEnd]

My cooler (AS500 Plus) came with a mystery tube of paste. I can't find any documentation on it, and there's nothing printed on the tube except "Deepcool"
I haven't gotten around to actually testing it yet since I'm still using up the Ceramique 2 I had lying around. Maybe I'll give it a shot the next time I clean my computer.

I also found a tube of Arctic Silver Matrix in a pile of my old stuff. Unfortunately there's not enough left to actually apply it onto anything.

 

[freeagent]

There should be a few left in that Matrix tube. You push it till its flat

But.. I would check for separation first..

 

[sLowEnd]

There should be a few left in that Matrix tube. You push it till its flat

But.. I would check for separation first..

The tube is basically flat. I pulled the plunger back slightly for the picture

 

[oiawshboiwsahnoidaws]

On a desktop (heatspreader) CPU similar temps are expected for the better pastes and max 75°C isn't really high to make a big difference. Nice to see they are similar for your CPU, actually it's a good result for the WX-130. Did you spread Koolingmonster, how was it compared to WX-130?

I tried AMeCh SGT-4 today on my laptop, unfortunately the results are not what I hoped for. PTM7950 and AeroCool Fuzion same temps as KOLD-01, these three are the best for me. Longevity performance could differ though.

I wanted to do the same application as i used with WX-130 so i ended up doing pea method for kooling monster again
Yeah i think with my pc/cooling/pressure i won't see much difference between the better pastes
I've been interested in ptm7950 and how it performs on a desktop pc and bought ptm7958 but that is only available in a paste, anyone has tips for applying ptmX paste? i only know it has a curing/drying time
maybe i just do the pea method with ptm7958 too... haha just kidding... or am i?

 

[oiawshboiwsahnoidaws]

I tried Thermalright's Heilos "Solid Thermal Interface Sheet" which is likely a PTM7950 re-brand
Kooling monster KOLD01 vs Thermalright Heilos @12700k / AK620 200W Load for 15 minutes with same conditions; fixed fans, same ambient temperature

KOLD01 vs HEILOS

docs.google.com

Interesting to see at start of load the temps are similiar but thermalright heilos improves over time
Max temp for KOLD01 ~86,3 and Heilos ~84,2
Not bad that PTM7950/PCM stuff, kinda expensive and was a pita to apply but really good performance (and longevity should be good i guess)

 

[Pouhon]

I tried Thermalright's Heilos "Solid Thermal Interface Sheet" which is likely a PTM7950 re-brand
Kooling monster KOLD01 vs Thermalright Heilos @12700k / AK620 200W Load for 15 minutes with same conditions; fixed fans, same ambient temperature

KOLD01 vs HEILOS

docs.google.com

Interesting to see at start of load the temps are similiar but thermalright heilos improves over time
Max temp for KOLD01 ~86,3 and Heilos ~84,2
Not bad that PTM7950/PCM stuff, kinda expensive and was a pita to apply but really good performance (and longevity should be good i guess)

Is Heilos double taped like PTM7950? It looks single side taped from product photos.

 

[jayjr1105]

That stuff is 99.99% PTM7950 based. It would have to be double taped, it's really sticky and would get ruined in that brown paper based packaging.

 

[oiawshboiwsahnoidaws]

Yep its double taped one transparent and one opaque side maybe different materials to improve separation?
Honestly for my region thermalright heilos seems a good product as it is cheaper than similiar sized ptm7950 i found online and as i read there may be fake ptm7950 out there if that is true one could think thermalright would have a higher chance to be legit/genuine