The Official Thermal Interface Material thread

[ThrashZone]

Hi,
LM well depends on the surface
Just using as thermal paste replacement on cpu cooler you shouldn't have any issue except maybe using to much and having it drip on anything under it.

Delid wise if the space is not tight and on a vertical case it may eventually ooze downward, damn gravity

Silicon lottery didn't want to change the z height so they loaded the LM on and with shipping/ handling/ computer vibrations my delid only lasted about 6 months before it had to be redone core temps were getting far apart in temp spread.
Similar to mercury so LM does not degrade.

 

[Assimilator]

I'm currently about to drop money on a Threadripper system, where the TDP is way higher than standard Ryzen, and wondering whether I need to invest in something better than MX-4. The thing is, I keep seeing people claiming that less viscous pastes are susceptible to pump-out, yet I've personally never experienced this in my many years of using MX-4 and MX-5 on CPUs. My suspicion is that it's because I always use large AIOs and thus the CPU never gets to the upper range of its temperature limit, i.e. the temperature delta remains lower than the pump-out temperature of the paste.

For anyone who has experienced pump-out, what paste were you using and what sort of application (temperatures)? I apologise if this has already been asked, but seriously there's no way I'm reading the entirety of a 40-page thread.

 

[TPCEA]

I'm currently about to drop money on a Threadripper system, where the TDP is way higher than standard Ryzen, and wondering whether I need to invest in something better than MX-4. The thing is, I keep seeing people claiming that less viscous pastes are susceptible to pump-out, yet I've personally never experienced this in my many years of using MX-4 and MX-5 on CPUs. My suspicion is that it's because I always use large AIOs and thus the CPU never gets to the upper range of its temperature limit, i.e. the temperature delta remains lower than the pump-out temperature of the paste.

For anyone who has experienced pump-out, what paste were you using and what sort of application (temperatures)? I apologise if this has already been asked, but seriously there's no way I'm reading the entirety of a 40-page thread.

With MX5 having been discontinued , did you mean MX6?

Have you considered a phase-change product like Honeywell PTM7950?

I don't get too scientific about thermal pastes. If you can't squirt it onto a CPU and get good results on the first try then it's not worth being patient with it, either it works or not.

The Noctua NT-H1 was a real disappointment. Very easy to spread but the syringe nozzle makes application difficult and it doesn't perform very well. Honestly I should probably throw it in the bin.

I have used Arctic MX-4 previously and thought it was good. The GD900 is far more affordable and performs better, but I'd use MX-4 if I were forced to use something else.

That's odd. I had no trouble at all using NT-H1.

 

[Assimilator]

With MX5 having been discontinued , did you mean MX6?

I used only MX-5 for years until it was discontinued, then switched to MX-4 because MX-6 wasn't yet available.

Have you considered a phase-change product like Honeywell PTM7950?

I am considering it, hence why I'm posting here. But I don't want to bother spending $$$ on a product that, by all accounts, is a PITA to apply - if I can just get away with good old MX-4.

 

[wNotyarD]

I am considering it, hence why I'm posting here. But I don't want to bother spending $$$ on a product that, by all accounts, is a PITA to apply - if I can just get away with good old MX-4.

Can't you just get away with your MX-4 irrespective of any other option? It's not like it won't work with a TR under an AIO block.
Sure, there may be something (marginally?) better, but you'll have to spend for it. I assume you have the MX-4 at hand already.

 

[Pictus]

I am considering it, hence why I'm posting here. But I don't want to bother spending $$$ on a product that, by all accounts, is a PITA to apply - if I can just get away with good old MX-4.

For longer lifespan even at higher temperatures, PCM is the way to go!
PCM is a plastic that melts and because of longer molecular chains it is more resistant to pump-out.

https://www.reddit.com/r/watercooling/comments/z8oa18
https://www.moddiy.com/products/Honeywell-PTM7950-SP-Super-Highly-Thermally-Conductive-PCM-Pad.html
https://www.newegg.com/p/2MB-00B8-00010

PCM needs minimum temperature and pressure.
https://www.caplinq.com/ptm7950-mos...t-works-in-vertical-applications-ptm7950.html
If the geometry of the CPU heat spreader does not match/fit well with the
cooler baseplate, such as a flat CPU heat spreader + a cooler with convex baseplate,
the performance of the PCM will probably be worse than with non-PCM thermal paste.

 

[Franz]

Can't you just get away with your MX-4 irrespective of any other option? It's not like it won't work with a TR under an AIO block.
Sure, there may be something (marginally?) better, but you'll have to spend for it. I assume you have the MX-4 at hand already.

GD007 search on ML for it and you can get one 3g for only R$14~25

 

[wNotyarD]

GD007 search on ML for it and you can get one 3g for only R$14~25

Wait, I'm not the one asking for thermal pastes better than MX-4. @Assimilator is.

 

[Count von Schwalbe]

Wait, I'm not the one asking for thermal pastes better than MX-4. @Assimilator is.

Actually, he is asking if MX-4 suffers from pump-out on large and hot IHS surfaces.

 

[wNotyarD]

Actually, he is asking if MX-4 suffers from pump-out on large and hot IHS surfaces.

Yes, but I understood he also asked if he could just get away with MX-4 even if it does suffer.

 

[lexluthermiester]

I'm currently about to drop money on a Threadripper system, where the TDP is way higher than standard Ryzen, and wondering whether I need to invest in something better than MX-4. The thing is, I keep seeing people claiming that less viscous pastes are susceptible to pump-out, yet I've personally never experienced this in my many years of using MX-4 and MX-5 on CPUs. My suspicion is that it's because I always use large AIOs and thus the CPU never gets to the upper range of its temperature limit, i.e. the temperature delta remains lower than the pump-out temperature of the paste.

For anyone who has experienced pump-out, what paste were you using and what sort of application (temperatures)? I apologise if this has already been asked, but seriously there's no way I'm reading the entirety of a 40-page thread.

If you have MX-4, use it. Unless you intend to be doing any extreme overclocking, MX-4 will perform very well. However, if you have some MX-5 left, use that instead as it's better than MX-4. Threadripper CPUs put out the heat, but not enough to make a huge difference enough to justify buying new TIM, unless you might run out. If you are going to do some OCing with your Threadripper, then get some KPx(KingPinExtreme). Simply put, it's the best of the best for non-liquid-metal TIMs.

Pump-out is a very controversial topic. I think it's total hogwash but some claim to have seen it. Regardless, there is no evidence MX-4, 5 or 6 suffers from it.

 

[wNotyarD]

Also, did you guys see the KryoSheet review W1zz made today?

Thermal Grizzly KryoSheet Review - Tested on RX 7900 XTX with 475 W

Thermal Grizzly KryoSheet is an interesting alternative to thermal paste. The pre-cut sheet doesn’t dry out over time and you don’t have to try spreading it in a thin layer. In our review we test thermal performance on the AsRock RX 7900 XTX, at up to 475 W.

www.techpowerup.com

 

[Borc]

I used only MX-5 for years until it was discontinued, then switched to MX-4 because MX-6 wasn't yet available.


I am considering it, hence why I'm posting here. But I don't want to bother spending $$$ on a product that, by all accounts, is a PITA to apply - if I can just get away with good old MX-4.

If there is a bad contact MX-4 is not a good choice, you would need something more viscous. Nobody can say if and how much you can gain with PTM7950/Gelid Heatphase or whatever rebrand, however it could potentially improve the temps a lot on a bare die chip. There are only a few pastes in the market who can be close to PTM7950, for example you could try out Thermal Hero Quantum. Some OEMs use pastes like DOWSIL TC-5888 or Shin-Etsu X-23-8117 but they are hard to buy.

 

[TPCEA]

If there is a bad contact MX-4 is not a good choice, you would need something more viscous. Nobody can say if and how much you can gain with PTM7950/Gelid Heatphase or whatever rebrand, however it could potentially improve the temps a lot on a bare die chip. There are only a few pastes in the market who can be close to PTM7950, for example you could try out Thermal Hero Quantum. Some OEMs use pastes like DOWSIL TC-5888 or Shin-Etsu X-23-8117 but they are hard to buy.

What about @freeagent 's favorite: SYY-157?

 

[freeagent]

What about @freeagent 's favorite: SYY-157?

My favorite is Thermalright TFX

But SYY is almost as good for a fraction of the price.

 

[TPCEA]

My favorite is Thermalright TFX

But SYY is almost as good for a fraction of the price.

I didn't know that. I don't think I've ever seen you recommend it.

 

[Pictus]

If there is a bad contact MX-4 is not a good choice, you would need something more viscous. Nobody can say if and how much you can gain with PTM7950/Gelid Heatphase or whatever rebrand, however it could potentially improve the temps a lot on a bare die chip. There are only a few pastes in the market who can be close to PTM7950, for example you could try out Thermal Hero Quantum. Some OEMs use pastes like DOWSIL TC-5888 or Shin-Etsu X-23-8117 but they are hard to buy.

The Shin-Etsu X-23-8117 is new to me, I have and like the X-23-7921-5.

My favorite is Thermalright TFX

It has the best smell in thermal compounds, eugenol!

https://hal.science/hal-03343073/document

 

[Borc]

What about @freeagent 's favorite: SYY-157?

Quite a good paste in my testing, although not as good as Thermal Hero Quantum or Kold-01, Aerocool Fuzion and of course PTM7950/Heatphase Ultra.

The Shin-Etsu X-23-8117 is new to me, I have and like the X-23-7921-5.

It's a relatively new Shin-Etsu. Framework Computer switched to this paste in their laptops. They even sell 0.5g tubes. I will test both soon btw.

https://twitter.com/i/web/status/1508532776022122497

Some sources say Lenovo uses Dowsil TC-5550 for their Lenovo X Power Cooling System. That's interesting because they usually use PTM7950 for their higher end laptops. Dowsil TC-5550 is especially designed for bare die application as a PCM alternative. Unfortunately it's impossible to buy for small customers.

 

[freeagent]

I didn't know that. I don't think I've ever seen you recommend it.

That is why I always say SYY lol.. for the price it’s pretty good

I am sure you can do better, but for me it’s good.

 

[wNotyarD]

It has the best smell in thermal compounds, eugenol!

Didn't someone bring up a Japanese apple-scented paste in this thread?

 

[68Olds]

Didn't someone bring up a Japanese apple-scented paste in this thread?

yup, it's a few pages back...

 

[Assimilator]

That is why I always say SYY lol.. for the price it’s pretty good

I am sure you can do better, but for me it’s good.

I have no trust in this SYY stuff when its thermal conductivity is supposedly 15.7W/mk, yet their very own listing for the 2g states only 4.5W/mk.

 

[freeagent]

I just go on thin

And it’s cheap

 

[lexluthermiester]

I have no trust in this SYY stuff when its thermal conductivity is supposedly 15.7W/mk, yet their very own listing for the 2g states only 4.5W/mk.

I completely ignore those kinds of numbers as they are completely meaningless without a standardized system of comparision. The only thing I pay attention to is actual, real-world performance.

 

[freeagent]

I have no trust in this SYY stuff when its thermal conductivity is supposedly 15.7W/mk, yet their very own listing for the 2g states only 4.5W/mk.

It’s funny how it has a completely different description here..