The Official Thermal Interface Material thread

[Noci]

But I do need to differ, as Ni has thermal conductivity (later TC) of ~58W/mK, Cu has TC of about ~380W/mK, while CPU IHS has TC of ~315W/mK.
Why has CPU IHS so much less TC then Cu? Well, because of the Ni plating, which has about 6~7 times less TC.
So putting pure Cu CPU IHS will get you something of more TC, but it is better to scrape down the Ni & to have rougher surface (not polished).
Why? Again, you can so more damage with replacing CPU IHS, then using fine sanding paper & getting rid of that Ni from the top.

With theoretical numbers you're absolutely right, but I can't recall any review or practical test where only the Ni has been removed (not lapped) from the IHS and the cooling solution gained 6 to 7 degrees Celsius difference in efficiency.
If it was this easy, don't you think the chipmakers would have omitted the nickel plating on the outside of the IHS, as that area is quite simple to mask during the Nickel application process (electrolytic or chemically).

Like mentioned before, the contact surfaces between the IHS & cooling plate being convex or concave, the applied TIM, the mounting pressure and even more variables have a way bigger impact than just this thin layer of Nickel. So in my opinion removing it is just a waste of time and voiding warranty on the chip.

Even if you think Noci is talking BS and you do remove the 58 W/mK Nickel layer, you still have to apply a TIM with an average TC of 5 W/mK and in the end there is no gain.

Just try to imagine heat being a liquid flowing through a piping system at a fixed pressure. This system has several restrictions in diameter, where the smallest diameter (in this case representing the lowest TC) is the biggest flow restriction.
Once you remove that one, then you go to the next one that is the smallest and so on until there is no flow restriction anymore. No use to remove the second one if the first one is still in the system.

Well, not sure...as Graphene I have looked about & found it is electrically conductive (so that is a no go) & also is very thermally polarized in its conduciveness (more conductive in line with the sheet or straws, less conductive perpendicular to sheets or straws). So although this is a good example & it is an interesting material, it can't solve anything!

That's why I mentioned fabrication & application techniques still have to mature, but many scientists are lyric about it's potential.

If you don't want an electrical conductive TIM, you're pretty much restricted to this average 5 W/mK TIM layer. Electrical conductive TIM's like Graphene and Gallium (liquid metal) have been proven solutions for the best results, but yes they do introduce certain risks.

Unless you're a hardcore OC'er, it's better to get educated on proper TIM application, perfect mounting of your cooler than trying exotic solutions that have a big risk factor and hardly an impact on the cooling efficiency.

Note, I'm talking about efficiency, not capacity that's another story.

 

[KLiKzg]

With theoretical numbers you're absolutely right, but I can't recall any review or practical test where only the Ni has been removed (not lapped) from the IHS and the cooling solution gained 6 to 7 degrees Celsius difference in efficiency.
If it was this easy, don't you think the chipmakers would have omitted the nickel plating on the outside of the IHS, as that area is quite simple to mask during the Nickel application process (electrolytic or chemically).

Problem with CPU is the most of them are mounted with "bare hands", usually not ESD gloves. & the acids & bacteria found on hands, do not play well with Cu.

Also, intel is shipped World wide. In order to have Cu not to oxidize while it gets to you, not going to happen. So to keep the CPU nice, they Ni plate it.

Yes, the thermal conduciveness is lowered, but Ni plays so well with Cu...they could also Zn plate it, but it would cost more...

Like mentioned before, the contact surfaces between the IHS & cooling plate being convex or concave, the applied TIM, the mounting pressure and even more variables have a way bigger impact than just this thin layer of Nickel. So in my opinion removing it is just a waste of time and voiding warranty on the chip.

Again, I advise not to void guarantee on the chip itself...but once the warranty is over, with opening of computer, then people can do what they want, even replacing TIM.

Even if you think Noci is talking BS and you do remove the 58 W/mK Nickel layer, you still have to apply a TIM with an average TC of 5 W/mK and in the end there is no gain.

Yes, that is why I use 10+W/mK TIM paste, as it does make a few °C of difference.

Just try to imagine heat being a liquid flowing through a piping system at a fixed pressure. This system has several restrictions in diameter, where the smallest diameter (in this case representing the lowest TC) is the biggest flow restriction.
Once you remove that one, then you go to the next one that is the smallest and so on until there is no flow restriction anymore. No use to remove the second one if the first one is still in the system.

You are right, TIM is the mayor pipe blocker...but also, thermal conductivity does not work, like the fluid dynamics - got mech.eng. degree to prove it!

That's why I mentioned fabrication & application techniques still have to mature, but many scientists are lyric about it's potential.

If you don't want an electrical conductive TIM, you're pretty much restricted to this average 5 W/mK TIM layer. Electrical conductive TIM's like Graphene and Gallium (liquid metal) have been proven solutions for the best results, but yes they do introduce certain risks.

Will come back to you, as I am starting to experiment a little...so keep on posted, maybe I can get some people here a solution.

Unless you're a hardcore OC'er, it's better to get educated on proper TIM application, perfect mounting of your cooler than trying exotic solutions that have a big risk factor and hardly an impact on the cooling efficiency.

Note, I'm talking about efficiency, not capacity that's another story.

I am not OC'er, just to mention here. But my computers run 24/7 on BOINC, doing valuable calculations. So any advantage in cooling is welcomed!

 

[Noci]

You are right, TIM is the mayor pipe blocker...but also, thermal conductivity does not work, like the fluid dynamics - got mech.eng. degree to prove it!

I'm aware that fluid dynamics and thermodynamics are two completely different disciplines in physics, each obiding to their applicable laws of physics.

But what I tried to establish here by using a metaphor, is for other readers (without an engineering level education in physics) to better/easier understand to point I was making.
My attempt of simplifying the complex thermal conductivity concept by suggesting some everyday example which most of us can relate to and even to a certain level visualize it to comprehend it.
Nothing more, nothing less, don't overthink it .

Will come back to you, as I am starting to experiment a little...so keep on posted, maybe I can get some people here a solution.

If you succeed in finding a way for limitless optimal usage of Graphene (including manufactring and application techniques), I'll most likely see you soon leading the list of Forbes 500 .
But do realise, you have tough competition as plenty of scientists, universities and companies are on it too.
Some would even call it the next thing to superconductivity .

 

[Shrek]

You are right, TIM is the mayor pipe blocker...but also, thermal conductivity does not work, like the fluid dynamics - got mech.eng. degree to prove it!

How does thermal conductivity not work?

 

[KLiKzg]

I'm aware that fluid dynamics and thermodynamics are two completely different disciplines in physics, each obiding to their applicable laws of physics.

But what I tried to establish here by using a metaphor, is for other readers (without an engineering level education in physics) to better/easier understand to point I was making.
My attempt of simplifying the complex thermal conductivity concept by suggesting some everyday example which most of us can relate to and even to a certain level visualize it to comprehend it.
Nothing more, nothing less, don't overthink it .

I do understand the metaphor. But even that analogy in metaphor is wrong, when you do look at how fluid behave in dynamics & how transfer of heat is transferred between layers through 2 different laws of physics, acting together in the same way to conduct the passage of thermal energy. That is why I did say, what was said.

After all, we are not in kindergarten here. This is a quite serious topic on a very serious testing site, which I very much like. So we need to communicate in a way that is more accurate & deliberate in transferring of information & knowledge.

If you succeed in finding a way for limitless optimal usage of Graphene (including manufactring and application techniques), I'll most likely see you soon leading the list of Forbes 500 .
But do realise, you have tough competition as plenty of scientists, universities and companies are on it too.
Some would even call it the next thing to superconductivity .

You never know...have some crazy ideas in my mind.

How does thermal conductivity not work?

It does work, just not in the same way as fluid dynamics...as you can't look at the thermal conductivity only & solely. Because you also have "heat emission", with thermal conductivity through touching surfaces.

The forms of energy transfer is different, formulas are completely different...yes, you can say one thing goes from A to B. But we are not kids here!

Not to mention that we do not understand the thermal conductivity! Why? We use too much constants, which is bad!
Everytime any science is using a constant, means it does not understand that science.
Example: gravity. We use gravity constant, for space equations & for Earth. Why? We do not understand the gravity, as we should. There are people working on the physical laws how to understand gravity, searching for graviton (as a particle), finding a way to explain & manipulate gravity... We still to not understand the gravity, so we all use just "gravity constants" for our everyday math & physics.

 

[Shrek]

Heat emission? of course heat travels through the TIM

I agree, thermal conductivity changes with temperature; but don't see that as a big issue.

For me the missing ingredient may be the interface between paste and surface... we seem to assume zero temperate drop at these interfaces.

 

[Noci]

After all, we are not in kindergarten here. This is a quite serious topic on a very serious testing site, which I very much like. So we need to communicate in a way that is more accurate & deliberate in transferring of information & knowledge.

No not Kindergarten here but a tech site indeed.
But a lot op peeps come here to learn/understand about a certain topic/subject or problem and I would not be suprised if there are a number of kids amongst them.

While quite some members here have a plethora of knowledge, this metaphor example is most likely unneeded for them.
But keep in mind that besides knowledgable members (and guests) we also do have a multitude of site vistors (being it members or guests) with less technical insights. If you just analyze a random number of posts with questions and even sometimes answers, you can often read between the lines the lack comprehension of the science involved.
And one of the main purposes of a techsite like this IMO, is to share knowledge/experiences and prevent from misinformation or confusion.

I'll try with another figure of speech to stay on topic;

So as I explained the TIM is the weakest link in the chain of a cooling system, hence this is why this very thread exists (the community searching for the best TIM and sharing experiences in their specific case scenarios).
All other nitty gritty about removing and or reinforcing other links in the chain, does not make the chain stronger, the weakest link still determinates the strength of the chain.

Hope this was a better metaphore than before .

 

[lexluthermiester]

I agree, thermal conductivity changes with temperature; but don't see that as a big issue.

In the context of PC's & such, it's not an issue.

 

[80-watt Hamster]

I do understand the metaphor. But even that analogy in metaphor is wrong, when you do look at how fluid behave in dynamics & how transfer of heat is transferred between layers through 2 different laws of physics, acting together in the same way to conduct the passage of thermal energy. That is why I did say, what was said.

After all, we are not in kindergarten here. This is a quite serious topic on a very serious testing site, which I very much like. So we need to communicate in a way that is more accurate & deliberate in transferring of information & knowledge.


You never know...have some crazy ideas in my mind.


It does work, just not in the same way as fluid dynamics...as you can't look at the thermal conductivity only & solely. Because you also have "heat emission", with thermal conductivity through touching surfaces.

The forms of energy transfer is different, formulas are completely different...yes, you can say one thing goes from A to B. But we are not kids here!

Not to mention that we do not understand the thermal conductivity! Why? We use too much constants, which is bad!
Everytime any science is using a constant, means it does not understand that science.
Example: gravity. We use gravity constant, for space equations & for Earth. Why? We do not understand the gravity, as we should. There are people working on the physical laws how to understand gravity, searching for graviton (as a particle), finding a way to explain & manipulate gravity... We still to not understand the gravity, so we all use just "gravity constants" for our everyday math & physics.

No not Kindergarten here but a tech site indeed.
But a lot op peeps come here to learn/understand about a certain topic/subject or problem and I would not be suprised if there are a number of kids amongst them.

While quite some members here have a plethora of knowledge, this metaphor example is most likely unneeded for them.
But keep in mind that besides knowledgable members (and guests) we also do have a multitude of site vistors (being it members or guests) with less technical insights. If you just analyze a random number of posts with questions and even sometimes answers, you can often read between the lines the lack comprehension of the science involved.
And one of the main purposes of a techsite like this IMO, is to share knowledge/experiences and prevent from misinformation or confusion.

I'll try with another figure of speech to stay on topic;

So as I explained the TIM is the weakest link in the chain of a cooling system, hence this is why this very thread exists (the community searching for the best TIM and sharing experiences in their specific case scenarios).
All other nitty gritty about removing and or reinforcing other links in the chain, does not make the chain stronger, the weakest link still determinates the strength of the chain.

Hope this was a better metaphore than before .

This discussion puts me in mind of the ol' Scientist vs. Engineer dichotomy. The Scientist (KLiKzg) is after the correct answer; the Engineer (Noci) is concerned with the effective answer.

Using the gravity example, the constant g doesn't fully describe the influence of Earth's gravity on its surroundings. But it, along with the published modulus of elasticity and such of a bungee cord calculated against a jumper's mass, allows the leap from span with confidence there's not a faceplant to terra firma waiting at the other end. Pi is also not 3.14, but that value is precise enough in a great number of instances.

When it comes to TIMs in PCs, I (neither engineer nor scientist by trade or profession, but more Engineer than Scientist by inclination) feel the Engineer's perspective is the more applicable (not to imply that of the Scientist isn't valuable). We're not developing the TIM, IHS nor heatsink/coldplate; merely combining these components the best we can. Given that manufacturer-provided conductivity numbers are both notoriously incomparable and an order of magnitude lower than those of the surfaces they're joining, we're kind of left with best practices, experimentation and sharing of experiences and data. Good thing we've got this thread, yeah?

In light of the above, it's probably no surprise that I consider physical properties of a TIM (material compatibility, service life, shelf stability, application characteristics) than 3C or even 5C of ultimate performance. I'd much rather work around the loss a few J/s of thermal transfer than need to re-TIM on a timescale of months or deal with LM.

 

[Shrek]

KLiKzg is a mechanical engineer

 

[Count von Schwalbe]

I woulda said engineer vs contractor.

 

[Noci]

Well this all does remind me of a 'food for thoughts' text I encountered once as a wall decoration in a metallugic laboratory, it went something like this:

Theory is when you know everything, but nothing works.
Practice is when everything works, but no one knows exactly why.
In our Laboratory we combine theory and pratice.
Result; nothing works and no one knows exactly why .

Guess this says it all .

 

[freeagent]

That is true. I have recommended TIMs that people hate, but they are awesome for me

 

[Shrek]

I like GD900

All other nitty gritty about removing and or reinforcing other links in the chain, does not make the chain stronger, the weakest link still determinates the strength of the chain.

I would not actually call the TIM the weakest link given that it can at best account for 2°C change (between TIMs)

 

[Noci]

Atm still using MX-4 to full satisfaction, but have these gems laying around for when I'll redo my custom loop.

Meanwhile I'll keep on monitoring this thread to see if there's a new revolutionary TIM that's worth of trying out.

 

[Shrek]

My issue is cost... I'm rather tight fisted... I mean parsimonious.

 

[KLiKzg]

I like GD900

View attachment 364377



I would not actually call the TIM the weakest link given that it can at best account for 2°C change (between TIMs)

This paste was used in for 1MW device, from my previous company...it has proven to be OK with heat dissipation in automotive industry!

 

[lexluthermiester]

This paste was used in for 1MW device

Did you say that right? 1MW = 1megawatt = 1 million watts

 

[Shrek]

Over 1000 HP, maybe brakes?

 

[lexluthermiester]

Over 1000 HP, maybe brakes?

No idea. I can't think of anything on a common road vehicle that can generate 1MW of heat.

 

[ShrimpBrime]

Over 1000 HP, maybe brakes?

1000HP is only 745,000 watts.

But there's no application for thermal paste in the automotive industry outside an ECM housing. On older diesels they would use fuel heatsink to heat and cool the ECM during summer/winter. This kept the ECM unit running roughly the same as the coolant temp.

 

[Shrek]

I wrote 'Over 1000 HP'

 

[ShrimpBrime]

I wrote 'Over 1000 HP'

OK then 1001 HP is 746,445 watts

 

[KLiKzg]

No idea. I can't think of anything on a common road vehicle that can generate 1MW of heat.

1MW of power...not heat! & yes, it is 1MW...

& I can think of...several vehicles with that much power!

1000HP is only 745,000 watts.

But there's no application for thermal paste in the automotive industry outside an ECM housing. On older diesels they would use fuel heatsink to heat and cool the ECM during summer/winter. This kept the ECM unit running roughly the same as the coolant temp.

Oh, this is for sure - not true...not true at all!

 

[ShrimpBrime]

1MW of power...not heat! & yes, it is 1MW...

& I can think of...several vehicles with that much power!


Oh, this is for sure - not true...not true at all!

For a CPU, ALL of the energy in electrical wattage is converted to thermal wattage. What you put in, comes out as heat.

And yes it is true!

HP is cetrifugal force converted to wattage. This is not the wattage put into the engine which is a liquid fuel. Engines are roughly only 30% effecient on average.

So, NO. There is no at home PC running a single cpu at 1Mw of power. Nor if HP measured by the energy put in, but what comes out.