How does one review the effectiveness of thermal paste?

[Zach_01]

That is why I used the word "seem"

Well, for this matter (cooling a CPU/GPU)... seems = wrong. Because you must completely ignore 1 variable/parameter to come to the conclusion that all TIMs are equal. Thats why every legit review is done by normalizing(equaling) all parameters between contestants except one. That would be fan rpm or temp.

 

[Deleted member 191766]

Exactly; that was my point from post 1

How does one review the effectiveness of thermal paste?

This might seem a silly question, one just reports the CPU temperature under the same load;
but the fan might be controlled to run such that the CPU temperature remains at a given set
temperature.
Under such feedback most all thermal pastes would seem to perform nearly the same,
toothpaste included.

Would fan speed then perhaps be a better measure?

 

[delshay]

Thats why, at least for Contactonaut, a huge red paper inside the package states the "Do not use with Aluminium" along with some Xs.

I use liquid metal for about 8~9months now. Nickel palated CPU IHS and copper coldplate of the AIO block/pump package. Its expensive compound not only for its price on purchase but also because you need to do several applications when copper coldplates are at least the 1 side of the 2 contact surfaces. If both sides are copper I assume you need to replace it sooner, at least in half time.

As of performance... yes it is the best TIM you can get. I saw about 5~6C reduction on max temp (100% load) compared to ArcticSilver5.
Does it worth it? That depends on how you see things. For me yes it does, and I knew from the biginning that with copper needs re-application after short(er) period of time and not just once. I dont mind the cost, its too small among other expenses.

I saw der8ouer talking about it saying that needs replacement after 2 months for at least 2~3 times before you leave it for long. I find that for the first replacement it needs to be done on the month, not 2. For the second its ok to leave it 2~3months.
I first replace it after ~50 days and it was a bit late. Gallium had saturated copper in such amount that whats left behind was almost like very thin layer of dried compound, like a thin leave.

View attachment 171570

View attachment 171571

View attachment 171572

It was a little hard to clean IHS but the AIO coldplate was a lot lot harder. I did some abrasion to remove it from copper surface from certain points/spots.

This is the final result.

View attachment 171573

Next application

View attachment 171574

After that I left it for 3 months and when I remove it again it was a lot easier for both as it wasnt dried this time. I'm on the 3rd month now (of 3rd application) and I will do it again next month. After that the copper saturation from gallium would be enough and slowed down for the TIM to last even longer.
You can see that AIO's coldplate is not as it was new, little scraped from "cleaning" process, but I did not see any cooling performance drop. I quess the high thermal conductivity of liquid metal makes up for it.

Another thing to consider other than cost, re-application and cleaning is the danger of shortcircuit components around the socket or the CPU itself. Needs extra care treatment, especially for beginners.

The danger here is, if you clean the PC with compress air the LM will get blown everywhere. It happened to me & landed on my fairly new R9 Nano a few years back. Did not see it. Turn PC on & BANG. Still could not see where the bang sound came from.

Turn PC on again, but this time it came on, but no display. Found R9 Nano to be super cold. So I took the card out then I spotted the LM on the card. I can now see what happened. The bang sound sent LM across a very small area of the card freeing the short circuit, this is why the PC came on the second time around. Force to buy an replacement R9 Nano which was still expensive at that time.

 

[Zach_01]

Exactly; that was my point from post 1

I dont know how the Guru3D test, we previously see was executed exactly.
If they let fan rpm to vary between different TIMs based on some rpm/temp curve (same curve on all) then this bring results closer together. But its a more real life approach. What they should've done is to report fan rpm also, but this could complicate things in the eyes of the user, depending on our own individual understanding of the issue here.

To get better results (as more distinquished between them) you can do 2 things, 1 or the other.

1. Set fan rpm same for all and see on what temp each ends-up
2. Set (a reasonably) target temp and see on what rpm each end-up

The danger here is, if you clean the PC with compress air the LM will get blown everywhere. It happened to me & landed on my fairly new R9 Nano a few years back. Did not see it. Turn PC on & BANG. Still could not see where the bang sound came from.

Turn PC on again, but this time it came on, but no display. Found R9 Nano to be super cold. So I took the card out then I spotted the LM on the card. I can now see what happened. The bang sound sent LM across a very small area of the card freeing the short circuit, this is why the PC came on the second time around. Force to buy an replacement R9 Nano which was still expensive at that time.

Yeah... you dont do that with LM in the first place, and second you apply less so that it doesnt come out the edges.

 

[Bill_Bright]

I prefer thermal grease that lasts indefinitely; I don't mind a few degrees hotter.

Thats Ok!
I'm dont have a mindset that every one must or should use it.
I prefer high performance...So?

"High performance" where?

Again, the range from worst to best typically is just a few degrees. In that Guru3D roundup, ~8°C was the widest gap and even then, the top temps, after running the CPUs at 100% for 30 continuous minutes (definitely NOT a typical real world scenario), the CPUs never breached any thermal protection thresholds or Tjunction ratings.

So in terms of "performance", even the least efficient TIM in that revue allowed the CPUs in the test to run at their highest rated performance levels.

So what does that mean? It means you can pick just about any TIM and be confident it will ensure efficient heat transfer to keep your CPU adequately cooled AS LONG AS, the TIM is properly applied, case cooling is properly setup to provide an adequate supply of cool air flowing through the case, and ambient temps are not so high they prevent all hopes of adequate cooling.

If you just have to have the lowest temps possible, that's fine! Just don't think it offers better performance. All it is doing is giving you bragging rights.

Because you must completely ignore 1 variable/parameter to come to the conclusion that all TIMs are equal.

I don't know of anyone who is claiming all TIMs are equal.

Thats why every legit review is done by normalizing(equaling) all parameters between contestants except one. That would be fan rpm or temp.

Wait! What? Except one? I don't agree. Legitimate reviews go to great lengths to level the playing field before actual testing. These steps include ensuring an equal ambient temp. They use the same CPU(s) and coolers/cooler fans. They allow each TIM to cure (even those that don't need curing). Guru3D allowed for 72 hours. They run those same CPUs, using the same programming to 100% utilization, an equal amount of warm-up time, which, in turn, pushes the fan speeds (the same fans used on all tests) to maximum RPM.

IF the maximum temp with a particular TIM does not cause the CPU fan to maximum speed, then that's a testament to the effectiveness of the TIM. Not a fault of the testing methods.

That would be fan rpm or temp.

I'm confused. "Or temp"? What temp? Ambient (room) temperature is fixed and controlled in these roundups. The CPU temp is the one variable that should NOT be fixed or controlled as that is the one variable you are testing to determine a TIM's effectiveness.

The danger here is, if you clean the PC with compress air the LM will get blown everywhere.

??? Then I would say you applied way too much TIM and/or you came in way too close with way too much pressure.

I use an air compressor all the time in my have. Have for decades. Note the recommended PSI for cleaning electronics is 40 - 80PSI. And you should always filter the air through a suitable moisture and particulate filter. And of course, always, as in EVERY SINGLE TIME, where eye protection and I recommend a quality dust filter mask too.

 

[delshay]

??? Then I would say you applied way too much TIM and/or you came in way too close with way too much pressure.

I use an air compressor all the time in my have. Have for decades. Note the recommended PSI for cleaning electronics is 40 - 80PSI. And you should always filter the air through a suitable moisture and particulate filter. And of course, always, as in EVERY SINGLE TIME, where eye protection and I recommend a quality dust filter mask too.

I was fairly new to LM at that time & I only knew the danger of short circuit & using it against some metals. Did not think LM will fall onto the GFX card as I did not remove the CPU cooler. LM will always be exposed just a little if you apply it to IHS & coldplate, which you need to do if user(s) go down this route. When you put the two together some LM will always spill over beyond the IHS, otherwise if user(s) were applying the LM with perfection, then you will be seeing a perfect shape of the IHS on the coldplate.

Yes, I did get to near the CPU with compressor hose, but I learnt my lesson, no LM to be applied between IHS & Coldplate, use only between die & IHS. One expensive fix. Man, did it hurt to buy a replacement card.

 

[Deleted member 191766]

I use an air compressor all the time in my have. Have for decades.

I use a blower but do recall that it is possible to develop static electricity. I recall some dry C02 was used to inflate a ballon and it ended up destroying the circuit due to the static developed.

DataVac has a special ESD blower.

 

[Zach_01]

"High performance" where?

Again, the range from worst to best typically is just a few degrees. In that Guru3D roundup, ~8°C was the widest gap and even then, the top temps, after running the CPUs at 100% for 30 continuous minutes (definitely NOT a typical real world scenario), the CPUs never breached any thermal protection thresholds or Tjunction ratings.

So in terms of "performance", even the least efficient TIM in that revue allowed the CPUs in the test to run at their highest rated performance levels.

So what does that mean? It means you can pick just about any TIM and be confident it will ensure efficient heat transfer to keep your CPU adequately cooled AS LONG AS, the TIM is properly applied, case cooling is properly setup to provide an adequate supply of cool air flowing through the case, and ambient temps are not so high they prevent all hopes of adequate cooling.

If you just have to have the lowest temps possible, that's fine! Just don't think it offers better performance. All it is doing is giving you bragging rights.

We are talking about TIM performance in this thread I think and not PC performance... Sorry for not being more clear about that.
Even though my ZEN2 CPU benefit from 5~6C lower temp for higher all core boost, I wouldnt consider it as reason to use the best TIM just for that. Anyway the benefit is +25~50MHz of the 3~6C lower temp across all work loads and that is realistically nothing. No real life gains.
I want lower temps just because I want them and for lower fan speed/noise too. ZEN2 with the chiplet design, the 7nm small die(s) = desner heat area, actually is benefited more than others in terms of lower temp, but it doesnt mean that working at 66C instead of 60C would hurt the chip. No I'm not saying that and I dont use LM for that.

I don't know of anyone who is claiming all TIMs are equal.

Havent you follow this thread from the beginning? and what the OP is trying to say with the fan rpm.
By the way you must have lost my point with the phrase:

"...you must completely ignore 1 variable/parameter to come to the conclusion that all TIMs are equal"

..which the OP was saying above. Even tho he ment "seems" to be equal, if you run fan rpm accordingly for same CPU temp. And I was trying to explain to him that this is wrong because for the same CPU temp it would reguire different fan rpm for each TIM, and so this implies different thermal characteristics of each TIM = different heat transfer performance.

Why do I have to repeat my self? Didnt you follow the conversation up to that point?

Wait! What? Except one? I don't agree. Legitimate reviews go to great lengths to level the playing field before actual testing. These steps include ensuring an equal ambient temp. They use the same CPU(s) and coolers/cooler fans. They allow each TIM to cure (even those that don't need curing). Guru3D allowed for 72 hours. They run those same CPUs, using the same programming to 100% utilization, an equal amount of warm-up time, which, in turn, pushes the fan speeds (the same fans used on all tests) to maximum RPM.

IF the maximum temp with a particular TIM does not cause the CPU fan to maximum speed, then that's a testament to the effectiveness of the TIM. Not a fault of the testing methods.

I'm confused. "Or temp"? What temp? Ambient (room) temperature is fixed and controlled in these roundups. The CPU temp is the one variable that should NOT be fixed or controlled as that is the one variable you are testing to determine a TIM's effectiveness.
??? Then I would say you applied way too much TIM and/or you came in way too close with way too much pressure.

Did you read first half of post#29?

Are we really doing this? We are saying the same things...

"Thats why every legit review is done by normalizing(equaling) all parameters between contestants except one. That would be fan rpm or temp"

This is excactly what Guru3D did. They normalize all parameters as you said with same cooler, same loads, curing, ambient temp, fan rpm to the max... except one. The CPU temp.

You can see this the other way around, although never done this way for graspable reasons.
Normalize all parameters, as same cooler, same loads, curing, ambient temp, CPU temp(to a high value like 85C but keep fan rpm vary to achieve the 85C target.

See my point and what I was trying to explain to OP?
TIMs are not equal. You either end up with different CPU temp or different fan rpm or both.

Are we saying anything different?

 

[Deleted member 191766]

Even tho he ment "seems" to be equal

I meant the exact opposite, that they would only look like they performed equally; it would be crazy to claim all thermal pastes are of equal performance.

 

[Zach_01]

I meant the exact opposite, that they would only look like they performed equally; it would be crazy to claim all thermal paste are of equal performance.

Ok, but in the end there are not if the fan speed is different. So how is that a valid point? ...or what exactly are you trying to say?

Sincere question(s)

 

[Deleted member 191766]

See post #1, I think it is all there.

 

[ZenZimZaliben]

You don't even need to use a CPU to test thermal paste. If you are just looking to get the delta. Get a hot plate and a cold plate and measure the heat transfer between. Swap out thermal pastes to see difference. That is how it is done in development before it even touches a cpu/motherboard. You have a plate that can be electronically dialed in to very exact wattage loads. With very accurate thermal probes.

 

[Deleted member 191766]

I agree, but for one factor, the clamping pressure; but one could build something based on the original heatsink.

 

[Bill_Bright]

When you put the two together some LM will always spill over beyond the IHS

Always? Ummm, nope. Sorry. Not true. "All" TIMs must be applied with as little as possible but still provide for complete coverage. LM TIMs must be applied in even thinner layers because of how well it spreads - but it still does not spread indefinitely. If yours spilled out and over, you applied too much.

Note what many do is paint over nearby electrical contacts with clear nail polish, or even epoxy resins - just in case.

Man, did it hurt to buy a replacement card.

LOL Well, I cringe to dredge up old memories of my own stupid (hindsight is 20/20), avoidable and expensive mistakes. Even as a formally trained, educated, and certified electronics technician, much of my own personal knowledge and experience comes via the hard way.

I use a blower but do recall that it is possible to develop static electricity.

Blowers like that in your image are for use around ESD sensitive electronics. Those are nice for house calls or occasional use, but they really don't provide enough pressure for thorough cleaning. So if you do a lot of electronics cleaning, a properly filtered air compressor works best. And of course, an air compressor can be used for all sorts of other tasks, like vehicle maintenance, roofing/nail guns, painting, etc.

Static discharges (voltage arcs) only happen when the different "poles" comes in close proximity to each other, and when the voltage has a high enough difference in "potential" to arc across the gap. With an air compressor, you should never get closer than a few inches anyway - which is way too big a gap for the voltage to arc across. But if concerned, simply grab the nozzle with your hand, extend a finger from that same hand, and touch bare metal of the computer case to discharge any built up static.

The bigger hazard with ESD comes from using a vacuum to clean out the dust because then you have to come in close with the nozzle. Plus, you are pulling in air and dust molecules and particles which then bang into the nozzle and that is what creates the static build up. So again, wrap your hand around the nozzle and plant an extended finger onto bare metal of the case interior.

When using an air compressor, you should ALWAYS - as in EVERY SINGLE TIME - use a inline moisture and particulate filter when cleaning electronics (or air-brush painting birthday cakes or faces) so there is less worry about dust and dirt particles knocking loose free electrons.

This is excactly what Guru3D did. They normalize all parameters as you said with same cooler, same loads, curing, ambient temp, fan rpm to the max... except one. The CPU temp.

TIMs are not equal. You either end up with different CPU temp or different fan rpm or both.

Are we saying anything different?

My confusion comes from you saying they are not controlling the CPU temp. I don't understand the point of even mentioning that because the whole point of the review is to see how CPU temps vary. So of course the temps are not controlled. Why would they be?

Are we saying anything different? Yes and no. I agree that TIMs are not equal. But I am saying the difference is small and thus insignificant for the vast majority of users under the vast majority of circumstances. Again, if the TIM is applied properly, the HSF assembly is mounted properly, and case cooling is properly setup, it does not matter which TIM you are using. Why? Because if you are so close to thermal protection thresholds that you "need" the few degrees a higher rated TIM might provide, you have failed to properly address those other conditions first.

 

[Zach_01]

one just reports the CPU temperature under the same load;
but the fan might be controlled to run such that the CPU temperature remains at a given set
temperature.

Even though I said it as an example its not how things work on a cooler (CPU or GPU). The reason is that there is no software that you can set a target fixed temp and let the fan ramp up or down to catch it. At least not a known one or even consumer targeted.

Coolers and fans work with rpm/temp curves. But as I said at post #29 this diminishes the temp results of different TIM applications as fan rpm is moving on the rpm/temp curve.

Thats why reviewers are doing tests with max loads and fixed fan rpm. To show on the worst case scenario how TIMs perform. In real life where the fan is ramping up and down on a curve, with various loads, CPU temp differences are smaller except maybe on max load and aggressive curves.

So when using the same rpm/temp curve... the better performing TIMs are keeping CPU temp and fan rpm little lower on the same time.

My confusion comes from you saying they are not controlling the CPU temp. I don't understand the point of even mentioning that because the whole point of the review is to see how CPU temps vary. So of course the temps are not controlled. Why would they be?

Are we saying anything different? Yes and no. I agree that TIMs are not equal. But I am saying the difference is small and thus insignificant for the vast majority of users under the vast majority of circumstances. Again, if the TIM is applied properly, the HSF assembly is mounted properly, and case cooling is properly setup, it does not matter which TIM you are using. Why? Because if you are so close to thermal protection thresholds that you "need" the few degrees a higher rated TIM might provide, you have failed to properly address those other conditions first.

I hope the above in this very post clear things further...

 

[phanbuey]

set your fans to 100% using bios. add paste to cooler, test. look at room temp, if it's about the same then repeat with new paste.

But this is a thermal paste thread... so there will be 150+ replies and we are about to get into some serious theoretical physics and laws of thermodynamics discussions.

 

[Bill_Bright]

I hope the above in this very post clear things further...

I understand. I hope you understand me now.

My concerns comes from being a tech with a shop. As a tech, I know TIM rarely ever needs to be replaced - as long as the cured bond is not broken. So I cringe when I see posters here and other tech sites so haphazardly tell other posters to pull the heat sink and reapply TIM when they post anything that even remotely might suggest heat problems. It's akin to back in the day when "format and reinstall" was commonly and haphazardly recommended by those pretending to be experts.

As a shop owner, I have seen way too many systems come in where the users damaged their CPUs or socket or motherboard when replacing the TIM. They might bend or even break a CPU or socket pin. Zap the CPU with ESD. Fail to unplug the PSU! I mean there's no reason to dismount the CPU when replacing the TIM, but some do it anyway. Or they mess up the HSF assembly mounting. They fail to clean off the old TIM. And one of the most common is using way too much TIM which results in even higher temps than before.

So I feel it is important for users to understand they don't need the absolute best TIM available. That TIM does not need regular replacing. And there are consequences to replacing TIM if one does not know exactly what they are doing, any why they are doing it.

 

[Deleted member 191766]

I agree and that is why I find it crazy that some only last a few months.

 

[Zach_01]

I understand. I hope you understand me now.

My concerns comes from being a tech with a shop. As a tech, I know TIM rarely ever needs to be replaced - as long as the cured bond is not broken. So I cringe when I see posters here and other tech sites so haphazardly tell other posters to pull the heat sink and reapply TIM when they post anything that even remotely might suggest heat problems. It's akin to back in the day when "format and reinstall" was commonly and haphazardly recommended by those pretending to be experts.

As a shop owner, I have seen way too many systems come in where the users damaged their CPUs or socket or motherboard when replacing the TIM. They might bend or even break a CPU or socket pin. Zap the CPU with ESD. Fail to unplug the PSU! I mean there's no reason to dismount the CPU when replacing the TIM, but some do it anyway. Or they mess up the HSF assembly mounting. They fail to clean off the old TIM. And one of the most common is using way too much TIM which results in even higher temps than before.

So I feel it is important for users to understand they don't need the absolute best TIM available. That TIM does not need regular replacing. And there are consequences to replacing TIM if one does not know exactly what they are doing, any why they are doing it.

Of course I understand. I was from the beginning. I was the one that probably misled the conversation trying to explain things to OP, after all we are not talking to my language.

At post #20 when I talked about LM was just because the OP had mention it and I thought to give an insight of my experience with it. I wasnt trying to persuade anyone to use LM. Its a special compound, not made for all and mostly not for all aplications. The 2C lower temp I saw when idling and 6C when fully loaded is not that great for sure. Especially when we are talking about sub 70C. I dropped from 30 to 28 when idling and from 66 to 60 when fully loaded (ambient 22C) coming from AS5 which is not so good anyway nowdays. With a little less fan rpm also but still... Do I needed it with that 88W CPU with no OC on AIO 280mm? No... a 100 times. Will I ever use other TIM than LM in future? Probably not as long as I dont have to, because I liked what I saw despite the high cost or the overall fuss about it (handling, cleaning, replacement). The replacement was needed because of the gallium absorption by copper. This will end eventually after copper saturation with gallium. I dont mind all this and I know my way around PC parts.

By all means all other "normal" TIMs do not need frequent replacement. The past 20 years I only replace paste when I had to dismount heatsink for any reason except now with LM. The longest I kept one TIM application was 4 years with minor impact on cooling capacity, and that was also AS5.

 

[Deleted member 191766]

Of course I understand. I was from the beginning. I was the one that probably misled the conversation trying to explain things to OP, after all we are not talking to my language.

Life happens...

Be careful about 'saturation' in the case of aluminum the gallium stays free and so saturation does not occur.

For $7 I got 3oz of thermal compound and that should last (based on silicone grease and zinc oxide);
diaper rash cream is also based on zinc oxide so probably not toxic.

In an emergency use diaper rash cream!

 

[Zach_01]

Be careful about 'saturation' in the case of aluminum the gallium stays free and so saturation does not occur.

If I had aluminium surface cooler I wouldnt use LM. It is said inside package as I said before. Plus I did a small research about it before I decide to use it.

Have you tried this super lube compound?

 

[Deleted member 191766]

My point was... saturation does not happen for aluminum so might not happen for copper.

I have the tube, but have not tried it yet.

 

[Zach_01]

Oh it does... Havent you see the photos I posted? After application, copper is permanently "stained with gallium as it absorbs it because of the opposite ions (+/-) of gallium/copper). Thats why whats left behind is drying little by little and there is the need to clean the remains and re-apply (30~40 days after first application) before it dries completely. By then its too late and the copper surfface needs hard lapping.

 

[Deleted member 191766]

And you are happy with

* high cost
* stains copper to the point of needing lapping
* might short out components
* needs replacing every few months

All for a few degrees?

It's just a question; things sometimes tend to get out of hand... it's just a question.

I respect your choice.

 

[Zach_01]

1. The cost is really nothing between my yearly expences. We are talking about 15~20$ for the first year, and after that is much less.
2. If replaced on time it does not damage copper surface. It stays exactly the same with different color. How does that effect anything?
3. I was scared the first time I apply it, and so I was very carefull with it. I saw a lot of videos about it, how its done, what not to do, or must do and so on... If you are carefull its nothing
4. It is what it is and I'm not bothered by this requirement. I'm doing things like that for 20 years now, assembling or disassembling my PC(s) for cleaning of anything else. Desktops or laptops. Never damage anything.

Yes I like this kind of stuff... so its worth it for me personally. I'm above the average (set and forget) user. I like trying things within "some" reason. My reason of course.