Thermal Paste Specification Comparison

[elmariachi]

Used Kryonaut around 7700K period and AQC Cryos Cuplex NEXT CPU Block and I pointed out this very low viscosity pastes breaks down quick especially when there is high thermal density as well. Looks like this hasn't changed from launch till now. Likewise for a Zotac 1070 Mini the thermals raised by around 10C over a short span of use (less than 6 months).

The varying measurements on TIM spec isn't a tell all especially with no long term testing. Especially in the age of internet influencers and prettied up specs to look good on paper.

Most popular reviewers do not do a long term review of pastes. Any fresh install of paste for reviewing is almost certainly to do well. I swapped pastes 3 times on my 3090 then to my 3080 Ti Asus Strix where I used Thermalright TFX. Within 2 or 3 weeks, the other pastes will start creeping up and the hotspot delta will widen. TFX until today for my 3080 Ti is 6.1-7 degrees hotspot depending on ambient variations. Kryonaut was increasing like 2-3 degrees every 2 weeks or so untill it slowly widened up to 20 degrees delta. Clearly, drying up and for those watercool guys down here, it is not fun draining the loop to keep doing this. You need long term stability for peace of mind.

 

[watzup_ken]

Most popular reviewers do not do a long term review of pastes. Any fresh install of paste for reviewing is almost certainly to do well. I swapped pastes 3 times on my 3090 then to my 3080 Ti Asus Strix where I used Thermalright TFX. Within 2 or 3 weeks, the other pastes will start creeping up and the hotspot delta will widen. TFX until today for my 3080 Ti is 6.1-7 degrees hotspot depending on ambient variations. Kryonaut was increasing like 2-3 degrees every 2 weeks or so untill it slowly widened up to 20 degrees delta. Clearly, drying up and for those watercool guys down here, it is not fun draining the loop to keep doing this. You need long term stability for peace of mind.

That is the limitation of reviews. They generally test it once, and never check back again. After all, it is not uncommon for test rigs to get parts swapped out very often. Only after using it ourselves will we figure out the pros and cons over time. But thank you for sharing. I got the SYY previously because you mentioned it, so I went to try it as an alternative to TFX. Cheap and good.

 

[elmariachi]

That is the limitation of reviews. They generally test it once, and never check back again. After all, it is not uncommon for test rigs to get parts swapped out very often. Only after using it ourselves will we figure out the pros and cons over time. But thank you for sharing. I got the SYY previously because you mentioned it, so I went to try it as an alternative to TFX. Cheap and good.

Awesome! SYY157 is slightly easier to spread but I still would heat up the paste on a metal spoon or small plate to get it to spread nicely. Temp difference between TFX and SYY157 is like 1-1.5 degrees only.

 

[watzup_ken]

Awesome! SYY157 is slightly easier to spread but I still would heat up the paste on a metal spoon or small plate to get it to spread nicely. Temp difference between TFX and SYY157 is like 1-1.5 degrees only.

I can live with that difference. As long as it is more durable and consistent with its thermal conductivity, I am good with it. So far, this paste is still keeping my Mini PC cool and quiet after a few months of usage. To me, this is a good test because the paste is applied on the bare die, and the temps can get quite high under load, i.e. 90s, with the stock thermal compound. Switching to MX4 dropped the temps to high 80s, and using SYY dropped it to close to mid 80s.

 

[elmariachi]

I can live with that difference. As long as it is more durable and consistent with its thermal conductivity, I am good with it. So far, this paste is still keeping my Mini PC cool and quiet after a few months of usage. To me, this is a good test because the paste is applied on the bare die, and the temps can get quite high under load, i.e. 90s, with the stock thermal compound. Switching to MX4 dropped the temps to high 80s, and using SYY dropped it to close to mid 80s.

From my experience, the breakdown of pastes are more common with direct die. IHS usually won't have much of a temp degradation based on what I experienced. Exception being Kryonaut which does break down past 80 degrees and performance of paste is not consistent after that.

For ease of mind, I just whack TFX for both CPU and GPU as I could have applied maybe KPX for my CPU for slightly better temps but I decided to just do TFX all around and my CPU has no issues with 5.2-5.3Ghz in games with 5.5Ghz single core boost with spike temps barely hitting 75 degrees. Average is mid-high 60s only.

 

[elmariachi]

My current temps after a month. Will record a video with my new CPU overclock (5.2Ghz-5.4Ghz in game depending on temps) so you all can see the performance of TFX. This is going to be slightly more than a month since I assembled my system. GPU temps and Hotspot remain unchanged till today.

 

[ragnarok95]

elmariachi, you say heat it up first, with a hair dryer first before spreading?

 

[elmariachi]

elmariachi, you say heat it up first, with a hair dryer first before spreading?

Yes. To make things easier, use a metal spoon. Squeeze about 1cm out onto the spoon then use a hairdryer over the paste to heat it up. The metal spoon will accelerate the heating up plus the paste will also be heated from below. Once it starts to looks soft, quickly use a spatula and stroke 1-2 times in 1 direction. Try to use a hairdryer on the lowest heat setting optimal would be around 65 degrees centigrade.

Repeat again untill you are able to complete the die. Do not worry too much about the edges if it's not much there. Once you mount the cooler down, the paste will be pushed out towards the edges. Hope this helps.

 

[ragnarok95]

Alright thanks! So the bird sai method doesn't work for this paste?

 

[elmariachi]

Alright thanks! So the bird sai method doesn't work for this paste?

It does work for IHS but Die better to even spread. But I just also spread even for IHS now given how much bigger the Die's have become today with everyone focussing on hybrid designs. So my preference is both nice and even spread for BOTH CPU and GPU.

 

[ragnarok95]

Okie thanks for the tips!

 

[elmariachi]

Gaming Session Last Night

Took this screenshot right 5 mins after gaming to show the min/max values. CPU hit peak 74 but if you look average in game is around 60s only. Both CPU and GPU pasted with TFX (Nice and Even Spread) more than a month ago with no variance in degradation on either CPU or GPU.

 

[watzup_ken]

Alright thanks! So the bird sai method doesn't work for this paste?

If the die is big, that drop in the middle method is not very effective. You can observe that when you remove the cooler and only the center is covered in the thermal compound. This is especially so if you have a very thick kind of thermal compound.
I don't use the heating method, but I will just waste more thermal compound in the process of trying to spread it as evenly as possible.

 

[elmariachi]

If the die is big, that drop in the middle method is not very effective. You can observe that when you remove the cooler and only the center is covered in the thermal compound. This is especially so if you have a very thick kind of thermal compound.
I don't use the heating method, but I will just waste more thermal compound in the process of trying to spread it as evenly as possible.

The heating effort is more effective because I find that it bonds to the surface better. Try spreading it as it is, and when you spread over again, it picks up the paste and that whole area is empty. When it's heated up, I find that the paste settles better into the micro pores of the die. This is evident when spreading.

Of course I can't prove that my method yields better results thermally. It's just my observation when spreading the paste. Effectiveness is in terms of spreading the paste and it settling onto the die.

 

[ragnarok95]

My CPU is not the 12000 series should be okie. I usually do 5 bird sai drop on my CPU, one in the middle and 4 sides.

GPU however i went for spread.

 

[elmariachi]

My CPU is not the 12000 series should be okie. I usually do 5 bird sai drop on my CPU, one in the middle and 4 sides.

GPU however i went for spread.

X pattern with 4 dots is much better if you wanna do it this way. The spread is better with no empty spots when spreading under the pressure of the cooler. But if thermals good then leave it. That's the purpose anyway.

 

[watzup_ken]

The heating effort is more effective because I find that it bonds to the surface better. Try spreading it as it is, and when you spread over again, it picks up the paste and that whole area is empty. When it's heated up, I find that the paste settles better into the micro pores of the die. This is evident when spreading.

Of course I can't prove that my method yields better results thermally. It's just my observation when spreading the paste. Effectiveness is in terms of spreading the paste and it settling onto the die.

Yeah, this is the problem with thick paste. It takes me like 3 or 4x the amount of time to apply the SYY and previously the Thermalright paste, than say MX4. As to the paste settling better, I feel if the mounting pressure is good, it should force the thermal paste into any uneven surface. Just my personal opinion.

 

[elmariachi]

Yeah, this is the problem with thick paste. It takes me like 3 or 4x the amount of time to apply the SYY and previously the Thermalright paste, than say MX4. As to the paste settling better, I feel if the mounting pressure is good, it should force the thermal paste into any uneven surface. Just my personal opinion.

Yeah thick pastes is like that. But I guess application methods are all personal preference. I prefer using the paste heated up during application its easier to be consistent with the spread though eventually this is a moot point because the cooler will spread the paste under pressure. Guess old habits die hard.

 

[ragnarok95]

Yes X pattern with 5 drops.

 

[Yongkit]

I felt more content and more comfortable using elmariachi method of spreading paste.

For example for cpu purposes, I find those preapplied thermal paste that come with aio like the ek aio basic was quite thin and when I lift up can easily see the CPU label like below (temperature also quite similar to my previous air cooler performance _idle temp mid 45c)



However if manually applied the paste the spread is better and can cover all the CPU wording like below (temperature performance are much better than air cooler, idle temp low 40c)