Thermal Paste Specification Comparison

[说好不打脸的]

Nice

I also curious how it perform in your R23 now?

how is your experience apply this paste compared to other CM paste?

Hmmm, i guess still the same as last time ba, I not very sure cos i didn't keep a record last year before i upgraded to Zen 3...

I find that the Cryofuze viscosity is a bit thicker compare to Mastergel 11w/mk, not those watery type of paste anymore. But not so thick until i unable to spread it evenly with a ATM card compared to some paste i tested before. The stickiness is good, able to stick well on the cpu IHS while spreading and doesn't come off easily when reapplying few times for even surface finishing. So far pretty satisfied, just hope the paste don't degrade so fast.

Btw, how's your temp holding up after 1 month? Good?

 

[Yongkit]

Hmmm, i guess still the same as last time ba, I not very sure cos i didn't keep a record last year before i upgraded to Zen 3...

I find that the Cryofuze viscosity is a bit thicker compare to Mastergel 11w/mk, not those watery type of paste anymore. But not so thick until i unable to spread it evenly with a ATM card compared to some paste i tested before. The stickiness is good, able to stick well on the cpu IHS while spreading and doesn't come off easily when reapplying few times for even surface finishing. So far pretty satisfied, just hope the paste don't degrade so fast.

Btw, how's your temp holding up after 1 month? Good?

many thanks for sharing

May i know the temp running cinebench same as previously you mentioned it is around 85c?

I guess only time can tell the different in long term usage.

I try to find the best data i can before switching to zen3..

Below was the test of the timespy with the cryofuze paste comparison between 18-Jan (76.38c) compared to 29-Jan (74.38c) both temp is very similarly perform at about 2c improvement for margin of error at later time perhaps the effect of paste is getting more effective over time:

18-Jan - Timespy CPU temp (76.38c)




29-Jan - Timespy CPU temp (74.38c)



in longer run maybe i can have better update on the heat performance for 5900x too.

 

[Lastwishes]

Anyway, what thermal paste do you guys recommend? Thinking to get a new tube. Is MX5 any good? KPX? Kryonaut Extreme?

Kpx and mx4 you're gonna have the same issue over time. Go for TFX one short and you don't need to keep opening up your card to change the paste. Been through that and it's not fun.

Echo what elmariachi said on TFX. I've opened my card up more than three times. Tried MX-4, NT-H1 and temps increased after a 2-3 months each time, with widening hotspot deltas (15c increased to about 20c).

Switched to TFX - initial results were the best of the lot, and there's still no significant deterioration until now (I think it's been about 6 months). Only 0.1c increase in hotspot delta after 6 months, which is very impressive.

 

[说好不打脸的]

many thanks for sharing

May i know the temp running cinebench same as previously you mentioned it is around 85c?

I guess only time can tell the different in long term usage.

I try to find the best data i can before switching to zen3..

Below was the test of the timespy with the cryofuze paste comparison between 18-Jan (76.38c) compared to 29-Jan (74.38c) both temp is very similarly perform at about 2c improvement for margin of error at later time perhaps the effect of paste is getting more effective over time:

18-Jan - Timespy CPU temp (76.38c)

View attachment 30584


29-Jan - Timespy CPU temp (74.38c)

View attachment 30585

in longer run maybe i can have better update on the heat performance for 5900x too.

Not the same under different environment, the current rebuilt 3800x stressing R23 was done under aircon 26-27℃ ambient at my workplace, end up slightly lower temp around 77℃ max this afternoon compare to earlier OCCT max temp in the morning, due to OCCT pull slightly more power under Extreme mode than R23 stressing.
The 85℃ in R23 i mentioned previously was paired with my current X570 Aorus Master and with the same D15 HS in the past at home, the stress test was done in the hot afternoon ambient of 32℃ during 2020 mid July.
But if we are to offset the ambient temp of 5℃ between both environment, plus higher ambient are less efficient in cooling the CPU compare to lower ambient (1-2℃ different). The end result of the current rebuilt 3800x will be around 77 + 5 + (1-2) = 83-84℃ under 32℃ ambient.

Thanks for sharing too, did not know that cryofuze need some curing time..

 

[it_geek]

Excellent thread.

Can attest to the breakdown of the Kryonaut on loads above 80 degrees. Used it for my older mobile workstation which always runs above 80 degrees.

The peak temps creeped back to 96 within a few days

I also realised how crappy the temp delta is with the Kryonaut on my i5-6500k. From 37 degrees can shoot to 67 on peak load now, that's a whopping 30 degree delta. True enough, initially the delta was around 24 degrees (37 - 61) when i first applied it... widening the delta after a few weeks really attests to the kryonaut not living up to its expectations.

Want to ask for advice from the bros for paste that has long term stability for machines that consistently run above 80 degrees.

Quite a few recommendations so I find it quite hard to pick.


Thermalright TFX as recommended by @elmariachi seems to be a decent long term paste, but can it still hold up on machines that consistently push beyond 80?

Also wonder how it matches up to the SYY-157 and the Kingpin KPX.

 

[elmariachi]

Excellent thread.

Can attest to the breakdown of the Kryonaut on loads above 80 degrees. Used it for my older mobile workstation which always runs above 80 degrees.

The peak temps creeped back to 96 within a few days

I also realised how crappy the temp delta is with the Kryonaut on my i5-6500k. From 37 degrees can shoot to 67 on peak load now, that's a whopping 30 degree delta. True enough, initially the delta was around 24 degrees (37 - 61) when i first applied it... widening the delta after a few weeks really attests to the kryonaut not living up to its expectations.

Want to ask for advice from the bros for paste that has long term stability for machines that consistently run above 80 degrees.

Quite a few recommendations so I find it quite hard to pick.


Thermalright TFX as recommended by @elmariachi seems to be a decent long term paste, but can it still hold up on machines that consistently push beyond 80?

Also wonder how it matches up to the SYY-157 and the Kingpin KPX.

Kpx has a chance of breakdown though not as fast as the rest. Syy-157 is probably a 10% lesser version of the TFX performance wise but much cheaper. All these thick pastes are like old-school pastes and requires a curing time. Problem is you need to keep it below 70 degrees for about a week before you get full thermal optimization. This is anywhere from 1 to 2 degrees.

I am not sure if straight away whack past 80 degrees the paste will degrade or not cause my system is well below 70 degrees and I usually run everything stock before overclocking to ensure everything is settled down. But I still think the TFX will be alot more consistent than what you're using now without a widening delta over time. Only way is to try one one workstation first and see for about 2 weeks. If works wonders, whack everything TFX.

 

[Lastwishes]

I've used TFX on both my GPU and CPU. I was aware of comments that there was a need to cure, but I didn't bother. Just used the system as per normal. Pretty sure I got both the CPU and GPU over 70 during the first week, but it never got over 80. No noticeable deterioration, especially for the GPU.

I find detecting deterioration for the CPU a bit trickier, since the PBO2 on the 5000 series will just push the chip to the max (90c) if you let it. I keep mine at set ppt/edc/tdc values, and haven't really seen much difference in the max temps under air-conditioned environments; on really hot days where I don't turn on the air-con, it can reach over 85 on one day and stay below 76 on the next when it's a cooler day.

 

[Yongkit]

I've used TFX on both my GPU and CPU. I was aware of comments that there was a need to cure, but I didn't bother. Just used the system as per normal. Pretty sure I got both the CPU and GPU over 70 during the first week, but it never got over 80. No noticeable deterioration, especially for the GPU.

I find detecting deterioration for the CPU a bit trickier, since the PBO2 on the 5000 series will just push the chip to the max (90c) if you let it. I keep mine at set ppt/edc/tdc values, and haven't really seen much difference in the max temps under air-conditioned environments; on really hot days where I don't turn on the air-con, it can reach over 85 on one day and stay below 76 on the next when it's a cooler day.

Agreed.

It is a reason I think logically I will test the initial temp with all default setting without PBO.

future next 6 months I will just enabled PBO and play around.

Coming Jun or Jul I intent to test the temp again with all reset to default to check what is the temperature like.

I am not sure if AIO will be different from air cooling fan in this manner?

1 thing for sure I will keep all the fans free of dust before next testing as dusty fans will degrade the cooling performance even it is running at the same rpm as before, just like household stand fan dusty will have less cooling air flow.

 

[kimsix]

kyronaut so bad? oh dear thought was so hyped. many watercooling comes with kyronaut

mine cpu about 9 months, so far not obvious bad peaks even though i got hit with stress tests.

 

[elmariachi]

The Kryonaut issue is mainly with GPUs. It dries up easily causing a degradation in your hotspot temps. CPUs not so much but it does gradually increase but at a much slower rate mainly because the IHS is more perfect than the direct die of the RTX 3000 series which are convex causing the thermal paste on the sides to dry up over time (weeks). TFX solves this problem because the paste is much thicker and it becomes like clay once cooled down. Upon heat it becomes semi-soft and the cycle is like that. Only downside is application of the paste is arse.

kyronaut so bad? oh dear thought was so hyped. many watercooling comes with kyronaut

mine cpu about 9 months, so far not obvious bad peaks even though i got hit with stress tests.

 

[Lastwishes]

Agreed.

It is a reason I think logically I will test the initial temp with all default setting without PBO.

future next 6 months I will just enabled PBO and play around.

Coming Jun or Jul I intent to test the temp again with all reset to default to check what is the temperature like.

I am not sure if AIO will be different from air cooling fan in this manner?

1 thing for sure I will keep all the fans free of dust before next testing as dusty fans will degrade the cooling performance even it is running at the same rpm as before, just like household stand fan dusty will have less cooling air flow.

I think your testing methodology is fine, but I'd say that ambient temperature will be a significant factor (air-conditioning may help to keep the conditions more consistent; but imo air-con on a rainy day vs one on a sunny day still can make a difference in most cases).

In the end it also depends on what you're testing - are you looking at the performance of the thermal paste, or whether performance degrades over time?

 

[Yongkit]

I think your testing methodology is fine, but I'd say that ambient temperature will be a significant factor (air-conditioning may help to keep the conditions more consistent; but imo air-con on a rainy day vs one on a sunny day still can make a difference in most cases).

In the end it also depends on what you're testing - are you looking at the performance of the thermal paste, or whether performance degrades over time?

True as aircon room will definitely affect the performance of the cpu cooling however nowadays I seldom on aircon unless the room temp reaching 31c.

Most of time my room ambient within 28c~30c which is still alright to me as I only need a stand fan for cooling myself.

I wish to test all factors in consideration of the thermal paste and also the CPU performance degradation over time.

 

[Lastwishes]

I wish to test all factors in consideration of the thermal paste and also the CPU performance degradation over time.

Can be tricky. If testing for CPU performance with PBO2 enabled, you would have to think through quite a bit on the way you run your tests and benchmarks for comparison.

Keeping CPU settings consistent would be good for testing for thermal paste performance, but some stuff to consider:

1. The difference between a good thermal paste and an average thermal paste can just be a few degrees.
2. Ambient temperature will be a significant factor from day to day, with no air-con.
3. The thermal paste degradation/break down that was being discussed in the thread can be anything from 5 degree to 15 degree, which also depends on what was the baseline.

For me I just look mostly at my GPU temps, where my card is capped to 850mv/1875mhz running 3dMark Timespy, and also running various games that I monitor occasionally as I play them over several months.

 

[Yongkit]

Can be tricky. If testing for CPU performance with PBO2 enabled, you would have to think through quite a bit on the way you run your tests and benchmarks for comparison.

Keeping CPU settings consistent would be good for testing for thermal paste performance, but some stuff to consider:

1. The difference between a good thermal paste and an average thermal paste can just be a few degrees.
2. Ambient temperature will be a significant factor from day to day, with no air-con.
3. The thermal paste degradation/break down that was being discussed in the thread can be anything from 5 degree to 15 degree, which also depends on what was the baseline.

For me I just look mostly at my GPU temps, where my card is capped to 850mv/1875mhz running 3dMark Timespy, and also running various games that I monitor occasionally as I play them over several months.

many thanks and will take note on all these points during testing.

 

[说好不打脸的]

kyronaut so bad? oh dear thought was so hyped. many watercooling comes with kyronaut

mine cpu about 9 months, so far not obvious bad peaks even though i got hit with stress tests.

My 1st encounter on Kyronaut got pumped out was on a mild overclocked 9900k with D15, D15 unable to cool itself off fast enough. After 2 months, CPU throttle until crazy due to overheating.
But if your temp is still good, no harm leave it and continue to use, don't waste it. Cos most prob is that your watercooling system is good and fast enough to draw away the heat from the paste from being cook between the CPU IHS and coldplate. But if you want to change to better durable paste like many bros shared here also can, it may give you a better peace of mind down the road. Currently i only start using the new CM Cryofuze, only time will tell me how the durability of the paste, but i still think that TFX is really a good paste being shared alot..

 

[kimsix]

My 1st encounter on Kyronaut got pumped out was on a mild overclocked 9900k with D15, D15 unable to cool itself off fast enough. After 2 months, CPU throttle until crazy due to overheating.
But if your temp is still good, no harm leave it and continue to use, don't waste it. Cos most prob is that your watercooling system is good and fast enough to draw away the heat from the paste from being cook between the CPU IHS and coldplate. But if you want to change to better durable paste like many bros shared here also can, it may give you a better peace of mind down the road. Currently i only start using the new CM Cryofuze, only time will tell me how the durability of the paste, but i still think that TFX is really a good paste being shared alot..

best is dont open hwinfo or gpuz.

i like TR items, low key but high performance, TFX didnt get hyped as kyronaut and we missed the boat

my TR hsf, this old red color version is powerful pwm and good noise signature, half price of noctua

 

[nexusscion]

After using so many diff type of thermal paste, I just stick to MX5 and done with it. Cheap and good and lasting. I dont need to keep changing the paste or scare dry up too fast.

 

[sglords]

ytd swapped out thermalright tf7(stock from aio) to thermalright tf9.
totally pia to apply using scraper and used more thermal paste than intended.
wiped away and reapplied with a X. done.

X method is the best out of all application. require zero effort and absolute risk free.

 

[watzup_ken]

The Kryonaut issue is mainly with GPUs. It dries up easily causing a degradation in your hotspot temps. CPUs not so much but it does gradually increase but at a much slower rate mainly because the IHS is more perfect than the direct die of the RTX 3000 series which are convex causing the thermal paste on the sides to dry up over time (weeks). TFX solves this problem because the paste is much thicker and it becomes like clay once cooled down. Upon heat it becomes semi-soft and the cycle is like that. Only downside is application of the paste is arse.

I feel the issue is not just limited to GPUs. GPUs tend to stick out more because they tend to run very hot. So likewise if the CPU is very hot, i.e. 80+ degrees for prolong period of time, the paste will breakdown/ dry up faster. So this may also apply to repasting for laptop GPU and CPU as they usually exceed 80 to 90 degrees easily.

 

[elmariachi]

I feel the issue is not just limited to GPUs. GPUs tend to stick out more because they tend to run very hot. So likewise if the CPU is very hot, i.e. 80+ degrees for prolong period of time, the paste will breakdown/ dry up faster. So this may also apply to repasting for laptop GPU and CPU as they usually exceed 80 to 90 degrees easily.

Breakdown and Drying up are entirely 2 different things. Breakdown I agree pass 80 or 90 most pastes breakdown over prolonged time.

My GPUs never reached even 50 degrees when they dried up. But hotspot increased tremendously in months. I can be like 50 degrees core but 75 degree hotspot. I was only 42 degrees core and and 57-58 degrees hotspot when I had it freshly installed.

So the drying up is mainly due to the convex die and running thinner pastes. Temps can be awesome but you're still drying up but not breaking down. I've already tested this so many times to tell you there is a difference running paste direct die and on an IHS. Direct die is alot trickier for long term stability despite having no breakdown even if you're well below breakdown temps.