2023/12/13 17:32:13
SM-71
So, not too long ago, my CPU started showing signs that the thermal paste may have dried out too much.   Seems odd since I'm running IC Diamond thermal paste which has never had a problem for me in the past.  But considering just how HOT the 13900KS runs, I figured that was probably the difference.
 
This is what it looked like when I removed the waterblock:

The colored area on the CPU seems to be permanently discolored.
 
Around the same time, I stumbled upon a review from Jays2cents, a review of KryoSheet Graphene thermal interface sheet from Thermal Grizzly.  In his review it looked like the KryoSheet was doing just as well as the thermal paste he had used on his test-bed.  I didn't realize how cold his test room was (19.6C = 67F) till after I had already tried out the KryoSheet....   
 
BTW, KryoSheet (graphene sheet) should not be confused with carbon/graphite pads.   Although Graphene comes from graphite it has significantly better thermal properties, in fact it has the highest thermal conductivity of any know natural or man made product:
Graphene = up to 4000W/mK
lab grown diamond = up to 3300W/mK
natural Diamond = up to 2200W/mK
Pure graphite = 25 to 470W/mK
.9999 pure silver = 429W/mK
.9999 pure copper = 398W/mK
 
HOWEVER.... Graphene only has that great conductivity when it is in a layer just a single atom thick   It would easily tear in such a thin layer & would basically be useless.  As you stack layers together to add strength, you also loose thermal conductivity.   This is something I did not know when I purchased the KryoSheet.  If you put together too many layers you can drop the thermal conductivity to as low as 6W/mK.  This would be a pad about as thick as a heavy weight business card (300GSM or about 0.43mm), which would be way to thick for an effective thermal interface material.  Thermal Grizzly claims the sheets are 0.2mm thick.  I'll measure when I replace it with more IC Diamond paste, but seems like it's thinner than 0.2mm. 
 
I don't have scientific instruments to test with, but using the IC Diamond thermal paste, my 13900KS in summer time heat with room temps around 78F, it was able to run a 10 minute Cinebench R23 run with a score of 41,697 at stock CPU settings.  That was with my Noctua Industrial PPM-3000 RPM high static pressure fans running at 100% speed on my Kraken X63 AIO radiator.

 
After putting the KryoSheet in place, and with those same fans running 100% on a freshly cleaned radiator, I cannot even finish a 10 minute run with the CPU clocked down to 5400MHz P-cores / 4000MHz E-cores.  It will finish a 30 minute stability run at 5000MHz P-core / 4000MHz E-core.   But even with clocks set that low, and with outside temps barely reaching 70F today, I'm still seeing temps hit 100C on 6 of the P-cores, and that is with a room temp of 71F after the 30 minute stability run. 
 
Looking at Hardware Monitor Pro while it's running, and with P-cores set to 5200MHz, I'm seeing it throttle down to 5000MHz about 90% of the time, occasionally dropping to 4900MHz on the P-cores.   I have to drop P-core to 4800MHz to keep temps below 100C with the KryoSheet in place.  I repositioned the water block twice to make sure the KryoSheet was in place correctly BTW, no change to results.   
 
 
It was running a 30 minute stability test at 5600MHz P-core / 4300MHz E-core with no problems while the old application of IC Diamond paste was in place, and would only see 100C on 1 or 2 cores during that run.  
 
Fortunately, I bought a 7 Karat tube of IC Diamond when I rebuilt this PC, so I have plenty to use!   
 
So basically, if you're running very cheap white thermal paste, KryoSheet is probably a better option because it is reusable if you're careful with it, and may even offer better temps....   But if you're already running a high performance thermal paste, skip the KryoSheet, and stay with the thermal paste, or switch to liquid metal.
2023/12/14 09:11:46
Cool GTX
Thanks for sharing your results with Thermal Grizzly KryoSheet vs IC Diamond thermal paste
 
It will be interesting to see if your original performance returns, once you switch back to IC Diamond thermal paste
2023/12/14 12:44:56
SM-71
Cool GTX
Thanks for sharing your results with Thermal Grizzly KryoSheet vs IC Diamond thermal paste
 
It will be interesting to see if your original performance returns, once you switch back to IC Diamond thermal paste


Indeed....  I've had pretty consistent results with IC Diamond, which is why I keep going back to it.  Just wish it was easier to apply.  Will probably be a few weeks before I swap, not doing anything in the short term that needs full performance.  
2024/01/28 20:41:15
ryouiki
I haven't used it in years (4790K era), but IC Diamond paste has/had? a bad habit of etching the CPU heatspreader/cooler cold plate (especially if nickle plated) to the point where the manufacturers would reject warranty claims.
 
I'd be curious if you lapped the cpu/cooler if the results would have been more in line with other results I've seen... I plan to test Kyrosheet in the near future on a 5950X (that I hate taking apart for yearly paste replacements) to see how it performs.
2024/08/20 17:58:18
SoDebonaire
After using the same Kryosheet on at least 15-20 different cpu mounts with multiple 13KS's and multiple 14KS's, consistently benching 42K+ points on R23 and also using a Kryosheet on my ASUS 4090 STRIX over 1 year+, that seems weird.  It was the best $20 I had spent in a long while.  No mess to clean, it never dried out, it's always perfectly flat and temps were equivalent or better than the Kryonaut, NT-H2, and PTM7950, I used prior.  Honestly don't see myself ever buying thermal paste again.
 
Coolers used with Kryosheet:
Corsair H150i Elite LCD 360mm AIO
Corsair H170i Elite LCD 420mm AIO
EK Direct Die 360mm AIO (Yes you can use the EK Direct Die AIO with stock cpu's. I posted a thread about it on reddit, where I show great results using a non-delidded 14KS with a Kryosheet.)
 

2024/08/21 04:17:17
rjohnson11
Thanks for the input.
 
2024/09/04 20:03:41
donta1979
I would just use an authentic Honeywell PTM7950 pad, first day it sucks, second day its a lot better, third day yup. If you want to cure it some more you got to let your hardware heat up to at least 80c to cure/bake some more. Never degrades, just gets better with time.
 

Use My Existing Forum Account

Use My Social Media Account