Overclock.net banner

Graphene & paste + 3770K De-lidded

3K views 23 replies 7 participants last post by  HitMe 
#1 ·
previously on graphene cooling Experiment with and shenanigans .... i did try to coat my waterblock with conductive graphene ink .
which gave me 1-5 °c lower temps ... i can be sure because i did not take any readings before...
since some have asked me to do this :
Quote:
Originally Posted by Torvi View Post

very nice score
smile.gif
Also have you considered deliding and using graphene ?

what about crafting solid graphene his also?
without any other intro let me give you the "résultat"

ok To make things clear this time i did take all temps before i change anything with fans under High
Cpu idle and load " prime95 "



ROOM TEMP (sensor N°5 set outside the case in front of 200mm fan ):


ok now 3770K De-lidding





then i did put the conductive graphene ink on the chip die
it was drying very fast i had to put the cpu ihs without taking pics

the thing is i left the graphene bottle open for months and the N-Butyl acetate dry out ...



i glued back the ihs with super glue and placed the cpu back and the waterblock

checked for any leak ... and stuff and i Start Up the Computer.

and i was full of regretted feelings ... .....
cpu temps did hit the 90°c i had to reboot to check again .. and later its 81 here's a screen :


i was totaly helpless and i did not had any TIM to replace the graphene

i decided to remove the ihs and check again and it all become clear

the ink was sooooo dry it did this :




so now i have to find a solution and all i have is unknown silicon paste ... i ordered 2 years ago from china

some thing like this :



so i decided to mix it with the ink and see what would happen :
i did lots of mixing and 5 drops to 1 drop was the most fluid


i put back everything back , had a little leak but is ok ...and testing :

MAX TEMP 51 °c (before : 63) :

this is under auto fan mod



this is under High Fan it bounce in 45-48...room temp 30-32 :



15-20 above room temp ... i dont know if thats good but please feel free to give your opinion
 
See less See more
12
#5 ·
Wait a second.. you measured non-delidded temperatures of your
cpu and then measured the self made TIM?

If so > well of course it is much better. Since intel is not soldering the IHS.
Some People even see a 15-20°C difference when replacing the original TIM
from intels with some good aftermarket ones.

The interesting part now would be to compare the graphene to other TIM pastes
and see if there is some magic
wink.gif


Thank ya for yo efforts.
 
#6 ·
Quote:
Originally Posted by smex View Post

The interesting part now would be to compare the graphene to other TIM pastes
and see if there is some magic
wink.gif
I'm interested in how it performs relative to other pastes as well. Assuming results are close to the relative cooling increase you had gotten from your other"graphene paste" results it'd be likely graphene would be better than just normal paste. It would still be great to have concrete numbers.

Recently I decided to grind up some of dat "useless" pyrolytic graphite that I had left over and mix it with paste just to see what happens. I mixed about 2 parts silicon paste to 1 part HOPG. I didn't really have a good grinder so the best I could grind it was into reasonably large flakes.

To my surprise when I applied it between cpu and cpuheatsink it.. actually worked well. Especially when one would compare this with the results I got from previous experiments as cpu-heatsink TIM (almost instant 100 degrees C cooking cpu).

Results

Original silicon paste

ambient temperature 28 C.

idle- 47 to 47.5 fanspeed- 13.7%
load- 66.5 fanspeed- 30%

Silicon + ground flaky HOPG paste

ambient 27 - 28 C

idle- 45 to 46 fanspeed- 9.8%
load- 64 fanspeed- 30%

So not only did my CPU not combust this time, it was also several degrees cooler than with the default thermal paste. The difference in temperature is too low for me however as one should take human error in account as well, so one test is not enough to conclude the difference is truely from the ground hopg. Still, at least my cpu didn't combust this time so there's improvement.

I've found a finer grinder with which I shall conduct the next experiment attempt.
 
#7 ·
Quote:
Originally Posted by ThijsH View Post

I'm interested in how it performs relative to other pastes as well. Assuming results are close to the relative cooling increase you had gotten from your other"graphene paste" results it'd be likely graphene would be better than just normal paste. It would still be great to have concrete numbers.

Recently I decided to grind up some of dat "useless" pyrolytic graphite that I had left over and mix it with paste just to see what happens. I mixed about 2 parts silicon paste to 1 part HOPG. I didn't really have a good grinder so the best I could grind it was into reasonably large flakes.

To my surprise when I applied it between cpu and cpuheatsink it.. actually worked well. Especially when one would compare this with the results I got from previous experiments as cpu-heatsink TIM (almost instant 100 degrees C cooking cpu).

Results

Original silicon paste

ambient temperature 28 C.

idle- 47 to 47.5 fanspeed- 13.7%
load- 66.5 fanspeed- 30%

Silicon + ground flaky HOPG paste

ambient 27 - 28 C

idle- 45 to 46 fanspeed- 9.8%
load- 64 fanspeed- 30%

So not only did my CPU not combust this time, it was also several degrees cooler than with the default thermal paste. The difference in temperature is too low for me however as one should take human error in account as well, so one test is not enough to conclude the difference is truely from the ground hopg. Still, at least my cpu didn't combust this time so there's improvement.

I've found a finer grinder with which I shall conduct the next experiment attempt.
you should not grind up the pyrolytic graphite , the core of its thermal conductivity its in a multi layers stacked together
you just lade it 70nm or more

friend try to make a heatsink

i have draw someting for you :

try to cut a smaller sheets and place them one next to the other until you cover the cpu
then put your heat sink
and if you can try to glue some RAM heatsinks on the extra sheets i'm sure it will be cooler
because as long as the sheet on the same size ans still layered together this will work
 
#9 ·
Quote:
Originally Posted by ThijsH View Post

Ooh use supplimentary strips around but not on the core of the cpu sort of like extra heatpipes. Didnt think of this, I like the idea. This is gonna be a pain to try as these sheets are hard to work with xD.
i'm sure you can do it and dont forget to add some ram heat sink on the tip of each sheet or glue it with TIM paste to the cpu cooler if air
 
#10 ·
So my heatsink block turned out to be alot bigger than the cpu so I made supplimentary heatsink to the sides of the heatsink instead directly on the cpu.





I left the heatsink+cpu with TIM still on it to open air for hours, couldnt be arsed to change it after I was done so you could probably shave 1-2 degrees off the new results.

Original silicon paste

ambient 28 C.
idle- 47 to 47.5 fanspeed- 13.7%
load- 66.5 fanspeed- 30%

Silicon + ground flaky HOPG paste

ambient 27 - 28 C
idle- 45 to 46 fan:9.8%
load- 64 fan:30%

HOPG used as supplimentary heatsink (+old tim)

ambient 29 C
idle 46-42 fan:10-14%
Load 66-66.5 fan:30%

test #2
ambient 35-36 --Dat hot day
idle 52-48 fan:30%
load 74 fan:30%

No spectacular changes, maybe one, two degrees difference. Still; if you're going for absolute lowest temps this could be something to try out for yourself.
 
#15 ·
Quote:
Originally Posted by ThijsH View Post

So my heatsink block turned out to be alot bigger than the cpu so I made supplimentary heatsink to the sides of the heatsink instead directly on the cpu.





I left the heatsink+cpu with TIM still on it to open air for hours, couldnt be arsed to change it after I was done so you could probably shave 1-2 degrees off the new results.

Original silicon paste

ambient 28 C.
idle- 47 to 47.5 fanspeed- 13.7%
load- 66.5 fanspeed- 30%

Silicon + ground flaky HOPG paste

ambient 27 - 28 C
idle- 45 to 46 fan:9.8%
load- 64 fan:30%

HOPG used as supplimentary heatsink (+old tim)

ambient 29 C
idle 46-42 fan:10-14%
Load 66-66.5 fan:30%

test #2
ambient 35-36 --Dat hot day
idle 52-48 fan:30%
load 74 fan:30%

No spectacular changes, maybe one, two degrees difference. Still; if you're going for absolute lowest temps this could be something to try out for yourself.
good job maybe i will try a water block with PG sheets as micro channels .. .

that will need some cnc machining....
 
#16 ·
Do it for science! I'm looking forward to your data on graphene as heatsink (if you decide to experiment).

Also please do mind that I tried this on an aluminium base air cooler. I wouldn't know know exactly what difference this makes for higher grade (water) coolers. We need more information through experimentation.
 
#17 ·
I'm not sure if this is what you were intending for, but graphene-based thermal paste does exist and it's one of the best thermal paste bases out there... likely better than CLU judging by their comparisons to AS5/MX-4. http://thermene.com

It's only $14.99 USD on amazon.com. If I weren't in a rush to get my build together again I would order some to my friend's mailbox in the States and grab a couple tubes
wink.gif
.
 
#18 ·
Quote:
Originally Posted by Diablosbud View Post

I'm not sure if this is what you were intending for, but graphene-based thermal paste does exist and it's one of the best thermal paste bases out there... likely better than CLU judging by their comparisons to AS5/MX-4. http://thermene.com

It's only $14.99 USD on amazon.com. If I weren't in a rush to get my build together again I would order some to my friend's mailbox in the States and grab a couple tubes
wink.gif
.
i dont know if you read the thread but this was tested and proven to be better than diamond ic 7 .

see the old test please
 
#19 ·
Quote:
Originally Posted by Diablosbud View Post

I'm not sure if this is what you were intending for, but graphene-based thermal paste does exist and it's one of the best thermal paste bases out there... likely better than CLU judging by their comparisons to AS5/MX-4. http://thermene.com

It's only $14.99 USD on amazon.com. If I weren't in a rush to get my build together again I would order some to my friend's mailbox in the States and grab a couple tubes
wink.gif
.
I shot them an email to see how that stuff handles low temps. Depending on its base it could be a valid replacement for AS Ceramique 2 that most TEC/Phase guys with aluminum cold plates are forced to use since we can't use Gallium based TIMs.

AS Ceramique 2 does not perform as well as many other TIMs, but is pretty much the only TIM proven to last in a 24/7 rig that sees sub ambient temps without shrinking, drying out, or otherwise failing...its good to -150c!
 
#20 ·
Quote:
Originally Posted by HitMe View Post

i dont know if you read the thread but this was tested and proven to be better than diamond ic 7 .

see the old test please
Oh, sorry I didn't realize
tongue.gif
. I was just trying to help.
 
#21 ·
Here is there response about temp range:
Quote:
Thermene does not freeze, shrink, or crack, or loose its liquid state below 0°C, which means it should work very well at the temperatures you are describing. If it doesn't, just tell us and we'll refund you.

Thanks,
Thermene
 
#22 ·
I heard about thermene. I haven't yet seen a study actually comparing thermene in real world tests, apart from the claims on their own site. There is also some youtuber who claims "thermene" ate through his cpu's HS and heatsink. It all seems a bit sketchy.

If you're gonna try this thermene feel free to post temps of before, after etc etc.
 
#23 ·
Quote:
Originally Posted by ThijsH View Post

I heard about thermene. I haven't yet seen a study actually comparing thermene in real world tests, apart from the claims on their own site. There is also some youtuber who claims "thermene" ate through his cpu's HS and heatsink. It all seems a bit sketchy.

If you're gonna try this thermene feel free to post temps of before, after etc etc.
Wow. I don't really see how it could they list the chemicals in it. I have heard about IC Diamond 7 scraping the IHS because they didn't grind the diamond shards down enough. That's the only case I've ever heard of something like that and it was only one batch. I would love to try it, but I don't have a passport and the States is the only place to get it. I've settled on Gelid GC-Extreme for now, but on future builds I hope to do better.
 
#24 ·
Quote:
Originally Posted by Diablosbud View Post

Wow. I don't really see how it could they list the chemicals in it. I have heard about IC Diamond 7 scraping the IHS because they didn't grind the diamond shards down enough. That's the only case I've ever heard of something like that and it was only one batch. I would love to try it, but I don't have a passport and the States is the only place to get it. I've settled on Gelid GC-Extreme for now, but on future builds I hope to do better.
thats true IC Diamond 7 did scrap my 3770k die
Gelid GC-Extreme , never use it ;'(
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top