View Full Version : HOPG or Highly Oriented Pyrolytic Graphite for short.
Ok, here is my question.
Has anyone ever heard or seen someone use this stuff for cooling cpu's?
I've looked around and it seems to have a great ability for thermal conductivity.
The source I talked to at one of the sites http://www.ntmdt.ru/products/HOPG/
said that the HOPG was rated for the following
for HOPG ZYA quality - 1800W/mK ± 20%
for HOPG ZYH quality - 1000W/mK ± 20%
Which puts it way up over copper or silver for example.
The only thing I was not able to find out was how fast it is able to radiate the heat it gathers away from itself. The person I talked with did mention that they use it in cooling systems for furnaces that produce silicon chips etc...
The reason I asked about this is that the above company make it into waffers that are perfect for mounting into the bottom of heatsinks, waterblocks and such. 10x10mm square or 20x20 with thicknesses rangin from 1 to 2 mm.
Drawback....is the cost. around 100 for a 10x10 high quality or the same for a lower quality 20x20 waffer. with minimum order of around 300US dollars.
Now hows that for interesting. Anyone out there have anything to add to this one. I'm also looking for a few takers on going in on buying some. I just can't afford around 420-450CDN dollars to buy a small packet of these waffers.
If anyone knows if these are able to radiate the heat they take in better or worse then say copper, let me know.
Orion
cowanrg
10-26-01, 11:58 AM
sounds interesting, bump?
What's with the bump thing?
BTW: They make the wafers in lower quality up to 50x50 not the 20x20 I mentioned before. A little more expensive, but whats 20 bucks when your already paying around 100 per wafer. :D
If anyone knows of a way a person could get a good bond between the PG and the heatsink, I'd like to know that as well. It's no good having the wafer there if it can't transfer to the heatsink right!
Orion
cowanrg
10-26-01, 12:40 PM
bump: posting a worthless message to a thread so its not buried at the bottom... since your post had 0 replies, it was quickly goes way down on the list of threads... so i replied, making it to the top of the pile again. just a way to get more people to look at this, since it could be pretty interesting.
sounds a BIT pricey though. i mean, you can get a silver coldplate for that. granted, this stuff COULD be better than silver, but by how much? we all know that the comparison benefit of copper to silver is really very very small. it's much more cost effetive to go from aluminum to copper, but going from copper to silver really doesnt get you that much for the extra $$. this stuff is more expensive than silver, therefore would have to be the god of coldplates. if it is, it still may not be wortht the money to shave off 2-3C from your chip temp. however, an entire heatsink made from this stuff could make the best HSF, because a low powered/silent fan could be used. so, for aircooling, it could be the ultimate, but for water/peltier setups, it would just be a very rich way to shave off a few degrees.
The Overclocker
10-26-01, 01:46 PM
Originally posted by ROH
What's with the bump thing?
BTW: They make the wafers in lower quality up to 50x50 not the 20x20 I mentioned before. A little more expensive, but whats 20 bucks when your already paying around 100 per wafer. :D
If anyone knows of a way a person could get a good bond between the PG and the heatsink, I'd like to know that as well. It's no good having the wafer there if it can't transfer to the heatsink right!
Orion
the bump thing is when ever you reply, the article goes to the top of the page, people simply put in bump
The Overclocker
10-26-01, 01:47 PM
graphite isn't very good for conduction heat, i wounder what this stuff is like
SavageHenry
10-26-01, 03:42 PM
Actually, it makes sense. Diamond also has great thermal conductivity (~2000 W/mK). . . Diamond is carbon with its molecules arranged in a certain matrix, so is graphite.
cowanrg
10-26-01, 04:14 PM
most natural materials are carbon-based... heck, humans are :-)
Christoph
10-26-01, 04:57 PM
That gives me an idea. Why don't you just stick your finger between the HSF and the core. JK:D
Actually, I remember reading that diamonds are a very very good conductor of heat, so this thing makes sense.
The Overclocker
10-27-01, 08:38 AM
but metals have free eletrons which aid with the transfer of around 40% of the heat
True, but I think it has something to do with the extreem uniformity of the graphite in this case that gives it the ability to remove heat as good as it does. It's basically uniform rows of graphite in layers.
The fact that it absorbs heat is not my worry. That according to the makers is done well. I'm more interested in how well it radiates heat.
As for the difference between it and silver. How much better it would be.
for HOPG ZYH quality - 1000W/mK ± 20% for the low quallity PG.
and silver if I remember correctly is somewere around the 500 mark with copper at the 300 mark and aluminum was down around 150 or 200W/mK. So it thermal canductivity is definately better by a fair amount. (if my brain is correct in it's aproximates. It's early) And the ability of the high quality PG is even that much better. But in a case like that DO the yealds in heat reduction deminish exponentially or something. I'll have to start looking around when I get home from work tonight.
Comments?
Orion
Carbon Brake disks in F1 absorb a lot of heat but they do not let go of it as easilly as it absorbs heat. It really depends on the arrangement of the molecules. You can also use graphite as an insulator.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.