Separate names with a comma.
Discussion in 'Overclocking & Cooling' started by SonDa5, Oct 8, 2012.
This would make the ultimate TIM in form of pad/grease/liquid application.
Yes graphene has high thermal conductivity, it's been known about since...graphene was discovered.
But the problem is, getting the material into some usable application. You can't just stick a wafer of graphene between the heatsink, as thats not how TIM works (and would be pretty hard to work with). Youd have to mix it with something, and break it up (reducing the efficiency) to make it conform.
Then you'd still have the main problem of any TIM. people not applying it properly.
I'll reply again.
Graphene is very cool....I think diamond is still better.
Carbon materials are very directional.Yes they are fantastic materials for thermal management , but , correctly engineering it is very expensive.The lasers used in underwater fiberoptic repeaters are on diamond chips....for thousands of dollars a unit.
It's really a waste to worry about 'super fantastic TIM'....I'd rather Intel make cpu's on diamond substrates.....
That is the challenge. A thin graphene pad would probably work well. It is flexible.
I'm sure somebody is going to get rich with a patent on a TIM for graphene.
I think you are right about that.
We are entering "the Carbon Age"
There's a video out showing nanotube manufacturing......very cool.......I'm sure a graphene machine would'nt be any harder.
It's flexible, but it still won't fill the microscopic holes that TIM is used to fill. You are just putting another solid between two already solid plates. I don't think that'll help very much.
It's not so much flexibility as plasticity which is needed. Graphene is more like dry paper (flexible but not plastic), TIM is more like jelly (plastic and flexible and will conform to molds). Need to make Graphene paper mache somehow.
Nope, you just need Intel to make the top of the die accurate to the 1/10,000th of an inch without any holes in it. Then you can get a 1/10,000th of an inch cpu cooler which mates perfectly to the die without the need for TIM at all.
A Graphene IHS will probably be the first step in using it for cpu cooling.
That's probably a decade away, realistically, factoring in the lack of mainstream demand for high performance (hot) chips that are increasingly being sidelined, business-wise, in favour of low power solutions for mobile devices.
Maybe it'll appear in supercomputing in about 5 years.
First Graphene TIM is here now. Not sure how good it works though.
Wow, $45 for 3ml. I'm sure it won't be long before someone gets some to do some tests with.
That is a little too high of price for me to just jump in without seeing how good this batch works.
Theoretically the stuff should work great if it is mixed pure but I'm not sure how it will work doped with another material.
Looking forward to some reviews.
Kyle should do a review over that TIM.
I'd like to see that.
I really think Graphene is next big jump in TIM technology.
Sweet Jesus thats pricey! Look forward to seeing how it performs
But hey, there are people out there who run 4 titans, $45 is practically pennies to them.
There are no more "big" jumps left in TIM. We're already to the point where delta T across the CPU to CPU cooler interface is in the 1/10ths of degrees using existing high end TIM, you really want to shell out $50 for an "improvement" that is less then instrument error?
They'd probably have to make CPU covers and heatsink bases out of nano-constructed surfaces for maximum perfection to make a difference... and perhaps utilize graphene in these rather than as TIM (actually, perhaps graphene to graphene to graphene would work well?).
Graphene still moving forward in electrical conductor design. Good stuff. Battery application is revolutionary.
Sweet necro !
So has Diamond CVD.As much as graphene is taking off because of it's cheaper cost,Diamond's still better.
They make ski's with graphene (lol) sweet carvers, but I call BS.
I'll take it!!
I have been using this https://www.digikey.com/products/en?keywords=p13691-nd in my Ryzen rig I built in July, so far it has worked very well. It's like 13 bucks for a 7" x 4.5" sheet. Just cut it with scissors and set it on top of the CPU with tweezers.
How well is the pad working over grease/paste?
I'm betting it's worse.
The purpose of a proper CPU TIM is not to form a layer between the CPU and the heatsink (as such a graphite sheet would). For the most efficient heat transfer as much metal-to-metal contact as possible between the two is best. The TIM is only there to fill in the minuscule gaps that exist when the two are joined together, as air is really a bad thermal conductor.
Any time you see a thermal pad or such used it's because it's cheaper, for ease of installation, and/or heat transfer isn't as critical.
It works at least as good as MX4, I can run a half an hour of Prim95 SmallFFT's with a max of 70C, CPU idles at 35C. Not an exhaustive test I know.
I missed this.... Graphene Shoe soles.
I'd love to have a pair of these just based on the technology of the material.
graphene bonded with a water set bond agent ( like use on dentures LOL) or a combo effect of super glue and the thin a hair graphene layer or few?
effectively allowing you to "peel" the graphene TIM pad (like those listerine pocket pack films) put on the cpu etc real careful and give a good firm press to "break" the small glassy seal on the "side down" use will make the solvent/bond agent seep through and make contact proper with heatsinks.. no delid etc needed... problem solved... should not be all that $$ either, probably .000001 per application ^.^
They do post it notes, they do similar for glass protectors, I have seen like such for glass crack fixing, why not graphene etc peel and stick "pads" ^.^
I call 1% profit world wide rights and will gladly be a tester and such as well.
I don't love linus but the test is simple enough. Product is available and appears to work just fine.
I may try it on my next build. https://www.amazon.com/Innovation-C...e+thermal+pad&qid=1557510857&s=gateway&sr=8-3
How many are needed for a threadripper? 4 of those?
for TR it might be worth using the product in josephrr's, link cut down to size.
I would be interested in a side by side of the two products, but now that I'm in an ITX build I don't mess with it unless I have to lol.