Slashdot Mirror
Sony's Thermal Sheet Good As Paste For CPU Cooling
An anonymous reader writes "Sony has demonstrated a thermal sheet that it claims matches thermal paste in terms of cooling ability while beating it on life span. The key to the sheet is a combination of silicon and carbon fibers, to produce a thermal conductive layer that's between 0.3 and 2mm thick. In the demonstration, the same CPU was cooled by thermal paste and the thermal sheet side-by-side, with the paste keeping the processor at a steady 53 degrees Celsius. The sheet achieved a slightly better 50 degrees Celsius. The actual CPU used in the demonstration wasn't identified. Sony wants to get the thermal sheet used in servers and for projection units, but I can definitely see this being an option for typical PC builds, too. It's certainly going to be less messy and probably a lot cheaper than buying a tube of thermal paste."
>> probably a lot cheaper than buying a tube of thermal paste
Normally these come free with the cooling fan IME. Otherwise, a tube of paste is like $5.
>> key to the sheet is a combination of silicon and carbon fibers
Paste MIGHT be cheaper.
..it puts a rootkit on your machine, too.
Thermal paste for a typical CPU installation runs in the pennies, but you do have to buy a whole tube. I'll bet you a dollar this sheet for a single CPU install runs in the dollars. This is Sony we're talking about, they need profitable revenue to offset their sinking Blu-Ray-PS/3 ship, among others.
The cesspool just got a check and balance.
...but can it be reused? Can you put it on an entirely different processor after being used? If so, then it would definitely be worth the money... if it were made by somebody other than Sony, that is.
and "Behold!" cried the archangel, "Sony has done something cool. Tremble with fear all ye nations...."
-1, Too Many Layers Of Abstraction
isn't the advice to have rather less than 2mm paste between the chip and heatsink?
2mm of thermal paste is WAY too much and will most likely be less effective than no paste at all. Ideally there should be no visible paste at all after the heat sink is applied. The paste's job is only to fill in the tiny air gaps made by the imperfections in the "flat" contact areas.
3M has had a thermal sheet which has outperformed paste for more than 10 years already.
How is this news?
No sig. Move along - nothing to see here.
Paste of some kind is extremely important. New retail CPUs with stock heatsinks though usually come with some sort of paste pre-applied onto the base of the heatsink.
It always seemed to me that applying paste was more of an art than it should be.
This looks like it would be a more repeatable process.
Yes, definitely.
It improves heat conduction by filling the small surface defects that would create miniature pockets of air. Air is a very good insulator, and very poor at conducting heat.
There is a noticeable difference, you can research it further, but as cheap as a tube of thermal paste is, why not spend an extra couple bucks on your shiny new processor?
So, does it outperform the high end pastes, like Arctic Silver, or the cheap crap that comes with the 9.99 heat sinks?
There is a margin at least as wide as the one they list between those two substances. If the paste they tested against is anything like the garbage that goes in a PS3 from the factory, I'd expect mud and spit would heavily outperform it.
Just another ignorant American.
It'll probably only work when it's placed between a chip manufactured with a genuine Sony Heat Spreader, and genuine Sony Heat Sink.
How is this news?
Sony article that doesn't contain the words "Leak", "Compromised", "Stolen" or "Litigation".
Join the Slashcott! Feb 10 thru Feb 17!
That was in reference to the sheet, not the paste. Presumably the same thickness rules don't apply there.
which is totally what she said
Answering several commenters above and adding some more information as well...
1. Apple tried this out from about 1997 to 2002 in their G3 and G4 laptops and some of the desktops. They tried a variety of "thermal tabs". They worked ok. Sometimes they're quicker to put on, other times they take more time. Some were brittle. They should be available from a variety of sources at this point, not just Sony. They were also used on some of the G5's and mac pro CPUs. They tried quite a few variations over the years, and the most recent on the early mac pros were considered highly hazardous materials and we were advised to wear gloves when handling them and to not let them be exposed to air for any length of time. They may still be using them but the procs come attached to the heat sink so I don't have to handle them directly. All the products I lift heat sinks from have been using regular compound for the last several years. So I assume they figure the tabs are good for manufacturing time but not the best idea for field-repair. They may have been using 3M as a supplier, I don't know.
2. They were more expensive than thermal compound but easier to store a bunch of them in a small box/envelope.
3. I tried to reuse them and mostly failed. They tend to bond to either the heat sink or the die, or both, and get torn up pretty bad when you lift off the heat sink. Usually have to scrape the bits off both surfaces with a plastic spudger before using a new one. Makes taking things apart for test swapout or inspection a bit more of a hassle and a little more expensive.
4. one advantage they had was no spillover. A few systems I've worked with wouldn't tolerate heat sink goop spilling too far over to the ballast resistors or caps mounted near the die on the package. For those you had to be very careful about how much compound you used so it wouldn't squish out and touch something it shouldn't and generate some capacitance that would cause wonky behavior from the cpu. These are idiot-proof that way for the most part. I've also been told about problems with getting an air bubble in with the compound and creating a pocket over the die with no compound on it - I've never had that happen to me personally but I've seen the effect a few times. This isn't possible with the tabs. I've also read cautions for not applying too much compound, as though if you put on too much it wouldn't squish out enough and would create too thick of a layer of compound between die and heat sink but I don't think that's likely to happen considering the viscosity of the compound and the torque of the heat sink.
5. Occasionally we'd get tabs that were the exact size of the die, or a little undersize, and those caused problems getting them on right with full coverage. I also watched a tech forget to replace the new tab, with the expected results, so you may run into a few oops moments when changing your technique.
I work for the Department of Redundancy Department.
At least not until the comments started.
So you're saying it's less of a thermal sheet and more of a thermal hospital gown?
Plenty fine here since I lap all of my heatsinks and tops of CPU die casing, and use a lapped copper shim as the go-between contact.
Thermal paste? Not needed here!
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
I bought a $3 tube years ago and I still have plenty left. You're only supposed to use a dot the size of a BB. If you smear it all over the CPU you're doing it wrong - aside from the mess, that's guaranteed to create air pockets.
(The cheap stuff is fine too. The expensive stuff may conduct heat better, but the layer of goo is so thin that it's only a fairly small percentage of your heat resistance.)
Still, these pads are interesting. It looks reliable and less prone to noob mistakes. Too bad it's Sony.
As others have said, 2mm doesn't apply to the paste. But anyway this is exactly why a sheet would be so much better. No more fretting about "do I have too much paste?" or "do I have too little paste?" or "I just had to pull the heatsink off again because I didn't line it up right the first time, do I need to start over and reapply paste?". Thermal paste is a messy PITA.
and the problem is that TFA is useless fluff devoid of any details on how the paste was applied.
Democracy is a sheep and two wolves deciding what to have for lunch. Freedom is a well armed sheep contesting the issue
Because even with 2 contact points a properly-lapped copper shim actually works better.
With thermal pads you have two contact points as well.
Ditto thermal grease. One side contacts the die, the other side contacts the heat sink.
"If introducing a shim or any sort of additional contact layer actually improved heat transfer"
What do you think thermal pads, thermal compound, and cooling plates are for?
"then why not sandwich 5 more shims? :P"
Because that'd just be a waste of money.
Hi, I design high-power LED systems that consume far more power than most CPUs. 300w array in a 30mm x 30mm package is just the latest breakthrough, and it's kept cool by....
A copper shim connecting the nickel-backed contact plate to a copper heat sink.
And still maintaining ~40C temps.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Man I can't stand those one-size-fits-nobody hospital gowns. Walking in them and feeling self-conscious about somebody looking at my pasty white butt cheeks. That's why I wear the slit to the front instead.