why does ceramique cost less than silver 5?

tano22

I'd go ceramique. That's the stuff I used and I idle at 41-44C and I've had it in less than 48 hours.

yup my idle in the new case (moderate fans all set to min) and the 480 set to auto 44% fan speed idles around 44 with ambient around 24'ishC (75'ish F) using OCZ Freeze ceramic

meh, difference is most likely only a deg or so. difference between crappy stock TIM and an "aftermarket" ceramic or silver can be more drastic (drastic still probably under 10 C difference) but when you get into brands and types your looking at only a deg or so 

how long would you run furmark to tell what your stable max is? and where would you tell the idle temps. when its at its lowest mark or after a certain amount of time

The only problem with IC Diamond is that the diamond particles will scratch the surfaces of the heatsink and GPU.

This may affect warranty coverage depending on how bad it is -- but typically -- the biggest issue with it is that it's difficult to spread across a GPU to ensure the entire surface is fully covered.

In my opinion -- you'd be better off with OCZ Freeze or MX-2 -- in that order.

very happy with the results of mine using freeze

I replaced the tim on 1 of my cards using Prolimatech.  Dropped my idle temps on that card by 5c.

While removing the 4 screws I noticed 1 of them was already loose.

Go figure my 480 had a screw loose.  I guess it matches it's owner.

Not sure if I will do my other card with the same tim or try one of my various others I have sitting around.  I have found the Prolimatech to be rather good.  Personally I find it better than Ocz freeze.  Works really well for my CPU.

Sure would be nice if Nvidia would just use and apply correctly a good quality tim from the beginning.  I can just picture a machine dropping a giant jelly bean size glob on the gpu before the heat sink goes on.  What a nightmare!

^^^^^^^^ Yeah, that's AS5 which was discussed in quite a few posts above.

  Doesn't void the warranty. Sometimes we are told to reapply it are selves by techs before we RMA a card. As far as AS5 I have used it on all my GPUs. Someone here said its capacitive, not conductive, in other words it hold an electrical charge if its touching something its not supposed to. Either way as long as you don't go nuts spreading it around you should be fine. Having it applied to three 275s for 7 months they all still work just fine. Waiting on OCZ Freeze though for my 470SC because I heard it was better. Stock paste I idle at 39-40 with Fan at 70 and OC Scanner it puts it at 90 degrees with 100% fan. Not sure if that good or not but I don't think it can get worse. Just stay the same or get better.

oscuro

diynoob

^^^^^^^^ Yeah, that's AS5 which was discussed in quite a few posts above.

Aw crap.  A friend of mine told me he recommends Arctic Ceramique over Silver because Silver, according to him "is slightly electrically conductive."  :(

Would Arctic Ceramique work too?

And whats the word on the warranty issue?

Not conductive -- the operative word is capacitive. BIG difference.

I used AS5 for the example. I've never had any issues. But like I've said many times, I'm not clumsy, I'm very neat, and borderline OCD.

If you're the opposite or even remotely the opposite of any of those things -- use something else like AS Ceramique.

I had some trouble with my second card (Orange PCB), almost 5 degree more than my top card (Always mesured with a constant room temp).

I replaced the stock TIM with some AS Ceramique and right now both card are @ same temp, GPU-Z; 36,30,30.8,30.3,44%,1700 and room temp constant 21.5.

The only thing is that i applied the compound the (HP) way, usually they recommend an S pattern but i found with some cheap stuff that it wasn't working pretty well with a die that big...

So i tried the line pattern but instead of a single line in the center of the GPU i've tried 4 thin lines starting from the edge of the GPU @ equal distance and this is the most effective way to apply it without air bubble.

Their is always chances to get alot of air bubble when you spread the paste on your die.

Ill probably do the top one later cuz once the burn-in period done, temp should be better than the top one.

I know, but theres is always greater chances to get air bubble when your spreading it.....Thats a fact, period.

And once most manufacturer are recommending everything but spreading i'm sticking with it...HP recommend five dots or S pattern.

I'm not saying that your technique is wrong but i always had better result with those technique and trust me i've tried the spread technique quite alot...
 
I've tried about 10 times with different stuff on my 480 before choosing the right pattern just to be sure and the 4 lines has been the most relevant between all technique.

mst-hilaire

I know, but theres is always greater chances to get air bubble when your spreading it.....Thats a fact, period.

And once most manufacturer are recommending everything but spreading i'm sticking with it...HP recommend five dots or S pattern.

I'm not saying that your technique is wrong but i always had better result with those technique and trust me i've tried the spread technique quite alot...
 
I've tried about 10 times with different stuff on my 480 before choosing the right pattern just to be sure and the 4 lines has been the most relevant between all technique.

HP recommends a "5 dot" or "S-pattern" for application of TIM on NVIDIA GF100 GPUs? Am I understanding you correctly? Is there a source for that?

Further, the die area is significantly larger than a CPU nearly necessitating full coverage of the heatspreader for maximum transfer.

Third, the applied pressure between the GF100 heatsink and the heatspreader is no where close to the pressure applied when mating a CPU heatsink and heatspreader. Thus, there will always be voids unless you specifically fill in the grooves or lap the heatsink. Coincidentally, the grooves themselves likely eliminate almost all potential voids because the air is pushed into them when they're mated and either held there or forced out.

The bottom line -- it's more important to ensure the entire heatspreader is covered with TIM sufficient to ensure complete contact with minimum thickness with the heatsink.

Everything you're sounding an alarm about is with regard to CPUs more so than a GPU -- specifically these GPUs.

To each his own -- if you like it your way, fine. But it's no coincidence I have some of the coolest air cooled 480s around with this method. This method is also preferred for the previous generation GPUs as well.

Finally, as I stated before -- voids will always exist with this setup, and most for that matter, it cannot be completely eliminated -- one of the primary goals with reapplying TIM is to ensure maximal contact coverage -- which you can do simply -- you can't similarly eliminate all voids. It's the lesser of two evils -- in this case -- the lesser of the evils is ensuring full coverage for maximum transfer over the entire surface area.

Would it be possible to lap the stock heatsink a bit since its got so many grooves? I've never lapped anything, just an idea

Has anyone used Notcua NT-H1 paste?  I don't have any of the brands mentioned here on hand, only a tube of NT-H1.

Any thoughts on how that would work on a GTX470?

http://noctua.at/main.php?show=productview&products_id=13&lng=en
 
EDIT: Nevermind, I checked and while the tube is a large one the amount remaining isn't much and I didn't use alot when I put it on my CPU.  Wasted materials I guess, I'll pick up some IC Diamond from MicroCenter when I get around to this.