All I have is MX-2.....you guys think that will work good ?

Thanks for the guide. I tried it on my GTX 470 with AS5 only to find it now runs 3-4° C hotter.  I'm not sure if my TIM application was bad, or I seated the heatsink poorly, or if it's just a lot warmer in here today. But I'll probably try it again later and hope for better results.
Even if I am a noob, thanks again for the guide.

heydabop

Thanks for the guide. I tried it on my GTX 470 with AS5 only to find it now runs 3-4° C hotter.  I'm not sure if my TIM application was bad, or I seated the heatsink poorly, or if it's just a lot warmer in here today. But I'll probably try it again later and hope for better results.
Even if I am a noob, thanks again for the guide.

AS5 has a cure time of 200 hours. They should equalize -- but double check everything just to be sure.

Thanks this really helped. One card is still hotter than the other but I figured out it's because it has a higher voltage set. Still this dropped my temps 5c.

artic silver ceramique dropped my idle temps about 7c with fan at 59% idle temp is 39c as opposed to 46c with the nvidia paste, now what impressed me the most is furmark extreme burn,fan at 75% max temp was 79c (oc clocks are 1500 and memory at 1900),rather impressed after replacing the tim

edit: also after testing games an benchmarks temps are barely breaking 70c, artic silver ceramique ftw!

efeman

I did this to my GTX 470 last Friday with AS5. I know there's a cure time, but I haven't seen any improvements in temps (around 45c idle, 90c load). I had taken a couple of photos when applying the TIM myself; it was my first time not letting the heatsink do the spreading work, and my first time on a GPU. Does it look right?

It might be a little thick but there's a number of reasons why you may not see a decrease in temperatures:

• Cure time
• Good factory paste application
• Inefficient air cooling/fan profile
• Card default voltage

Overall, your temps appear to be just fine. 45*C idle isn't too bad assuming it's the default fan profile. 90*C in vantage or with OC Scanner/Furmark or in-game? If it's 90*C with OC Scanner/Furmark with the default fan profile that's better than me.

efeman

I did this to my GTX 470 last Friday with AS5. I know there's a cure time, but I haven't seen any improvements in temps (around 45c idle, 90c load). I had taken a couple of photos when applying the TIM myself; it was my first time not letting the heatsink do the spreading work, and my first time on a GPU. Does it look right?

Meh, better than me. I put AS5 on a few days ago an actually increased my temps a bit. I idle at ~49° C.

heydabop

efeman

I did this to my GTX 470 last Friday with AS5. I know there's a cure time, but I haven't seen any improvements in temps (around 45c idle, 90c load). I had taken a couple of photos when applying the TIM myself; it was my first time not letting the heatsink do the spreading work, and my first time on a GPU. Does it look right?

Meh, better than me. I put AS5 on a few days ago an actually increased my temps a bit. I idle at ~49° C.

Yeah, that's why I only recommend it for folks who have temps outside the norm. Look at this way though -- that paste will get crusty in about 6 months and lose most of it's performance ability. The AS5 will continue performing long after the crusty paste hardens.

So it's not all bad news.

Let me add a warning to those who like to just use a really thin layer of paste.

One of my two 480's seemed like it might benefit from replacing the paste, so I popped the heatsink off and cleaned things up. That's when I noticed that the heatpipes were actually recessed into the base of the heatsink compared to the "strips" of base material between the heatpipes. In other words, I could lay a small metal ruler across the heatsink and see a gap between the ruler and the heatpipes. If I'd just used a really thin layer of paste it wouldn't have even been contacting the heatpipes.

After thinking about things for a bit I decided to solve the problem by lapping the heatsink base. Yeah, it probably voided the warranty, but I figured if I ever had to return the card that I'd just swap heatsinks with the unmodifed one from my other card.

I did gain a couple of degrees improvement from doing all this, but really if I had it to do over again I wouldn't bother changing anything. The gain was minor. I really think that the thermal resistance of the heatsink to air is large enough to swamp out the minor gain from improving the thermal interface between the heatsink base and the fermi heatspreader.

mwparrish

heydabop

efeman

I did this to my GTX 470 last Friday with AS5. I know there's a cure time, but I haven't seen any improvements in temps (around 45c idle, 90c load). I had taken a couple of photos when applying the TIM myself; it was my first time not letting the heatsink do the spreading work, and my first time on a GPU. Does it look right?

Meh, better than me. I put AS5 on a few days ago an actually increased my temps a bit. I idle at ~49° C.

Yeah, that's why I only recommend it for folks who have temps outside the norm. Look at this way though -- that paste will get crusty in about 6 months and lose most of it's performance ability. The AS5 will continue performing long after the crusty paste hardens.

So it's not all bad news.

Well that's good. And it's not like I'm getting dangerously hot right now anyway.
Thanks again for the guide.

i had a question my temps on idle were 47c.and 70c playing just cause.so i opened the card up to look and sure enough glob of paste ulgy,so i cleaned it up and used as5 i  first applied to heatsink to fill in cracks so it would be  flush with heat sink then scraped of remaining as5.then i applied to gpu and nice 35c on idle 60-63c playing just cause.
but when i install second monitor idle shoots to 47c does anyone else see temp's raise like that with second monitor?when new drivers come out will they fix that

Excellent post, grats on the BR very well deserving. Taking the time to help others is what this hobby is all about. Glad we have you here on the forums!

Do you or anyone have an opinion on application of TIM, I had always used the grain of salt method of old days. However now that I am water cooling I searched around and found people using credit cards to spread a thin layer all over the chip(s).

I did this myself about 3-4 hours ago for my NB/SB and CPU before installing water blocks. Did a quick power on test to ensure I didn't brick my board somehow foolishly and all is well. Went to bios really quick and seems to have worked out pretty good.

With the little drop in the center I always found it did not spread evenly now spreading carefully with credit card (cut up into a small piece) I was able to get a very thin layer on the entire chip area.

Pentium777

Excellent post, grats on the BR very well deserving. Taking the time to help others is what this hobby is all about. Glad we have you here on the forums!

Do you or anyone have an opinion on application of TIM, I had always used the grain of salt method of old days. However now that I am water cooling I searched around and found people using credit cards to spread a thin layer all over the chip(s).

I did this myself about 3-4 hours ago for my NB/SB and CPU before installing water blocks. Did a quick power on test to ensure I didn't brick my board somehow foolishly and all is well. Went to bios really quick and seems to have worked out pretty good.

With the little drop in the center I always found it did not spread evenly now spreading carefully with credit card (cut up into a small piece) I was able to get a very thin layer on the entire chip area.

Yeah, with CPUs... if you're not using Indigo Xtreme, I recommend the rice grain method in the long direction of the die with a flat polished heat sink. If it's a HDT sink (pipes exposed with grooves between pipes and surface) then I recommend filling in those cracks first then two thin rice grains so it spreads evenly across the pipes. This works best to eliminate voids.

For GPUs, it's spread with a card only for me. Voids aren't as big of an issue because the heatsink is held on with as much pressure anyway to trap air. Also, the GPU die usually takes up more space relative to the heatspreader so you want to adequately cover it all for optimal thermal transfer.

Just an update to the above post. 12:45 PM in South Florida, air conditioner set to 74 F.

GPU Temps @ Idle (EVGA GTX 480 SCx2):

41 C
44 C

Great guide and well deserved BR! I can't wait to get back from deployment and check out my before and after on my new 480s waiting for me when I get home! (after all the QT with the wife and kids of course!)