Overclock.net › Forums › Cooling › Air Cooling › Thermal Paste Reapplication of GPU (not CPU) HDT Coolers
New Posts  All Forums:Forum Nav:

Thermal Paste Reapplication of GPU (not CPU) HDT Coolers - Page 6

post #51 of 59
Thread Starter 
Quote:
Originally Posted by hammong View Post

You are confused, agreed. =)

The factory probably had a roll applicator on the assembly line, they put down a "stripe" of TIM in approximately the right place and rely on the die to heatsink pressure to spread it out. It's not physically some person squiring a blob on the die and spreading it out, but it's the same net effect.

Not sure why temps are higher than they were in the past. What temps are you seeing now, and are they are the same fan speed % and ambient room temperature?

Greg

Seeing around 77C with stock speeds when they were around 64C before. Yes, same fan speed and room ambient temp (around 32C).

Quote:
Originally Posted by Bearybear View Post

http://www.coollaboratory.com/en/produkte/liquid-ultra/

I've used in on my GTX 670's die/cooler, my 3770K's die and on my CNPS12X CPU cooler which also has direct heat-pipe contact: http://content.hwigroup.net/images/products/xl/130363-2.jpg

I honestly don't think I'll ever go back to using any sort of paste type thermal compound, I saw a 30C temp drop after switching from NT-H1 which I believe is mainly because it's not only undoubtedly better but it's so much easier to apply correctly unlike paste.

Only downside to this stuff is that it can be a pain to remove and clean up, but it not like you plan to take your graphics card apart every once in a while.
Please, I washed my graphics card in hot water. Also, nearly every person that de-lids their CPU uses Liquid Ultra directly on the die, never seen anybody short anything etc.

Oh that, yes I know that paste. How much does that cost?

Also, is the application method for this paste always by spreading it in any surface (IHS or bare lid)?

How hard is it to remove it? Can you use 90% alcohol just like removing ordinary TIM?

What type of GTX 670 did you try applying it into?

I heard that this paste damages the heatsink base and the GPU die surface when used, right?
post #52 of 59
Price and some resellers should be listed on their website.

For application I always paint a thin as possible layer on both surfaces, for example: one droplet on the die and paint to cover the whole surface, another droplet on the cooler and paint the whole area that will be in contact with the die and then put the cooler on as quickly as possible, get it aligned and get it on nice and tight.

I've reapplied the stuff once from my GTX 670 and it came off easily enough with the included alcohol wipe. I personally don't bother buying anything special to remove TIM, I usually use some deodorant spray I have lying around does the same job, but I won't recommend it because who knows what could happen. The main thing worth noting is that Liquid Ultra dries on like a paint (it's comparable almost to a very soft solder) and doesn't stay wet.

The one in my sig rig, it isn't direct heat pip contact like your though, however my CPU cooler is and I've had no issues with that.

I've heard that too, though not experienced it, I'd believe any damage caused to the die would likely be caused by the way it's removed; using the alcohol wipe and a lint free cloth (the sort that you clean glasses/spectacles with) left my die as good as new with a mirror shine. The cooler surface though had sort of a brushed finish out of the box (wasn't a completely flat fully polished surface) and I noticed it had made the surface look darker because the TIM gets into the 'scratches' in the brushed surface, that said, a few minutes with the wire wool and a good clean with a lint free cloth left it looking better than it did originally. I think damage is the entirely wrong term for that, it's just a discolouration because of the rougher surface.

I'd totally recommend the stuff personally as it solved my issues and worries with thermal compounds, how ever don't take my experience and the final answer, look into other cases.
post #53 of 59
Thread Starter 
Quote:
Originally Posted by Bearybear View Post

Price and some resellers should be listed on their website.

For application I always paint a thin as possible layer on both surfaces, for example: one droplet on the die and paint to cover the whole surface, another droplet on the cooler and paint the whole area that will be in contact with the die and then put the cooler on as quickly as possible, get it aligned and get it on nice and tight.

I've reapplied the stuff once from my GTX 670 and it came off easily enough with the included alcohol wipe. I personally don't bother buying anything special to remove TIM, I usually use some deodorant spray I have lying around does the same job, but I won't recommend it because who knows what could happen. The main thing worth noting is that Liquid Ultra dries on like a paint (it's comparable almost to a very soft solder) and doesn't stay wet.

The one in my sig rig, it isn't direct heat pip contact like your though, however my CPU cooler is and I've had no issues with that.

I've heard that too, though not experienced it, I'd believe any damage caused to the die would likely be caused by the way it's removed; using the alcohol wipe and a lint free cloth (the sort that you clean glasses/spectacles with) left my die as good as new with a mirror shine. The cooler surface though had sort of a brushed finish out of the box (wasn't a completely flat fully polished surface) and I noticed it had made the surface look darker because the TIM gets into the 'scratches' in the brushed surface, that said, a few minutes with the wire wool and a good clean with a lint free cloth left it looking better than it did originally. I think damage is the entirely wrong term for that, it's just a discolouration because of the rougher surface.

I'd totally recommend the stuff personally as it solved my issues and worries with thermal compounds, how ever don't take my experience and the final answer, look into other cases.

Thanks but when I look at reviews I see that the difference between the PK-1 and the Liquid Ultra is just around 2-3C?
post #54 of 59
Yeah, 2-3c
But what is the air temp going into your cooler compared to your room temperature? In most cases it's at least 3c and often much more... and every degree warmer the air going into cooler is translates into a degree warmer the GPU is. wink.gif
post #55 of 59
Thread Starter 
Quote:
Originally Posted by doyll View Post

Yeah, 2-3c
But what is the air temp going into your cooler compared to your room temperature? In most cases it's at least 3c and often much more... and every degree warmer the air going into cooler is translates into a degree warmer the GPU is. wink.gif

And your point is that every degree counts, right? smile.gif
post #56 of 59
Quote:
Originally Posted by kevindd992002 View Post

And your point is that every degree counts, right? smile.gif
True, but that's not really my point. biggrin.gif

My point is many cases are 5-8c warmer than room with user worrying about which TIM to use for a 1-2c improvement.... when they could setup case to run 3-6c cooler and therefore CPU would be 3-6c cooler... instead of only gaining 1-2c with new TIM. rolleyes.gif
post #57 of 59
Quote:
Originally Posted by kevindd992002 View Post

Seeing around 77C with stock speeds when they were around 64C before. Yes, same fan speed and room ambient temp (around 32C).

I hate to say this out loud, but I don't think you have a problem here....

77C is well within normal temp range for a GPU under load, ESPECIALLY with a really high 32C (almost 90 F) ambient temp. How long ago was the 64C temp? Different season, room ambient temp maybe? You're looking at a T delta of only 45C with the room now, and probably 35-38C delta with the in-case temps. That's not bad for air cooled GPU bro.

Greg
post #58 of 59
Thread Starter 
Quote:
Originally Posted by doyll View Post

True, but that's not really my point. biggrin.gif

My point is many cases are 5-8c warmer than room with user worrying about which TIM to use for a 1-2c improvement.... when they could setup case to run 3-6c cooler and therefore CPU would be 3-6c cooler... instead of only gaining 1-2c with new TIM. rolleyes.gif

Exactly smile.gif
Quote:
Originally Posted by hammong View Post

I hate to say this out loud, but I don't think you have a problem here....

77C is well within normal temp range for a GPU under load, and the 64C temp sounds like the temp sensor on the GPU itself might be reading low, ESPECIALLY with a really high 32C (almost 90 F) ambient temp. You're looking at a T delta of only 45C with the room, and probably 35-38C delta with the in-case temps. That's not bad for air cooled GPU bro.

Greg

I would believe that, yes, if I hadn't tried at least three Gigabyte GTX 670's before with the same 64-65C max load temp each time I run Heaven. So this has got to be the TIM application.
post #59 of 59
Quote:
Originally Posted by kevindd992002 View Post

Thanks but when I look at reviews I see that the difference between the PK-1 and the Liquid Ultra is just around 2-3C?

Depends which review you read, I've already said in regards to my CPU that "I saw a 30C temp drop after switching from NT-H1 which I believe is mainly because it's not only undoubtedly better but it's so much easier to apply correctly unlike paste."
You're looking at numbers based on 'best record set' by people that have likely tried and tested to find the best method and have likely done it hundreds of times over. A professional artist might compare pencils and recommend the best one but if I can't draw well in the first place, my sketches won't get any better.

Also bear in mind that a GTX 670 throttles at 70C and above, so I'd say every degree counts, especially if the card is over-clocked.
Edited by Bearybear - 1/7/14 at 11:00am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Air Cooling
Overclock.net › Forums › Cooling › Air Cooling › Thermal Paste Reapplication of GPU (not CPU) HDT Coolers