PDA

View Full Version : CPU Thermal Paste Question


nV`andrew
01-17-06, 07:36 PM
Hey, the X2's come with thermal paste instead of pads right? It looks like thats what mine came with(so the heatsink could hold on the processor).If so, do i have to reapply it often and how often? Also, which paste should i use?

Thanks,
andrew

Clay
01-17-06, 07:41 PM
You only need to re-apply if/when you remove the heatsink at some point in the future. Most people will recommend Arctic Silver as the thermal paste of choice.

jAkUp
01-17-06, 08:57 PM
Yea, I never liked the thermal tape stuff, I always use AS5.

hemmy
01-17-06, 09:05 PM
Honestly, the shin-etsu AMD uses on the bottom of the hs/fans is about as good as AS5

Roadhog
01-17-06, 09:12 PM
i highly doubt it.

Rytr
01-17-06, 09:15 PM
I like the shin-etsu pads for memory heatsinks if I have them, otherwise, I use Ceramique. Always use AS5 for CPU and GPU.

Ihmemies
01-18-06, 12:23 PM
I've this age-old tube of AS3. I've used it to like dozen processors, gfx cards etc. and it never ends. Great stuff.

Richteralan
01-18-06, 07:07 PM
The new Shin Etsu X23-7763D's performance is on par with AS5 or slightly better according to some people.
I bought a tube of it and will try to see how it compares with AS5.

TierMann
01-19-06, 02:11 AM
When I got my CPU+MOBO+RAM combo from Monarch it was re-assembled with shin-etsu. I replaced it with AS5 and didn't see a single degree difference. Pads suck though, but Shin Etsu or AS5 paste works great.

Capt. Picard
01-19-06, 02:36 AM
How can you say that the pads suck. Did you use them, measured the temp, took it off and cleaned it, then put on the paste and measured it again.

TierMann
01-19-06, 03:12 AM
MY XP 3000+ had a pad. I ripped it off, cleaned the heatsink and applied some AS5 which dropped the temp ~18c from what it used idle at and ~22c from the load temp. They're ok if you need to stick a ramsink on, but for the CPU or GPU they do a better job of keeping the chip from getting scratched than helping with temps.

seeker
01-21-06, 11:34 AM
I know nothing about pads, but I used a cheap Chinese generic paste and then replaced it with AS5, but it doesn't run any cooler now.

Capt. Picard
01-21-06, 11:42 AM
I know nothing about pads, but I used a cheap Chinese generic paste and then replaced it with AS5, but it doesn't run any cooler now.

Yeah I think that the fact that one brand runs signifigantly cooler than the other is a load of hogwash.

I use a cheap generic American paste called toothpaste.

Roadhog
01-21-06, 11:54 AM
no, its not. Ever read reviews on thermal paste? You can get a few C lower with using the best paste...

It's just common sence that a pad wont be as good as name brand paste.

Capt. Picard
01-21-06, 12:05 PM
Now why does a respectable company like AMD include something as cheap as this then as standard issue.

Nutty
01-21-06, 12:24 PM
Because its good enuff, and easier than paste for them to distribute. But by no means the best. Thats like saying why do they come with bog standard aluminium coolers, instead of top of the line copper ones ? Because its good enuff. Doesn't mean its the best tho.

jAkUp
01-21-06, 12:26 PM
Because its much cheaper when being mass produced. Also, it is easier for the average user to install if the tape is already put on the heatsink. AS5 is very expensive.

Really, the difference in temps is usually only a couple of degrees though.

seeker
01-21-06, 01:16 PM
In the review that I read, AS5 was noticeably cooler than it's competitors, but my Chinese generic only cost me $2.00, instead of $10.95, like AS5. I wouldn't have minded the extra cost, if I had seen a difference in temperature. Then again, perhaps my Chinese generic is AS5 by another name, because it looked exactly like it.

OWA
01-21-06, 02:54 PM
Maximum PC recently compared the various brands. Here is their table of results.

Capt. Picard
01-21-06, 03:04 PM
How much do you put on? I know it shouldn't squirt out and come into contact with the pins, but do you put on enough to just before that point or do you apply only a small dollop.

Nutty
01-21-06, 03:11 PM
What I do is squeeze on the heatsink block, then use a needly to pick a tiny bit up and spread it as ultra thingly as possible on the cpu.. Then rub in the stuff on the heatsink with a plastic bag over yer finger. Then remove all excess. Then stick em together.

The Key to sucessful application is to remember, that you dont want a thick layer of gue between the metal surfaces, you're just trying to fill in imperfections in each surface to get 100% pure contact between them.

seeker
01-21-06, 03:58 PM
Maximum PC recently compared the various brands. Here is their table of results.
I don't think that those numbers are very realistic, because the website that I had seen before quoted lower numbers for both idle and loaded. I know that my numbers have always been lower than those, regardless of the paste that I used.

ViN86
01-21-06, 04:26 PM
MY XP 3000+ had a pad. I ripped it off, cleaned the heatsink and applied some AS5 which dropped the temp ~18c from what it used idle at and ~22c from the load temp. They're ok if you need to stick a ramsink on, but for the CPU or GPU they do a better job of keeping the chip from getting scratched than helping with temps.
your temps changed 18-22C by just changing the paste? ehh that doesnt sound right

OWA
01-21-06, 04:41 PM
I don't think that those numbers are very realistic, because the website that I had seen before quoted lower numbers for both idle and loaded. I know that my numbers have always been lower than those, regardless of the paste that I used.
The numbers really shouldn't matter all that much. There are lots of reasons why they could be different from what you see. The only real thing that matters in a test like this is that they applied the paste and measured the temps in a consistent manner each time.

Clay
01-21-06, 07:48 PM
How much do you put on? I know it shouldn't squirt out and come into contact with the pins, but do you put on enough to just before that point or do you apply only a small dollop.
http://www.arcticsilver.com/arctic_silver_instructions.htm (skip down to Step #9 on that page)

The "very thin layer" technique is no longer what Arctic Silver recommends with newer CPUs (Intel P4 or Athlon64). I've always done this simple "rice grain" approach and had excellent results...makes a lot of sense too if you read their explanation.