Blob vs spread (applying thermal paste)

[unholygregor]

Hi guys,
My new Coolermaster V8 is on its way somewhere in the post along with some Arctic silver 5. I should have it in the next five days so I was wandering what is the best way to apply thermal compound?

-Some people say its better to put a pea sized blob in the middle and then put the heatsink on it, which will spread it out and create a vacumn.
-Others say its best to spread it with your finger in a plastic bag, as you can then be sure you have even coverage.

Can anyone give me some advice as to which way is better any why?
I like the blob method but I dont like the way you cant be sure you have a full coverage when doing it.
Thanks,Greg.

 

[fanaticvo0_20]

The blob method to me makes more sense.
i have been a I.T technician for 3years now, i have used this method alot of times.
The Reason i Stick behind this method is because, When you apply the paste in the center of the cpu and exsert force equaly, the paste is spred evenly and forced into the textured surface between the Heatsink and Cpu. (Thats Why lapping a Cpu Brings the Temp down because you Removing a Layer of metal and evening out the rough texture of the cpu)

So if you just Spreading you will then get Uneven patches of thermal past that when the heatsink is put on it sits with patches that are not making Contact with the Heatsink.

Hopes this Clears things up.

 

[LoneWolf_53]

+1 Can't explain it any better than that.

 

[unholygregor]

Thanks for the advice, thats how i'l apply my new Thermal Paste then

 

[fanaticvo0_20]

Glad to have helped

 

[OvrClkr]

There are a few different ways of doing this. My way is a tad different than many others, I just put a dab of paste on the cooler itself and spread evenly...here is a guide that might help you out.....

http://www.frozencpu.com/resource/r8/How_to_Apply_Thermal_Paste.html

I do not recommend JUST putting a dab on the CPU and slapping the cooler on because if you use too much chances are it might pour onto the board and if the paste is conductive like AS5, it will damage the board.

 

[overshocked]

The way most pros do it is 5 or so small dabs of the TIM all around. (this is usually done with "white" TIMs as they are not conductive as ovrclkr said.

BTW OvrClkr, i love the new quote!

 

[rsmith13]

I put a small dab on both the cpu and heatsink, then spread it with my finger. I then put the heatsink on the cpu and slide it around to make sure that the compound is spread evenly. I lift the heatsink off and check to make sure there is even contact. Then I put the heatsink back on , slide around again and clamp down.

 

[LoneWolf_53]

If I understood all the various info out there correctly that method is not advised since it can create pockets of bad contact.

 

[fanaticvo0_20]

I do not recommend JUST putting a dab on the CPU and slapping the cooler on because if you use too much chances are it might pour onto the board and if the paste is conductive like AS5, it will damage the board.[/quotemsg]



Well Then that Whould be His Problem, But if you were a person with Common Sence You Whould Put a Tiny Blob and Slap on the Heatsink!

''Arctic Silver 5 will not separate, run, migrate, or bleed.''

Direct from There Website!

 

[jsc]

Best instructions that I have seen:
Suggestions for applying thermal compound:
http://benchmarkreviews.com/index.php?option=com_content&task=view&id=170&Itemid=1&limit=1&limitstart=0

 

[randomizer]

''Arctic Silver 5 will not separate, run, migrate, or bleed.''

Direct from There Website!

They don't know their own products very well. I opened a brand new tube the other day and it ran almost like water. I had to squeeze some out until the thin, watery material was gone and then clean it all off so I could start over. I wasted a good 5% of the tube.

@unholygregor: It really depends on the TIM. Some are quite runny, like AS5, while others are thick and abrasive, like IC Diamond. The latter require large amounts of pressure or they simply will not spread. I think ICD requires 50psi minimum for optimal spreading.

 

[jsc]

I wasted a good 5% of the tube.

Fortunately, even a small tube goes a long way.

 

[OvrClkr]

The way most pros do it is 5 or so small dabs of the TIM all around. (this is usually done with "white" TIMs as they are not conductive as ovrclkr said.

BTW OvrClkr, i love the new quote!

hehe, yea I was bored

yo, have you used this new Gelid compound yet?

GELID Solutions GC-Extreme
http://www.newegg.com/Product/Product.aspx?Item=N82E16835426020&cm_re=Gelid%3f-_-35-426-020-_-Product

if not you should give it a try... lowered my temps by -5/-6 vs. AS5 and MX-2 [:lectrocrew:6]

 

[overshocked]

Ive never had a probem with AS5 running, any way i took this from their website

"Not Electrically Conductive:
Arctic Silver 5 was formulated to conduct heat, not electricity."

@ovrclkr

This stuff is much better than both ( :

http://www.svc.com/1366-x1.html

 

[OvrClkr]

Ive never had a probem with AS5 running, any way i took this from their website

"Not Electrically Conductive:
Arctic Silver 5 was formulated to conduct heat, not electricity."

@ovrclkr

This stuff is much better than both ( :

http://www.svc.com/1366-x1.html

Yea, I looked into that over at frozen CPU but it is just too dam expensive for 2 applications

The Gelid Extreme is working just fine. Got the B50 at 4.215Ghz and it will not reach 52c no matter how hot my room is... I even cranked up the heater last night to 80F and the temps stayed the same running large FFT's for over 4 hours.... With my room temp at 70F or lower my LOAD temps hover around 43-45c.... Need a case though, looking at the 902 vs. the haf...

 

[AsAnAtheist]

My method for applying thermal compound depends on several factors for me:
1. Viscosity
2. Electrical conductivity (Arctic Silver 5 is electrically conductive, not exactly your standard gold conductivity but it's enough to cause shorts).
3. Heatsink surface's shape, and size.

This is keeping surface area for heat conductivity, this has NOT being tested properly. I have applied thermal compound using different methods and have yet had to re-seat a heat sink or re-apply thermal compound because of faulty placement of heat sink or thermal compound following these 3 procedures.

Viscosity is measured by placing it in a clean surface, and taking a cotton swab over it. If it is not easily spread apart by the cotton swab, it will then go with a small amount in the middle, letting the heatsink even it out through pressure, and thermal cycles. The reason why I would do this is if the material is extremely "sticky" and has high viscosity then it would be improbable for it to leak out of the heatsink and will prevent the formation of air bubbles. If the material is less viscous and is spread more easily, then I would have to go with a spread method covering about 3/4ths the CPU's heat spreader to form an extremely thin layer. This will allow it to disperse the easiest. I would then immediately with a cotton swab clean off a good 2mm off the edge. Now I know this sounds weird but there is a reason why I say to spread the thermal compound into a very thin layer if it is easy to spread apart. That is because since it is spread out, the pressure for the heat sink will matter very little, and will only help to even the spread, not so much thin it, this is also prevent air bubbles because of the thinness and any leaking by removing most of the thermal compound from around the edges.

Electrical conductivity is quite frankly easily researched, most thermal compound companies will list it there.

3. Heatsink surface's shape and size is incredibly important when applying your thermal compound.
If the surface is small, then the best approach would be a spread by user method regardless of thermal compound. If the size is large then a dab in the middle would be preferred however would be dependent also on the thermal compound's viscosity. All in all you need to find a heat sink to come close to matching the size of your CPU's heat spreader.

 

[OvrClkr]

My method for applying thermal compound depends on several factors for me:
1. Viscosity
2. Electrical conductivity (Arctic Silver 5 is electrically conductive, not exactly your standard gold conductivity but it's enough to cause shorts).
3. Heatsink surface's shape, and size.

This is keeping surface area for heat conductivity, this has NOT being tested properly. I have applied thermal compound using different methods and have yet had to re-seat a heat sink or re-apply thermal compound because of faulty placement of heat sink or thermal compound following these 3 procedures.

Viscosity is measured by placing it in a clean surface, and taking a cotton swab over it. If it is not easily spread apart by the cotton swab, it will then go with a small amount in the middle, letting the heatsink even it out through pressure, and thermal cycles. The reason why I would do this is if the material is extremely "sticky" and has high viscosity then it would be improbable for it to leak out of the heatsink and will prevent the formation of air bubbles. If the material is less viscous and is spread more easily, then I would have to go with a spread method covering about 3/4ths the CPU's heat spreader to form an extremely thin layer. This will allow it to disperse the easiest. I would then immediately with a cotton swab clean off a good 2mm off the edge. Now I know this sounds weird but there is a reason why I say to spread the thermal compound into a very thin layer if it is easy to spread apart. That is because since it is spread out, the pressure for the heat sink will matter very little, and will only help to even the spread, not so much thin it, this is also prevent air bubbles because of the thinness and any leaking by removing most of the thermal compound from around the edges.

Electrical conductivity is quite frankly easily researched, most thermal compound companies will list it there.

3. Heatsink surface's shape and size is incredibly important when applying your thermal compound.
If the surface is small, then the best approach would be a spread by user method regardless of thermal compound. If the size is large then a dab in the middle would be preferred however would be dependent also on the thermal compound's viscosity. All in all you need to find a heat sink to come close to matching the size of your CPU's heat spreader.

Nice...

+1

 

[overshocked]

Yea, I looked into that over at frozen CPU but it is just too dam expensive for 2 applications

The Gelid Extreme is working just fine. Got the B50 at 4.215Ghz and it will not reach 52c no matter how hot my room is... I even cranked up the heater last night to 80F and the temps stayed the same running large FFT's for over 4 hours.... With my room temp at 70F or lower my LOAD temps hover around 43-45c.... Need a case though, looking at the 902 vs. the haf...

Word...

Its waaaay to expensive.

 

[jonpaul37]

Hmmm..., what about a Xiggy HDT s1283 where there are 3 copper pipes & 2 aluminum spacers between the 3 copper pipes? What then would you do in that situation?

 

[OvrClkr]

Word...

Its waaaay to expensive.

You can get a CCF for a few bucks more... :lol:

JohnnyLucky uses it and he says it performs as adverstised, I guess if you have cash to burn it should be worth a try....

 

[delluser1]

Hmmm..., what about a Xiggy HDT s1283 where there are 3 copper pipes & 2 aluminum spacers between the 3 copper pipes? What then would you do in that situation?

Read the link that jsc provided.

 

[Conumdrum]

The HDT is done a few diff ways. One is to cover the base, making sure you fill in the deep spots, scrape off the extra and then use a blob. Or make three lines, one down each heatpipe. the lines are VERY thin BTW. It's covered in Youtube vids and pics on many forums if you dig deep enuff.

Important point. Even my MX 2 paste (I have done over many with it) when using a new tube (I have used 3 tubes on CPUs and GPUs) there was a slimy goop at first. Since I'm experianced, I know to squeeze the first bit onto something else besides a CPU chip or GPU chip. It's the small things you learn to make your life easier. The only way is to do it over many years, you kinda learn what is right.

You also get better at reading things and deciding whats crap and not.

Soo, I'm done.

 

[darkfa8]

in a pinch I bought some of the Dynex stuff from Best Buy, spread a thin layer over the entire mating surface of the HDT-1283, carefully placed it on the CPU (AMD Pehon II X3 720), in the process of installing the retention bracket the cooler slid around a little - locked the cooler down.

With a 800MHz multiplier overclock I'm seeing ~30C idle and 53C under load with Prime95 running a blended, 4-thread torture test for 15 minutes.

I think I may improve my cooling by upgrading the HDT-1283 fan to a higher CFM. The only technical difference between the HDT-1283 and Dark Knight 1283 is the fan and the Dark Knight was higher in the ranking on Frostytech's chart. I may also step up to a higher CFM fan for the 120mm case exhaust too.

My temps aren't terrible, but lower would be great if I can do it for under $40 or so.

I've read so many reviews on thermal compound and application. Many have stated that the difference between compounds is neglible and instead the difference in application, coolers, temperature conditions are more likely to effect your system.

For those who have just and ONLY changed compounds and then realized a temp drop, maybe your application was different, maybe it was thinner, maybe you cleaned out some dust...idk.

I was advised by one retailer to clean off the Dynex and use MX2 instead and that'd I'd definitely see a difference. IDK that it's worth the trouble, rather improve the physical airflow within the case - may be more beneficial.

My system is a Rosewill mATX mid-tower, no discrete GPU at the moment, HDT-1283 blowing into the CPU cooler and up into the OCZ Fatal1ty 550 PSU 120mm fan that sucks air up through the PSU and out the back. Also have the stock Rosewill 120mm case fan that blows out the back of the case. The left side of the case has vent holes that are open.

 

[OvrClkr]

in a pinch I bought some of the Dynex stuff from Best Buy, spread a thin layer over the entire mating surface of the HDT-1283, carefully placed it on the CPU (AMD Pehon II X3 720), in the process of installing the retention bracket the cooler slid around a little - locked the cooler down.

With a 800MHz multiplier overclock I'm seeing ~30C idle and 53C under load with Prime95 running a blended, 4-thread torture test for 15 minutes.

I think I may improve my cooling by upgrading the HDT-1283 fan to a higher CFM. The only technical difference between the HDT-1283 and Dark Knight 1283 is the fan and the Dark Knight was higher in the ranking on Frostytech's chart. I may also step up to a higher CFM fan for the 120mm case exhaust too.

My temps aren't terrible, but lower would be great if I can do it for under $40 or so.

I've read so many reviews on thermal compound and application. Many have stated that the difference between compounds is neglible and instead the difference in application, coolers, temperature conditions are more likely to effect your system.

For those who have just and ONLY changed compounds and then realized a temp drop, maybe your application was different, maybe it was thinner, maybe you cleaned out some dust...idk.

I was advised by one retailer to clean off the Dynex and use MX2 instead and that'd I'd definitely see a difference. IDK that it's worth the trouble, rather improve the physical airflow within the case - may be more beneficial.

My system is a Rosewill mATX mid-tower, no discrete GPU at the moment, HDT-1283 blowing into the CPU cooler and up into the OCZ Fatal1ty 550 PSU 120mm fan that sucks air up through the PSU and out the back. Also have the stock Rosewill 120mm case fan that blows out the back of the case. The left side of the case has vent holes that are open.

Get rid of that Dynex paste and get some AS5 or MX-2... That Dynex paste is crap

I re-pasted last night with MX-2 and this was the outcome :

LOAD temps are around 44/45c at 4Ghz