Replaced Thermal Compund

[vSyNd1c4t3]

Hi guys, I have a Corsai h100i, and I had it in Push fan configuration, and decided to turn it into a Push/Pull config.

 

In the process of adding the extra fans, I ended up having to take off my h100i from the CPU.

 

Before I did this I was running 4.7 at v1.262 with it never going above 70degrees C.

 

I replaced my thermal compound and added the new fans, but after booting and running tests to see the possible temp difference, it actually got worse...

 

I was reaching 82+ degrees C with the new configuration.

 

I figured I did something wrong with the thermal paste, so I took it back off, cleaned it up, and tried again...

 

Not getting 82+ degrees C anymore, but I am still getting much higher then before I did this new configuration. Getting 78 degrees C max after I replaced the thermal paste the second time.

I've only just replaced the paste last night, and so maybe it still needs time to settle, but I don't think its going to drop another 10 degrees from settling...

 

I put on a grain of rice worth's of paste, so I'm certain I'm not putting too much.

 

The paste I am using is Cooler Master ThemalFusion 400.

 

Should I try to replace it again? With even less paste?

 

I want my old temps back. :unsure:

 

 

[Gofspar]

I'd get some better paste like Gelid GC Extreme or Noctua NT-H1.

 

Gelid GC Extreme in the BEST no matter what people say.

[techguru]

try this

put a little on...and spread it out...almost like you are painting it on the surface of the cpu...cover the whole top o the cpu

 

then tighten down the cooler a bit more

 

also make sure the liquid is proper circulating

[vSyNd1c4t3]

I'd get some better paste like Gelid GC Extreme or Noctua NT-H1.

 

Gelid GC Extreme in the BEST no matter what people say.

I was thinking about getting new paste, but is it going to make a 10+ degree difference? D:

 

 

try this

put a little on...and spread it out...almost like you are painting it on the surface of the cpu...cover the whole top o the cpu

 

then tighten down the cooler a bit more

 

also make sure the liquid is proper circulating

 

I was thinking about trying this next, but wasn't sure if it was a good idea, as I don't see a lot of people using this method. (could be wrong)

 

h100i is known for it's unstable mounting? Should I get some extra washers?

 

It feels fine on there, but I think a 1/2- 2/3 turn could loosen the whole thing up.

 

I don't want to BEAR DOWN on my screws and break/crack my MoBo... very scary thought.

[Gofspar]

I was thinking about getting new paste, but is it going to make a 10+ degree difference? D:

 

 

 

I was thinking about trying this next, but wasn't sure if it was a good idea, as I don't see a lot of people using this method. (could be wrong)

 

h100i is known for it's unstable mounting? Should I get some extra washers?

 

It feels fine on there, but I think a 1/2- 2/3 turn could loosen the whole thing up.

 

I don't want to BEAR DOWN on my screws and break/crack my MoBo... very scary thought.

Gelid GC Extreme using the pea method would drop temps by like 8+. and drop even more over time from warming up and cooling down causing it to spread more over time.

 

I got mine yesterday and ill show you my results, I got these using the physics test in 3Dmark.

 

Before Gelid GC Extreme /w Kraken X61 and delidded i7 : 69c

Gelid GC Extreme 1st day 58c

2nd day of Gelid GC Extreme: 53c

[techguru]

painting it on the surface of the cpu is the proper way to do it

most people just glob it on cause its faster and easier

[Ertman]

painting it on the surface of the cpu is the proper way to do it

most people just glob it on cause its faster and easier

Ummm... I don't know about proper. Could you link me to a quality resource that says that?

[techguru]

what kinda resource.....

www.howtoapplythermalcompund.gov....?

 

it is the proper way to do it...any good tech knows this..or should know it 

[SeanBond]

Gelid GC Extreme using the pea method would drop temps by like 8+. and drop even more over time from warming up and cooling down causing it to spread more over time.

 

I got mine yesterday and ill show you my results, I got these using the physics test in 3Dmark.

 

Before Gelid GC Extreme /w Kraken X61 and delidded i7 : 69c

Gelid GC Extreme 1st day 58c

2nd day of Gelid GC Extreme: 53c

16 degree drop? Yeah maybe vs no thermal paste. But any reputable website will tell you even the worlds best thermal paste is only a few degrees off the mediocre stuff. You more than likely did something wrong the first time around.

[Gofspar]

16 degree drop? Yeah maybe vs no thermal paste. But any reputable website will tell you even the worlds best thermal paste is only a few degrees off the mediocre stuff. You more than likely did something wrong the first time around.

Yet again im delidded. it makes a bigger difference when you are direct die LOL

[vSyNd1c4t3]

16 degree drop? Yeah maybe vs no thermal paste. But any reputable website will tell you even the worlds best thermal paste is only a few degrees off the mediocre stuff. You more than likely did something wrong the first time around.

 

Let's not start a war guys.

 

My technical first time around is when it did the best, staying under 70degrees C with my overclock even while under max load.

 

It was after I added my push/pull fans and had to replace the compound that it starting getting over 70degrees.

 

I honestly think it's the way I am applying, first time (after adding push/pull) I'm guessing I put too much, so I did less the second time (after adding push/pull) and saw some improvement, but when I first put on paste (before adding push/pull) I was getting almost 12+ degrees difference.

 

SO YEA...

[vSyNd1c4t3]

painting it on the surface of the cpu is the proper way to do it

most people just glob it on cause its faster and easier

kk, got this done, and guess what!? PERFECT! Now I actually have a decrease in temp from my original thermal paste.

I will always spread my thermal paste from now on.

 

16 degree drop? Yeah maybe vs no thermal paste. But any reputable website will tell you even the worlds best thermal paste is only a few degrees off the mediocre stuff. You more than likely did something wrong the first time around.

Yup, going with a grain of rice and not spreading it seemed to cause heating issues, so I did the spread method and now it's great. Thanks for your input.

 

Ummm... I don't know about proper. Could you link me to a quality resource that says that?

 

Seems to have been the fix for me, went from 78 degrees C to 63 degrees C... Guess the proof is in the science!

 

Yet again im delidded. it makes a bigger difference when you are direct die LOL

 

Next time I buy thermal paste I will put your paste to the test, I looked up the numbers on your paste choice, and it seemed really good from real reviewers. So it will be the next type I get, thanks again for your help!

OK guys IT'S FIXED! Reapplied paste, and and made sure to spread a thin layer over the whole cpu, and now it's flawless.

My push/pull setup dropped me about 6 degrees C from my original Push only setup. So I'm very happy about that, and am now thinking of doing a bigger overclock!

 

Thanks everyone for your suggestions and input! :lol:     

[Queek]

kk, got this done, and guess what!? PERFECT! Now I actually have a decrease in temp from my original thermal paste.

I will always spread my thermal paste from now on.

 

Yup, going with a grain of rice and not spreading it seemed to cause heating issues, so I did the spread method and now it's great. Thanks for your input.

 

 

Seems to have been the fix for me, went from 78 degrees C to 63 degrees C... Guess the proof is in the science!

This finding surprises me.  TIM has such good thermal conductivity that there should be no practical difference between any two compounds.  I could have sworn that LTT did a video on this not too long ago, but I can't seem to find it.  For many intel desktop cpu's, the best method of applying TIM is a vertical line.  Source: http://linustechtips.com/main/topic/4405-how-to-apply-thermal-paste-correctly/ as per arcticsilver's website http://www.arcticsilver.com/methods.html.  Spreading the TIM can trap air bubbles, which diminishes the contact area and makes the heat transfer worse (air is worse than TIM for conducting heat [citation needed]).  

 

Commonly, large temperature differences are a result of the quality of the mount (and hence the quality of the contact and heat transfer).  I have had problems in the past with H100's where a particular orientation of the cpu block did not give good contact, and my temperatures spiked.  You can see the quality of the contact by looking at how evenly spread the TIM is on the cpu after you remove the cooler.  If it's evenly spread, you should have had good contact, if it's splorped off to one corner, that was bad contact.  It took several tries to get the temperatures down, and even mounting it multiple times in the same orientation gave varying results (this was on a 7 year old motherboard though, so warping was probably an issue).  In my case temperatures went from ~75°C to 35°C just from improved mounting.  

 

I would hazard a guess that it's not the TIM spreading that improved your temperatures, but the mounting of the cpu block.  

[SatansHollow]

This finding surprises me.  TIM has such good thermal conductivity that there should be no practical difference between any two compounds.  I could have sworn that LTT did a video on this not too long ago, but I can't seem to find it.  For many intel desktop cpu's, the best method of applying TIM is a vertical line.  Source: http://linustechtips.com/main/topic/4405-how-to-apply-thermal-paste-correctly/ http://www.arcticsilver.com/intel_application_method.html#as per arcticsilver's website.  Spreading the TIM can trap air bubbles, which diminishes the contact area and makes the heat transfer worse (air is worse than TIM for conducting heat [citation needed]).  

 

Commonly, large temperature differences are a result of the quality of the mount (and hence the quality of the contact and heat transfer).  I have had problems in the past with H100's where a particular orientation of the cpu block did not give good contact, and my temperatures spiked.  You can see the quality of the contact by looking at how evenly spread the TIM is on the cpu after you remove the cooler.  If it's evenly spread, you should have had good contact, if it's splorped off to one corner, that was bad contact.  It took several tries to get the temperatures down, and even mounting it multiple times in the same orientation gave varying results (this was on a 7 year old motherboard though, so warping was probably an issue).  In my case temperatures went from ~75°C to 35°C just from improved mounting.  

 

I would hazard a guess that it's not the TIM spreading that improved your temperatures, but the mounting of the cpu block.  

Im with you on this

[Ertman]

Seems to have been the fix for me, went from 78 degrees C to 63 degrees C... Guess the proof is in the science!

Glad it worked out.

It seems the what is the right way is not what is necessarily proper, but what works. I have manufacturer instructions for CPU coolers and thermal pastes that differ. I have also personally applied both ways, with no observable difference, spread was marginally warmer but not enough for it to beyond error. I wonder if the spread method makes up for any mistakes when applying the cooler, which the only thing that I have seen that seems to make a difference.

[Queek]

I wonder if the spread method makes up for any mistakes when applying the cooler, which the only thing that I have seen that seems to make a difference.

It would more likely introduce imperfections rather than remove them.  It would be nigh on impossible to have a perfectly flat surface from spreading, and those imperfections will by definition will trap air.  If you have a continuous line or blob, the pressure of applying the cpu block will force the TIM outward, but will not trap bubbles, as the TIM will displace the air as it is pushed flat.  

[Ertman]

It would more likely introduce imperfections rather than remove them.  It would be nigh on impossible to have a perfectly flat surface from spreading, and those imperfections will by definition will trap air.  If you have a continuous line or blob, the pressure of applying the cpu block will force the TIM outward, but will not trap bubbles, as the TIM will displace the air as it is pushed flat.  

That what I have always been told, and demonstrated. Clearly the OP did better with the spread method. I was just wondering if it would mask an improper mount.

[vSyNd1c4t3]

That what I have always been told, and demonstrated. Clearly the OP did better with the spread method. I was just wondering if it would mask an improper mount.

 

Reading around, seems the h100i has mounting issues sometimes, with the tubes pulling the cpu around while trying to get it screwed in place.

 

I would like to see a great visual video demonstration of the differences between glob and spread.

 

like a huge scale proportion with clear, flat, and thick glass. so you can see everything as it happens in slow motion.

 

see the results of what has more air pockets and such.