Bad thermal paste aplaction?

[cowpker4life]

I don't really know where to post this but anyways. Replaced thermal paste on my laptop went from 80-85c for 100%load on cpu using burnintest replaced with mx-4 thermal paste retested got 64c a few days later notic the fans are running pretty high rpms compared to what they were the other day check temps was running 80c for the cpu while gaming so closed the game down loaded up burnintest and see my temps are back to 80-85c. Question is the thermal paste just getting settled in bad application or some other reasons room temps have remained similar around 60-70f could high humidity have a big effect on cpu temps by chance maybe? A change in 10f should not be 15-20c change in temps at most 5-10c should be the diffrance. Just asking cause don't really wanna take it all apart again and don't want to have temps that high.

[D13H4RD]

How did you apply the paste?

 

For laptops, it depends on the size of the die. Don't apply too little but also don't apply way too much.

[cowpker4life]

3 minutes ago, D13H4RD2L1V3 said:

How did you apply the paste?

 

For laptops, it depends on the size of the die. Don't apply too little but also don't apply way too much.

Went with a line but temps should not incress upwards of 20c a few days after the paste has been changed it should only incress if worse paste or did not apply correctly within a little bit of testing not a few days later

[cowpker4life]

22 minutes ago, D13H4RD2L1V3 said:

How did you apply the paste?

 

For laptops, it depends on the size of the die. Don't apply too little but also don't apply way too much.

when i was changing paste i did notice a werid plastic cover that had acut out for the die it was on both my guess it is to keep paste from shorting something out so left it there might look about taking it off tho.

[markr54632]

Haha I have posted two topics on this now. What you are experiencing is the "pump out" effect. Basically on bare die applications this effect is greatly exaggerated. The thermal paste migrates from the center of the die to the edge. The thinner the paste the faster this happens. Mx-4 is one of the worst pastes you can use on a bare die application NT-h1 is the only one I have tried that pumps out faster.

 

Arctic silver 5 has held up for 6 months or so in my experience in laptops. I am currently trying IC diamond as it is advertised to not pump out.

[cowpker4life]

3 minutes ago, markr54632 said:

Haha I have posted two topics on this now. What you are experiencing is the "pump out" effect. Basically on bare die applications this effect is greatly exaggerated. The thermal paste migrates from the center of the die to the edge. The thinner the paste the faster this happens. Mx-4 is one of the worst pastes you can use on a bare die application NT-h1 is the only one I have tried that pumps out faster.

 

Arctic silver 5 has held up for 6 months or so in my experience in laptops. I am currently trying IC diamond as it is advertised to not pump out.

so pretty much repasting it will not do anything unless i get some paste that dose not "pump out" well damn. going wait until tomorrow see if temps drop tonight at all and then see how the spread is in the morning if temps are pretty bad. any idea on leaving or removing the plastic guard thing on the cpu/gpu that has the cut out for the die's. also if the die was roughed up a tad giving more friction for the paste to stick wouldnt that help? as there is a thin layer of i forget what covering the die and that is the shinny stuff you see. might just screw it and get some liquid metal paste but dont really know how to aplly it as i see people put it on the die and heat sink and laptop heatsinks are not the easiest to keep from sliding around and moving the paste making it a good chance of the liquid metal leaking over the edges.

[markr54632]

Please dear god do not try to "rough up the die" that is bad. and yes liquid metal will always be risky at best.

 

Leave the guard in place. It will make no difference, and can make application of liquid metal easier and less risky should you choose to be braver than I and try it. Liquid metal is the only thing that doesn't pump out. With non conductive pastes its a matter of finding the paste that pumps out the slowest. You are going to have to repaste fairly regularly.

 

85 c is not a bad temp for most performance laptops. Unfortunately your temps will just degrade rapidly with the mx-4.

 

If you just want to get factory performance back and not have to worry about it try ceramique. Its not the best performer, but lasts pretty good. If you want a good performance paste AS5 holds up for around 6 months and degrades slowly after this.

 

The two I have left to try are IC Diamond, which is currently on both of my cheaper laptops and Gelid GC Extreme.

 

The Gelid extreme supposedly pumps out rather slowly and holds up for a year or so.

 

The IC diamond has a few drawbacks, but is supposed to last the longest. It scratches the die, which I have experienced. It also supposedly stains the heatsink, which I have not experienced, but even if it does copper is copper. It doesn't matter.

 

Honestly, pick your poison and repaste with something other than mx-4 or nt-h1. If you can find a good article on the topic please send it my way as I have spent weeks researching this topic finding very little information, so I have had to do alot of testing.

[cowpker4life]

2 minutes ago, markr54632 said:

Leave the guard in place. It will make no difference, and can make application of liquid metal easier and less risky should you choose to be braver than I and try it. Liquid metal is the only thing that doesn't pump out. With non conductive pastes its a matter of finding the paste that pumps out the slowest. You are going to have to repaste fairly regularly.

 

85 c is not a bad temp for most performance laptops. Unfortunately your temps will just degrade rapidly with the mx-4.

 

If you just want to get factory performance back and not have to worry about it try ceramique. Its not the best performer, but lasts pretty good. If you want a good performance paste AS5 holds up for around 6 months and degrades slowly after this.

 

The two I have left to try are IC Diamond, which is currently on both of my cheaper laptops and Gelid GC Extreme.

 

The Gelid extreme supposedly pumps out rather slowly and holds up for a year or so.

 

The IC diamond has a few drawbacks, but is supposed to last the longest. It scratches the die, which I have experienced. It also supposedly stains the heatsink, which I have not experienced, but even if it does copper is copper. It doesn't matter.

 

Honestly, pick your poison and repaste with something other than mx-4 or nt-h1. If you can find a good article on the topic please send it my way as I have spent weeks researching this topic finding very little information, so I have had to do alot of testing.

the copper is copper is true but you might have to sand it down as it gets darker and turns to green forget what it called to tried anyways to care but it loses conditioned and pretty sure it loses heat transfer my dad has some as5 that i can prob get form him but i honestly will prob just go with some liquid metal as it gives best cooling and shouldnt pump out..every 6 months is not bad but would be nice to not have to do it ever. also its not the actually die it is screacthing but instead a protective layer forget where i heard this from but giving ic diamond has diamond in it not to surprised. also a lot of people repaste GPU dies for desktops without liquid metal and never hear anything about them having the so called pump out problem. but thanks for the info and i can not find a article on this really everyone says bad application reseed the heatsink or something along them lines.

[markr54632]

If it was a bad application it would have shown up immediately.

 

As for the copper, it doesnt degrade or oxidize the copper. It just leaves a darkish grey stain.

 

Yes you are scuffing a protective layer, however it is very thin. There is a very fine line between scuffed and non functional.

 

As for liquid metals, they perform fantastic. They also have potential to migrate. They are very conductive. They corrode aluminum. They are risky. There are plenty of tutorials on how to prepare your laptop for liquid metal to lower risks. Some use tape, some use lacquer, some use gum eraser, to prevent migration. 

 

Personally I actually move my laptops frequently, and not all of my laptops have copper heatsinks. So, not really an interest for me. I would feel much better using it on a delidded processor that spends most of its life stationary.

[cowpker4life]

14 minutes ago, markr54632 said:

Please dear god do not try to "rough up the die" that is bad. and yes liquid metal will always be risky at best.

 

Leave the guard in place. It will make no difference, and can make application of liquid metal easier and less risky should you choose to be braver than I and try it. Liquid metal is the only thing that doesn't pump out. With non conductive pastes its a matter of finding the paste that pumps out the slowest. You are going to have to repaste fairly regularly.

 

85 c is not a bad temp for most performance laptops. Unfortunately your temps will just degrade rapidly with the mx-4.

 

If you just want to get factory performance back and not have to worry about it try ceramique. Its not the best performer, but lasts pretty good. If you want a good performance paste AS5 holds up for around 6 months and degrades slowly after this.

 

The two I have left to try are IC Diamond, which is currently on both of my cheaper laptops and Gelid GC Extreme.

 

The Gelid extreme supposedly pumps out rather slowly and holds up for a year or so.

 

The IC diamond has a few drawbacks, but is supposed to last the longest. It scratches the die, which I have experienced. It also supposedly stains the heatsink, which I have not experienced, but even if it does copper is copper. It doesn't matter.

 

Honestly, pick your poison and repaste with something other than mx-4 or nt-h1. If you can find a good article on the topic please send it my way as I have spent weeks researching this topic finding very little information, so I have had to do alot of testing.

for liquid metal i think it might be best to take the protive plastic off clean all the paste away then i see people put tap right up next to the die gernally 4 peices 1 one each side then if any gets over the edge when you pull the tap off it isnt there and all of it is on the die my biggest problem would be putting a little bit to much or not enough as I wouldnt want to put any on the heatsink its self but very video of people using liquid metal they put it on the heatsink and the die so theres is 0% room for error when putting the heatsink back on.

[cowpker4life]

1 minute ago, markr54632 said:

If it was a bad application it would have shown up immediately.

 

As for the copper, it doesnt degrade or oxidize the copper. It just leaves a darkish grey stain.

 

Yes you are scuffing a protective layer, however it is very thin. There is a very fine line between scuffed and non functional.

 

As for liquid metals, they perform fantastic. They also have potential to migrate. They are very conductive. They corrode aluminum. They are risky. There are plenty of tutorials on how to prepare your laptop for liquid metal to lower risks. Some use tape, some use lacquer, some use gum eraser, to prevent migration. 

 

Personally I actually move my laptops frequently, and not all of my laptops have copper heatsinks. So, not really an interest for me. I would feel much better using it on a delidded processor that spends most of its life stationary.

liquid metal shouldnt move much tho after it settles into place tho right? as once the heatsink is all the way down it should move very little as it would be on the microscopic level along with once the paste is all flattened out by spreading/pressure of the heatsink it shouldnt move much either right?

[cowpker4life]

dont know how legit this is but quick glanced through it and they seem to have found a way to speed up testing the pump out effect

https://www.researchgate.net/publication/3899700_An_accelerated_reliability_test_method_to_predict_thermal_grease_pump-out_in_flip-chip_applications

[markr54632]

Honestly, I dont know. I have very little experience with liquid metal. I can tell you all of the laptops I have experienced are low pressure mounts. There is a very good laptop review forum where you can find answers to all of these questions.

 

Having a conductive fluid on a portable item is not for me.

 

I have actually encountered that article. Thank you though. Another good one can be found by googling this bare-die-testing-a-delidded-3770k-an-h100-and-9-different-tims I dont want to link another forum as its usually not acceptable. It is an old article and the pictures are gone, but informative.

[cowpker4life]

2 minutes ago, markr54632 said:

Honestly, I dont know. I have very little experience with liquid metal. I can tell you all of the laptops I have experienced are low pressure mounts. There is a very good laptop review forum where you can find answers to all of these questions.

 

Having a conductive fluid on a portable item is not for me.

okay. might have to get some clear non-conductive nail polish to cover some parts of the stuff. my cpu i belive it is has no visible transmitters on the pcb what surprises me. reading around still and seeing what people say in one tread someone has put LM on 3 differnt laptops and has not had that problem yet

[cowpker4life]

hearing CLU is the best idea for thermal paste as it has the highest viscosity out of all of all the LM at the time of that post there might be a better one by now but who knows

[cowpker4life]

1 hour ago, markr54632 said:

Honestly, I dont know. I have very little experience with liquid metal. I can tell you all of the laptops I have experienced are low pressure mounts. There is a very good laptop review forum where you can find answers to all of these questions.

 

Having a conductive fluid on a portable item is not for me.

well i have some nail polish and CLU thermal paste on the way will be here in around a week lol wish me luck

[cowpker4life]

7 hours ago, markr54632 said:

If it was a bad application it would have shown up immediately.

 

As for the copper, it doesnt degrade or oxidize the copper. It just leaves a darkish grey stain.

 

Yes you are scuffing a protective layer, however it is very thin. There is a very fine line between scuffed and non functional.

 

As for liquid metals, they perform fantastic. They also have potential to migrate. They are very conductive. They corrode aluminum. They are risky. There are plenty of tutorials on how to prepare your laptop for liquid metal to lower risks. Some use tape, some use lacquer, some use gum eraser, to prevent migration. 

 

Personally I actually move my laptops frequently, and not all of my laptops have copper heatsinks. So, not really an interest for me. I would feel much better using it on a delidded processor that spends most of its life stationary.

well repasted again with mx-4 and prob will 1 or 2 more times until i get the new paste just to keep temps down kinda for the mean time back to 69c or so max in 100% load much better then the 80c although gaming still 80c as gpu/cpu are in the same heatsink because laptops. thanks for helping me with this problem ill let you know how the LM gose but every fourm is saying get the thickest thermal paste pretty much for raw dies as it will not have the pumping effect nearly as bad.

[markr54632]

You never did say what laptop have. Make sure the heatsink fits well before trying the liquid metal. A piece of carbon or pressure paper should let you know. If it fits well and makes even contact the liquid metal will be a good oOrion. If it doesnt a thicker compound will perform better.

[cowpker4life]

On 9/16/2017 at 0:18 PM, markr54632 said:

You never did say what laptop have. Make sure the heatsink fits well before trying the liquid metal. A piece of carbon or pressure paper should let you know. If it fits well and makes even contact the liquid metal will be a good oOrion. If it doesnt a thicker compound will perform better.

acer predator g5 793 72nz same heatsinke and chassy as the g9 just dosnt have a sub the drive/added cooling fan and lower end gpu alog with no rbg keyboard stuck to only red and most likely a slower refresh rate or low res display then the higher end g9s

[Erfsh]

Very interested in where your experience reveals later. Keep us up to date with the CLU, if you ever do it.

[cowpker4life]

On 9/19/2017 at 3:19 AM, Erfsh said:

Very interested in where your experience reveals later. Keep us up to date with the CLU, if you ever do it.

well its going be here in a day or two and should be done by Friday ill test then then a week or so later as mx-4 would drop temps by 20c on cpu maybe a few on both gpu and cpu under 100% load but it needs a bigger heat-sink/fin stack to get much cooler then it can with stock paste for cpu and gpu. but a few days after the mx-4 temps on cpu were up to 80c like stock so hopefully it will stick around the 60c mark with CLU but who knows ill temp test with both cpu and gpu under 100% load to see if the CLU makes a difrance there but dont think it will as the cooling setup cant dispate more heat then it currently is able to.

[cowpker4life]

On 9/20/2017 at 9:07 AM, cowpker4life said:

well its going be here in a day or two and should be done by Friday ill test then then a week or so later as mx-4 would drop temps by 20c on cpu maybe a few on both gpu and cpu under 100% load but it needs a bigger heat-sink/fin stack to get much cooler then it can with stock paste for cpu and gpu. but a few days after the mx-4 temps on cpu were up to 80c like stock so hopefully it will stick around the 60c mark with CLU but who knows ill temp test with both cpu and gpu under 100% load to see if the CLU makes a difrance there but dont think it will as the cooling setup cant dispate more heat then it currently is able to.

well i got the LM a week after it was on campus bc they lost it then found it and came here end of the day but havnt seen temps above 70c with intel stress test and heaven up at the same time and dif was a good idea to use nail polish or something as i did f**k up a little on the LM and got some on the pcb of the chip but had that layer in between so works like a champ. there brushes kinda suck to spread the paste and its really pretty thick but over all not to bad. overwatch max temp i saw was around 73c or so so far. under 100% fan 60-65c. if osmeone is going to do this i suggest getting a kinda decant paint brush from a store to spread it with tho because there brushes the bristle all stick out and do not clump up.