Overclock.net banner
Cancel
Create thread New Forums

Should I use thermal paste between the die and the IHS after delidding ?

Jump to Latest
1 - 20 of 29 Posts

L3gi0n

· Registered
Joined
·
44 Posts
Discussion starter · #1 ·
I have heard that I must use liquid metal between the die and the IHS after delidding and that if I want to use thermal paste between the die and the IHS, I shouldn't even delid.

I have also heard that using thermal paste between the die and the IHS causes the pump-out effect which is bad....is this true ?

I have an i5-6600K and an i7-7700K that I may want to delid but I don't wanna use liquid metal.

Do you recommend delidding and using thermal paste between the die and the IHS ? isn't the thermal clay that intel uses completely dried out by now? Will my temps improve if I delidd and use thermal paste under the IHS ?

If you recommend delidding, which thermal paste do you recommend for between the die and the IHS ?
 
#2 ·
From experience, the temps will improve for a short while, and when the new paste is pumped out it will quickly go to ****.
If you refuse to use LM, then I strongly suggest you leave the IHS alone, the stuff Intel use stock is way better at resisting pump out than traditional off the shelf thermal paste.
TLDR: Use LM or forget about it.
 
#3 ·
L3gi0n said:
I have heard that I must use liquid metal between the die and the IHS after delidding and that if I want to use thermal paste between the die and the IHS, I shouldn't even delid.

I have also heard that using thermal paste between the die and the IHS causes the pump-out effect which is bad....is this true ?

I have an i5-6600K and an i7-7700K that I may want to delid but I don't wanna use liquid metal.

Do you recommend delidding and using thermal paste between the die and the IHS ? isn't the thermal clay that intel uses completely dried out by now? Will my temps improve if I delidd and use thermal paste under the IHS ?

If you recommend delidding, which thermal paste do you recommend for between the die and the IHS ?
Click to expand...
Kyrosheet option.
 
#5 ·
I tested regular thermal paste GD900 (similar to MX4) versus PTM7950 (hybrid pad/liquid whatever it is) and Liquid Metal (LT-100)

So PTM7950 was waaaaay better than GD900, no comparison there. But GD900 still couldn't handle what LT-100 (liquid metal). I tested this back and forth several types with several applications. Prior to this I considered using PTM7950 on the die instead of LM, but now, now not anymore, only LM it will be.
 
#6 ·
massivex said:
I tested regular thermal paste GD900 (similar to MX4) versus PTM7950 (hybrid pad/liquid whatever it is) and Liquid Metal (LT-100)

So PTM7950 was waaaaay better than GD900, no comparison there. But GD900 still couldn't handle what LT-100 (liquid metal). I tested this back and forth several types with several applications. Prior to this I considered using PTM7950 on the die instead of LM, but now, now not anymore, only LM it will be.
Click to expand...
Sadly, LT-100 is the worst liquid metal I've experienced. Also, liquid is a bit much said to describe it.
 
#7 ·
ArneR said:
From experience, the temps will improve for a short while, and when the new paste is pumped out it will quickly go to ****.
If you refuse to use LM, then I strongly suggest you leave the IHS alone, the stuff Intel use stock is way better at resisting pump out than traditional off the shelf thermal paste.
TLDR: Use LM or forget about it.
Click to expand...
There is one more option: actual solder. It’s wildly hard to do though and may take a special solder formulation, so it’s effectively still true.
 
#8 · (Edited)
massivex said:
I tested regular thermal paste GD900 (similar to MX4) versus PTM7950 (hybrid pad/liquid whatever it is) and Liquid Metal (LT-100)

So PTM7950 was waaaaay better than GD900, no comparison there. But GD900 still couldn't handle what LT-100 (liquid metal). I tested this back and forth several types with several applications. Prior to this I considered using PTM7950 on the die instead of LM, but now, now not anymore, only LM it will be.
Click to expand...
It's a lot more fun and many times cheaper to make your own liquid metal. Just get 50g of gallium, 20g of indium, 10g of tin and mix it together in a crucible or an oven and you have instant homemade galinstan. Then you can experiment with trace amounts of Bismuth and Antimony to help increase oxidation resistance (no idea how much, but definitely not more than the amount of tin added--so, let's say 5g each, not sure if it's a good idea to go 10g antimony 10g bismuth and 10g tin, but if you get it cheap, why not become a mad scientist while getting great temps at the same time? :)). The hardest part is getting the mixture into a primary syringe for long term storage and then into a syringe that allows controlled application without too much coming out (the conductonaut syringes work well for this).
 
  • Rep+
  • Helpful
Reactions: MrFox and massivex
Save
Reply Quote Rep+ Rep+
#10 ·
Falkentyne said:
It's a lot more fun and many times cheaper to make your own liquid metal. Just get 50g of gallium, 20g of indium, 10g of tin and mix it together in a crucible or an oven and you have instant homemade galinstan. Then you can experiment with trace amounts of Bismuth and Antimony to help increase oxidation resistance (no idea how much, but definitely not more than the amount of tin added--so, let's say 5g each, not sure if it's a good idea to go 10g antimony 10g bismuth and 10g tin, but if you get it cheap, why not become a mad scientist while getting great temps at the same time? :)). The hardest part is getting the mixture into a primary syringe for long term storage and then into a syringe that allows controlled application without too much coming out (the conductonaut syringes work well for this).
Click to expand...
Whoazy. My first ever reply from Falkentyne, I'm honored.
Scientific testing and hobbies aside, would this yield even better results than commercially bought LM? I buy mines from Aliexpress, after testing the cheap LT-100 (3g / $10) vs Conductonaut, I didn't see a point to ever pay expensively again. I saw they also have on aliexpress 25g srynges with for $25~$30.
" Then you can experiment with trace amounts of Bismuth and Antimony to help increase oxidation resistance" , thanks for making me feel stupid I had to google up all that atimony and bismuth.
I don't know if I'd trust myself mixing chemicals that much, but as long as there shouldn't be much risk on the die itself then I don't see the harm in it.
 
Save
Reply Quote Rep+ Rep+
#11 ·
massivex said:
Whoazy. My first ever reply from Falkentyne, I'm honored.
Scientific testing and hobbies aside, would this yield even better results than commercially bought LM? I buy mines from Aliexpress, after testing the cheap LT-100 (3g / $10) vs Conductonaut, I didn't see a point to ever pay expensively again. I saw they also have on aliexpress 25g srynges with for $25~$30.
" Then you can experiment with trace amounts of Bismuth and Antimony to help increase oxidation resistance" , thanks for making me feel stupid I had to google up all that atimony and bismuth.
I don't know if I'd trust myself mixing chemicals that much, but as long as there shouldn't be much risk on the die itself then I don't see the harm in it.
Click to expand...
Conductonaut extreme apparently has extra metals in it to help preserve longevity but I'm not paying $25 for 1 gram to test that theory. I'd rather just have fun and mix stuff around myself, although it's rather difficult to get indium and gallium in proper sizes to test out mixing very small amounts, (while getting small samples of Tin shot and the other metals are easy), since they usually come in chunks. You would have to be a badass metallurgist to create properly sized samples from large chunks (which are cheaper per gram). I heard LT-100 isn't very good but I don't know how it's worse than basic galinstan, unless they went crazy on the indium content (it's the indium that hardens like that, IIRC Gallium by itself starts melting at 30C).
 
Save
Reply Quote Rep+ Rep+
#12 ·
Falkentyne said:
Conductonaut extreme apparently has extra metals in it to help preserve longevity but I'm not paying $25 for 1 gram to test that theory. I'd rather just have fun and mix stuff around myself, although it's rather difficult to get indium and gallium in proper sizes to test out mixing very small amounts, (while getting small samples of Tin shot and the other metals are easy), since they usually come in chunks. You would have to be a badass metallurgist to create properly sized samples from large chunks (which are cheaper per gram). I heard LT-100 isn't very good but I don't know how it's worse than basic galinstan, unless they went crazy on the indium content (it's the indium that hardens like that, IIRC Gallium by itself starts melting at 30C).
Click to expand...
You're absolute right I'm not paying $25 for that. Experimenting mixing your your own could be fun, I'll look into it! I actually only read good things about LT-100? But perhaps those could be the "normies" using it happy with their concurrent results. Have not seen seen anything explicit about it.
So if not LT-100, what's another cheap alternative? Thermalright SilverKing is also quite cheap, only a few dollars more expensive than LT-100 for 3g.
I'm also interested in buying those 25g syringes on Aliexpress.
So what are the instructions
  • 50g of gallium, 20g of indium, 10g tin
  • and then add trace amounts of antimony and bismuth?
 
Save
Reply Quote Rep+ Rep+
#13 ·
massivex said:
You're absolute right I'm not paying $25 for that. Experimenting mixing your your own could be fun, I'll look into it! I actually only read good things about LT-100? But perhaps those could be the "normies" using it happy with their concurrent results. Have not seen seen anything explicit about it.
So if not LT-100, what's another cheap alternative? Thermalright SilverKing is also quite cheap, only a few dollars more expensive than LT-100 for 3g.
I'm also interested in buying those 25g syringes on Aliexpress.
So what are the instructions
  • 50g of gallium, 20g of indium, 10g tin
  • and then add trace amounts of antimony and bismuth?
Click to expand...
yes maybe 5g of Bn and 5g of An
Then you need some way to heat the mixture and stir it so it gets mixed fully. The gallium will melt the indium as it lowers the melting point of everything.
 
  • Rep+
Reactions: massivex
Save
Reply Quote Rep+ Rep+
#14 ·
Alexandrus said:
Sadly, LT-100 is the worst liquid metal I've experienced. Also, liquid is a bit much said to describe it.
Click to expand...
Well obviously I'd value older member's opinion on this forum over some random redditors. I'm also not THAT experienced with liquid metal so if you have lots of experience with different once I'll take your word for it.
I actually recently ordered another 3g batch but oh well it's only $10.
So in the mean what's a good budget friendly alternative? How's Thermalright SilverKing sound? If so I'll order a batch and do a comparison on the direct die block.
P.S. You have the time or energy then please share your experience with LT-100.
 
Save
Reply Quote Rep+ Rep+
#15 ·
I would be also interested why LT-100 is bad. I had 3 different LMs in use (no name china ebay, Conductonaut and LT-100) and they behave the same, look the same and temperatures are the same. That china stuff was 5 years or older when I used it first. But still worked like it should.
 
#16 ·
X909 said:
I would be also interested why LT-100 is bad. I had 3 different LMs in use (no name china ebay, Conductonaut and LT-100) and they behave the same, look the same and temperatures are the same. That china stuff was 5 years or older when I used it first. But still worked like it should.
Click to expand...
Not talking about the heat transfer w/mk.
If there is too much indium in the mix, if gallium gets absorbed or the mix oxidizes, that indium is going to get as hard as a rock.
If you've ever applied galinstan to a nickel plated (NOT electroplated!!) surface, you will notice that that liquid metal will remain liquid for months. Because the gallium will remain in the alloy so it keeps its properties. Then you see the problems with the Iceman direct die block mentioned by that one user in the 14900K thread, because they use electroplated coating.
 
  • Rep+
Reactions: X909
Save
Reply Quote Rep+ Rep+
#17 ·
massivex said:
Well obviously I'd value older member's opinion on this forum over some random redditors. I'm also not THAT experienced with liquid metal so if you have lots of experience with different once I'll take your word for it.
I actually recently ordered another 3g batch but oh well it's only $10.
So in the mean what's a good budget friendly alternative? How's Thermalright SilverKing sound? If so I'll order a batch and do a comparison on the direct die block.
P.S. You have the time or energy then please share your experience with LT-100.
Click to expand...
Not to worry, I have 2 or 3 packs of 3g each, I was fooled by the marking as well, I only used it once to test against Conductonaut, Conductonaut Extreme, DC2 Pro and Thermalright Silver King, while these were quite close, the LT-100 was ... well, let's just say far away, sadly.
However, I only tested between the die and IHS, I would never used LM between the IHS and block or heatsink.
Silver King is quite good and consistent.
 
#18 ·
Alexandrus said:
Not to worry, I have 2 or 3 packs of 3g each, I was fooled by the marking as well, I only used it once to test against Conductonaut, Conductonaut Extreme, DC2 Pro and Thermalright Silver King, while these were quite close, the LT-100 was ... well, let's just say far away, sadly.
However, I only tested between the die and IHS, I would never used LM between the IHS and block or heatsink.
Silver King is quite good and consistent.
Click to expand...
That's great to hear I'll get a batch of SilverKing next time and when I got time try and mix my own for the fun of it.
Testing on the die is exactly what I'm looking for and it's what I tested as well. To be honest I didn't use the LT-100 for over a year, until just recently when I dellided a few CPUS and I was wondering 'is this how liquid metal is supposed to look like?', something just liked a bit off from what I remember compared to Conductonaut which I used previously.
 
Save
Reply Quote Rep+ Rep+
#19 ·
Falkentyne said:
Conductonaut extreme apparently has extra metals in it to help preserve longevity but I'm not paying $25 for 1 gram to test that theory. I'd rather just have fun and mix stuff around myself, although it's rather difficult to get indium and gallium in proper sizes to test out mixing very small amounts, (while getting small samples of Tin shot and the other metals are easy), since they usually come in chunks. You would have to be a badass metallurgist to create properly sized samples from large chunks (which are cheaper per gram). I heard LT-100 isn't very good but I don't know how it's worse than basic galinstan, unless they went crazy on the indium content (it's the indium that hardens like that, IIRC Gallium by itself starts melting at 30C).
Click to expand...
Digikey sells 20g of LM for $55 before shipping.

https://www.digikey.com/en/products.../chip-quik-inc/TC4-20G/15195126?s=N4IgTCBcDaIMwEYCsBaAKgYQCwrABgHEUA5AERAF0BfIA

However, the mad scientist in me will be trying your Galinstan method whenever I run out of this 20g tube.
 
#20 ·
massivex said:
That's great to hear I'll get a batch of SilverKing next time and when I got time try and mix my own for the fun of it.
Testing on the die is exactly what I'm looking for and it's what I tested as well. To be honest I didn't use the LT-100 for over a year, until just recently when I dellided a few CPUS and I was wondering 'is this how liquid metal is supposed to look like?', something just liked a bit off from what I remember compared to Conductonaut which I used previously.
Click to expand...
I am sure LT-100 is better than most regular TIMs when used between IHS and die, should be plenty for older CPUs such as the 7700K and 8700K.
 
1 - 20 of 29 Posts
About this Discussion
28 Replies
12 Participants
Last post:
Alexandrus
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling