Overclock.net banner

1 - 4 of 4 Posts

·
Registered
Joined
·
1,074 Posts
Discussion Starter #1
I'm going to delid an 8086k, and have two ways to go about doing it: I can use liquid metal with the stock Intel IHS, or I can use non-liquid metal TIM with a custom copper IHS since liquid metal will react with copper. I'd rather go with using non-liquid metal TIM with a custom copper IHS, but am worried about "pump out" occurring. This is when the rapid thermal cycles cause the TIM to expand and get "pumped out" from atop the chip, causing temperatures to rise. I've read cases of this happening with the stock Intel IHS, but has anyone had good long term results with a custom copper one?
 

·
Registered
Joined
·
202 Posts
People who get pump out did it wrong. You need to brush on your liquid metal and brush until the liquid metal grips to the die covering the whole surface and the same on the are above the due on the IHS. Making sure you haven’t put on too much liquid metal it should only be a thinner brushed layer not balls and pools.

8086k is a super easy chip to apply liquid metal to its its not like haswell with all the surface mount part that no one applied conformal material to.
 

·
High Clocker
Joined
·
3,440 Posts
Your using way too much.
Iv used LM with copper for ages and had no problem. Its alloy you cant use it with. Those copper IHS are a waste of time tho.
Stock IHS is nickle plated copper. So at most youll gain 0.2c your better off going naked die if you make adapters for your setup.
 

·
Iconoclast
Joined
·
30,437 Posts
He's not talking about using liquid metal TIM, but more conventional TIM after a delid.

Anyway, this is going to depend on the quality and viscosity of the TIM. In general, the thicker/more viscous the material, the more resistant it is to pump out, and that's one of the reason the stock TIM is more properly called a putty than a paste.

Stock vs. custom IHS won't matter much.

Also, liquid metal TIMs do work with copper; they will stain it, but there won't be a harmful reaction that hurts performance over time or eats away at the metal. It's aluminium you really need to avoid with LM TIM.

If you still want to avoid LM for whatever reason, expect a smaller temperature improvement and select something that last forever...Prolimatech PK-3, TC Quantum, and higher conductivity Dow Corning or Shin-Etsu stuff perform well and won't be pumped out. Plenty of other TIMs will also be viable, at least over the mid term (a few years) and may well perform better for that time. The only TIMs that are categorically unacceptable are those with a lot of metallic fillers or 'watery' ones.
 
1 - 4 of 4 Posts
Top