Die-electric grease to assist heat dissipation?

[FireDemon]

Kind of a random thought I've had.

In the past, I've packed a CPU socket with die-electric grease on a phase change setup to prevent condensation. I was wondering though if anyone has tried it with an above ambient cooling setup. Die-electric is often used in automotive applications for both corrosion prevention and heat dissipation (Ever had a bulb burn your socket out? Pack it with the stuff and I bet the problem goes away). Curious if the added heat dissipation through the socket/board and the other side of the CPU could potentially improve temps by a degree or two.

 

[Supercharged_Z06]

Most dielectric grease products are actually heat resistant while at the same time not being electrically conductive. They really aren't designed to conduct or transfer heat. Their main purpose is to keep moisture out of electrical connections to keep them from corroding... and to keep stuff like spark plugs from seizing. They are also usually designed to withstand/hold up to high temps (~300F).

Heatsink/thermal paste compounds on the other hand are specifically designed just for heat transfer... and they do a much better job of it.

I did some googling:
Dielectric grease usually sports a thermal conductivity rating that is quite low. Somewhere between .5 and 2.5 W/(m-K)
A good silver based thermal paste provides a thermal conductivity rating of 8 W/(m·K)
And the newer liquid metal based compounds for heatsinks have a thermal conductivity that's often greater than 13 W/(m·K)

So the bottom line is thermal paste will do a much better job of transferring heat away from your CPU... dielectric grease, not so much.

With a phase change setup or if cooling with liquid nitrogen, sure, I can see applying dielectric grease wherever you want to keep moisture buildup at bay... but not so much between the actually contact plate between the CPU and heatsink - more around it.

 

[Nobu]

Most dielectric grease products are actually heat resistant while at the same time not being electrically conductive. They really aren't designed to conduct or transfer heat. Their main purpose is to keep moisture out of electrical connections to keep them from corroding... and to keep stuff like spark plugs from seizing. They are also usually designed to withstand/hold up to high temps (~300F).

Heatsink/thermal paste compounds on the other hand are specifically designed just for heat transfer... and they do a much better job of it.

I did some googling:
Dielectric grease usually sports a thermal conductivity rating that is quite low. Somewhere between .5 and 2.5 W/(m-K)
A good silver based thermal paste provides a thermal conductivity rating of 8 W/(m·K)
And the newer liquid metal based compounds for heatsinks have a thermal conductivity that's often greater than 13 W/(m·K)

So the bottom line is thermal paste will do a much better job of transferring heat away from your CPU... dielectric grease, not so much.

With a phase change setup or if cooling with liquid nitrogen, sure, I can see applying dielectric grease wherever you want to keep moisture buildup at bay... but not so much between the actually contact plate between the CPU and heatsink - more around it.

Can't put those in the socket, tho...well, maybe the non-conductive stuff, but they would probably not allow the pins good contact in the socket.

So, the takeaway I guess is: Dieletric grease probably wouldn't hurt, but may not really help either. Especially considering copper-copper heat conductivity is pretty good already.

 

[FireDemon]

Most dielectric grease products are actually heat resistant while at the same time not being electrically conductive. They really aren't designed to conduct or transfer heat. Their main purpose is to keep moisture out of electrical connections to keep them from corroding... and to keep stuff like spark plugs from seizing. They are also usually designed to withstand/hold up to high temps (~300F).

Heatsink/thermal paste compounds on the other hand are specifically designed just for heat transfer... and they do a much better job of it.

I did some googling:
Dielectric grease usually sports a thermal conductivity rating that is quite low. Somewhere between .5 and 2.5 W/(m-K)
A good silver based thermal paste provides a thermal conductivity rating of 8 W/(m·K)
And the newer liquid metal based compounds for heatsinks have a thermal conductivity that's often greater than 13 W/(m·K)

So the bottom line is thermal paste will do a much better job of transferring heat away from your CPU... dielectric grease, not so much.

With a phase change setup or if cooling with liquid nitrogen, sure, I can see applying dielectric grease wherever you want to keep moisture buildup at bay... but not so much between the actually contact plate between the CPU and heatsink - more around it.

Agreed. Which is I always have to remove the ICMs from HEI style distributors and replace the die-electric grease that most manufacturers still haven't caught on about, with actual TIM where the module contacts the distributor base for heat dissipation. I've been stranded twice now because of that very issue before I figured out what the problem was.

I'm not talking about the IHS side of the CPU though. I'm talking about where it plugs into the socket. Die-electric is a rather poor heat transfer material compared to an actual TIM but it's still many times over a better conductor of heat than plain air is at ~.025 W.

It might not be the best conductor, but there's a reason it eliminates hot spots and burns when packed into bulb sockets.

 

[Supercharged_Z06]

Agreed. Which is I always have to remove the ICMs from HEI style distributors and replace the die-electric grease that most manufacturers still haven't caught on about, with actual TIM where the module contacts the distributor base for heat dissipation. I've been stranded twice now because of that very issue before I figured out what the problem was.

I'm not talking about the IHS side of the CPU though. I'm talking about where it plugs into the socket. Die-electric is a rather poor heat transfer material compared to an actual TIM but it's still many times over a better conductor of heat than plain air is at ~.025 W.

It might not be the best conductor, but there's a reason it eliminates hot spots and burns when packed into bulb sockets.

The only real drawback I see is the mess... and because it is not electrically conductive, like Nobu said, I could also see it acting as an insulator - contributing to and preventing good pin contact. Trust me, if it did work and helped with the heat, all us overclocking enthusiast whackos would be slathering that stuff all over our CPU sockets.

 

[cdabc123]

The only real drawback I see is the mess... and because it is not electrically conductive, like Nobu said, I could also see it acting as an insulator - contributing to and preventing good pin contact. Trust me, if it did work and helped with the heat, all us overclocking enthusiast whackos would be slathering that stuff all over our CPU sockets.

It has been slathered into and around sockets on sub ambient cooling builds. I would suspect it has a positive impact on temps although a small one. Next do we have to cool the back of the motherboard to make use of out improved thermal conductivity between the CPU and socket?

 

[FireDemon]

It has been slathered into and around sockets on sub ambient cooling builds. I would suspect it has a positive impact on temps although a small one. Next do we have to cool the back of the motherboard to make use of out improved thermal conductivity between the CPU and socket?

Yep I thought of that too

I've personally never had die-electric prevent good contact on....well, anything. Including a past CPU setup where it was packed into the socket. Of course at -20F, I can't exactly say whether or not it helped with heat dissipation.

I'm half tempted to try it on an old setup just for S&Gs. Anyone have a socket A motherboard laying around they'd care to donate to science? I have a Barton core that hasn't seen any excitement since high school.

 

[Nobu]

To be clear, I'm sure dielectric grease would not cause issues with contact. Other, more viscus tim (pastes especially) were more what I was referring to.