Heat sink interface material

[mark0310]

Hi,I have also posted this enquiry to the Polymers forum because they may be equally able to assist....there may be a better way of doing this?
My issue is that I have an application that requires an electrically insulating (breakdown voltage better than 2000 vac)material with excellent thermal conduction (better than 0.5 C-in2/W)to be used between a long pcb 150mm x 25mm and it's heat sink (has same mating area).We are not set up to use applied greases, also the PCB has European voltage to both sides, so I have looked at the usual options of thin (circa 0.15mm) adhesive backed materials (Polyimide, Silicon) from suppliers such as Kunze, Bergquist, Chomerics and Aavid but the one big problem I have is that I cannot maintain a pressure between the pcb and the heat sink, so with the adhesives I am reliant upon the glue strength of the adhesive. The performance figures from the suppliers all show that the thermal conductivity is much lower when there is little pressure applied.I have no knowledge of phase change materials, if I can get this insulation to behave as for the polymer (polyethylene??)from a glue gun then after application (heating up and cooling down in a low temperature oven)the bond strength between the pcb and heat sink should be good. I've enquired of 2 of the suppliers and they do market phase change materials, but they're soft at room temperature....so no good...any suggestions?

 

[jimbofitz]

I have seen ceramics used in this application for both dielectric properties and heatsinking capability. If the ceramic substrate is metallized with copper/nickel on both sides it can be soldered to eliminate the epoxy problem. I don't have any exact details but I believe this method is popular in IC packaging.

Hope this helps.

 

[IRstuff]

I don't get what's holding the board to the heatsink right now. Seems to me that if you have no means of attachment, you need to respin the board and modify the heatsink.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[mark0310]

Jimbo...thanks, I hadn't considered this soldering idea....maybe a practical problem given length of pcb.

IRStuff....You're mostly right, the fixture of the PCB to the heatsink is a problem...what should resolve this is if I can achieve adequate adhesion between the insulating material and the PCB/Heatsink, the majority of adhesive backed materials that are available unfortunately are very pressure sensitive....I'm hoping there's a melt adhesive variant...but haven't been able to locate one myself.

 

[IRstuff]

But, what's holding the board to the heatsink right now?

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[Compositepro]

Do not rely on an adhesive bond to hold a heatsink in place. Even solder joints fail due to the stresses of thermocycling.
Heatsinks are almost always held in place with mechanical connectors that maintain contact pressure but allow slight movement. Thermal greases and pad are simply used to fill the very tiny gaps between the contacting surfaces. Almost anything in these gaps is a better conductor than air.

How do you expect a hot melt adhesive to work on parts that get hot?

 

[mark0310]

IR Stuff: I need the adhesion between the PCB and Heatsink to hold them together, in assembly I can use a longitudinal clamp to fix their location but this would be a temporary measure as the adhesion cures....

Compositepro:I need to rely on the adhesion, the application is using the heatsink to keep the pcb as close to room temperature as possible because it has in service heat sensitive components upon it, max 50C, and the pcb actually generates very little power as thermal loss, 3 watts, so we're not worried about thermal cycling....we just cannot find a manufacturer of a phase change insulator that would give a good bond in the 20 to 50C range.....It doesn't seem too much to ask?

 

[lumenharold]

How about this stuff? Thermally condustive electrically insulative epoxy:

http://www.arroyoinstruments.com/manuals/ArroyoThermalEpoxy.pdf

Harold
SW2011 SP1.0 OPW2011 SP1.0 Win 7 Ultimate
BOXX 8550 Xtreme Dual Xeon 5680 @4.2Ghz
nVidia Quadro 4000

http://www.lumenflow.com

 

[mark0310]

lumenharold; Thanks very much, you're right this would work, unfortunately we're looking at 5000+ examples so hoping there's a solution that's free of variables.

I was hoping there would be a heat activated phase change material that would double as a strong adhesive. I just cannot find one!

Thanks again.