Pro Audio Design Forum

A simple spring clamp should provide adequate thermal bonding and won't risk damaging the devices. Even power transistors on big power amps use spring clamps nowadays not screws.

JR

mouser has these. http://www.mouser.com/ProductDetail/Wak ... K7i5qEg%3d Pretty pricey. If you use brass it would be quick to cut even with a hand saw.

Maybe a piece of 16 ga solid core bent tight around them. That would be quick to try and to do if it worked.

I was thinking something more like this


this is for larger devices but I suspect it could work. http://uk.rs-online.com/web/p/heatsink- ... s/1311238/

JR

I may make something like this from Paul Kemble's site:


Feranti E-line thermal coupling from the Practical Electronics Gemini

This is what Paul suggested. These are the same "e-line" transistors as the ZTX851.
They're flat on both sides with beveled edges.

Source: http://paul-kemble.tripod.com/sound5l.html

I don't see an image just a placeholder from tripod

JR

The embeded image was a gif so I captured it and made it a jpg.


Feranti E-line thermal coupling from the Practical Electronics Gemini

I would like for at least one of the metal pieces to be thin enough to flex and not over stress the epoxy packages.

Modern power amps use spring clamps on power devices for decent thermal coupling and very high reliability. I actually did a bunch or research into this as 2 screw metal devices were phasing out and being replaced by plastic devices and spring clamps in large scale manufacturing.

In the factory we had lots of quality issues with screws... (low torque, cross threaded, etc), modern reliability using springs is quite good.

JR

I was thinking of using SilPad in the sandwich.

Maybe some spring steel... That flexes,,, theses aren't moving any power to speak of, so you mainly want good constant tension.

Do whatever is handy.

JR

I've used a small piece of Cu cut from pipes to clip the transistors together, gunking with heatsink compound and finishing of with a bit of heat shrink for quick & dirty applications.

The application was a ribbon mike 'cloudlifter' type device for ribbon mikes where it was necessary to keep Vbe matched.

Worked well.

For http://www.aes.org/e-lib/browse.cfm?elib=3941 which required rather better thermal match, the devices were clamped between two small square Aluminium rods with more heatsink compound. Can't remember if we covered the assembly with thermal insulation.