Soldering to glass

Soldering HV components to a ceramic heatsink with indium is a fascinating idea. For those who often run things hot, its low 150C limit may be an issue, but alloys may be in order.

You can indium-solder to glass and ceramics. It's much easier if you have an ultrasonic soldering iron. Those cost kilobucks unfortunately, but they sure are fun.

Cheers

Phil Hobbs

Another approach (which I have used) is to fire a patch of silver or silver/palladium thick film ink onto an alumina substrate and then solder onto that metallic coating.

John

Seems to me that epoxy would work as well, maybe better. If the parts are flat, epoxy with some compression would flow out to make a very thin layer. The glass or ceramic may well dominate theta anyhow.

DPAK thick-film resistors are great for heatsinking.

With what? I'd expect that to fall apart when you ablated the binder.

The indium solder technique is hermetic, which helps in many cases.

Cheers

Phil Hobbs

A possible concern, is that indium forms a transparent oxide that is conductive. How does one debug a HV circuit when conductive tracks on the board aren't carbonized, but... transparent and invisible?

The thick-film inks are a blend of powdered glass and metal particles, so when the ink is fired the glass particles fuse to the glass binder in the alumina. (Most commercial alumina is made by sintering a mixture of alumina with a small percentage of glass.) The final result is a mostly metallic coating strongly bonded to the surface of the alumina. Tin/lead solder wets it nicely when the metal is silver or silver/palladium. The glass component means that it can be hermetic. I used this technique in an implantable medical application a long time ago. Indium is nice, but as already mentioned it is unsuitable for high temperatures.

John

So does tin.

John

The vapour pressure of indium is insignificant at soldering temperatures. It boils at around 2000 C.

Cheers

Phil Hobbs

Okay, so it's basically a mixture of metal and frit, like cermet. Good to know.

Cheers

Phil Hobbs

Right, who wants metal melted under their high-voltage resistor?

Use epoxy.

What about Gallium? Doesn't that wet glass too?

George H.

Gallium is really good at making other metals brittle by grain boundary attack.

John

I second this. I've soldered onto glass without exotic solder using an ultrasonic soldering iron. Didn't even use flux.

Yup. It's liquid near room temperature, though, and is pretty reactive--it gets dross all over it.

Cheers

Phil Hobbs

I think alumina can serve as an insulator, while improving thermal characteristics. Simple copper-on-alumina fabs are already available.

Glass? Who needs it?

RL

AlN conducts heat almost as well as aluminum alloy, and it's an excellent HV insulator.

Porcelain-on-steel is a common heavy-duty circuit substrate; i've got a few relay boards in my junk box of that construction. It also tolerates environments (like oil baths) that few polymers can handle, and has excellent heatsinking capability.

Available -- yes, one can buy DBC boards, but it might be too expensive for one-offs.

It is dirt cheap and easily available.

Best regards, Piotr