Ultrareliable transistor mountings

[Oznog]

I have been toying with the idea of replacing the diode battery isolator on my car with a MOSFET implementation. With a 160 amp alternator and some extremely high loads that thing can drop like 150 watts or more. It says it's rated for that high a current but the heatsink hardly looks sized for that. And the thing drops over 1v under load so the alternator is working harder than it should.

I'd like to use an op amp to drive a MOSFET- or several- to make a cooler running device with little voltage drop. The op amp is there so that we can just amplify the drop across the drain-source, if it's positive the MOSFET needs more drive. This can reduce the loss to something comparable to the op amp's offset voltage.

The thing I worry about here is they still need to be sinked, but a failure of the insulator would be catastrophic to say the least. This is a very high vibration, high heat environment. Fuses are quite large and expensive and it is very tricky to keep their resistance low. I know and love silicone pads, is there any truly bulletproof space-age insulator out there that would be better for such a critical app?

 

[awright]

Transistor Mountings

Consider mounting the MOSFETS directly onto a large heat sink, insulating the heat sink from the chassis with discrete, heavy duty insulators, and surrounding the heat sink with a shield to prevent accidental shorting to the chassis by tools and misplaced junk. The shield could be a screen, allowing free convective cooling, or a tube directing forced cooling air over the heat sink.

Direct contact with the heat sink using a heat conductive compound is always thermally superior to the best electrical insulating hardware.

awright

 

[Oznog]

Re: Transistor Mountings

Consider mounting the MOSFETS directly onto a large heat sink, insulating the heat sink from the chassis with discrete, heavy duty insulators, and surrounding the heat sink with a shield to prevent accidental shorting to the chassis by tools and misplaced junk. The shield could be a screen, allowing free convective cooling, or a tube directing forced cooling air over the heat sink.

Direct contact with the heat sink using a heat conductive compound is always thermally superior to the best electrical insulating hardware.

awright

Hmm... I'd initially discounted that idea because the heat sink needs to be pretty free and open, but now that you mention it, it wouldn't need to be as large since the reduced thermal impedance between the die and sink makes the sink more effective.

I might want to run some numbers on that one. Not sure how I would go about actually building such a shield though. It'd be nice to have a fan but such things aren't generally reliable for under-the-hood environments. Too much heat, dust, rain, and crap. I'd be nice to route the power into the cab and switch it there but the routing is way too far and it'd need a firewall port.

 

[zevon8]

You may be able to use one of the isolated tab FET's. Many of the standard type by International Rectifier come in an isolated version. Most have the same part number, just an " I " inserted, such as IRFIZ44.

Some are also available in stud-mount.