Transistor choice for switching inductive loads

Any such devices spring to mind?

SMAJxxA or P6KExxA would fit well enough for a start; the SMTs go up to SMC (3kW) and SMD (5kW) sizes, and the THT parts go up, well, as high as you care to spend on them (but if you need /that/ much, you should try to use a MOV instead..).

Make sure you get enough power (peak and average) dissipation, and current rating, for whatever it might have to handle.

Those are pretty much the jellybean parts, but everyone makes their own flavors too. Any will do, as long as the ratings are sufficient for the purpose.

Pick a nominal voltage rating that's higher than the worst case supply/load voltage, and pick a transistor that's rated for more than the worst case clamping voltage. Typical example is a SMCJ12A or 15 on a

For what it's worth, this works better than a clamp diode, for relays and solenoids and such. The higher (but still safely clamped) flyback voltage means quicker demagnetization, shaving off a millisecond or so of actuator time. Assuming that's a concern, of course!

Tim

I'm persuaded. The clamping voltage in my case would need to be 48V - 50V but I can't find any for this value range. Can they be wired in series? A couple of 24Vs to give 48V for example?

Den mandag den 8. februar 2016 kl. 19.07.08 UTC+1 skrev Tim Williams:

for fuel injectors and such something like this is used:

-Lasse

Series is okay, yes.

Parallel is kinda-sorta-okay too, but only for surges that really push things into breakdown, where the ESR is significant. If you have an application that experiences a lot of lower-energy transients, you may eventually toast the lesser volted part, even though a pulse of much higher energy might be safely sustained by both. So it would be an option for lightning protection, but not so much for clamping regular switching spikes.

Tim

Thank you, Tim. I'm going to scope the rail that's getting chopped later today. It's nominally 45VDC but I'm curious as to why the manufacturer chose a transistor that appears on the face of it to be so over-specified voltage-wise for the job - and yet it still failed. At any rate, I think your clamping suggestion would be a wise move.

Then perhaps the problem is over-current like the transistor coming out of saturation? Can you post a sketch of that part of the circuit with values etc.

piglet

My goodness, you're a helpful fellow, Piglet! Fortunately that won't be necessary. I spent around 2 hours trawling through my transistor stash this morning looking for any of the part numbers you kindly suggested. Sadly I found none. The nearest thing I could find was a BD139. I had no doubts about its voltage and power-handling capabilities, but was unsure whether or not the little BCY71 oscillator would be able to drive it without an additional buffer stage. Happily my fears were unfounded and the scope works fine again. Big thanks to Piglet, Tim and everyone else who chipped in here with helpful suggestions. :)