Greg Mayman wrote in a message to Roy J. Tellason:

 -=> Roy J. Tellason said to Greg Mayman
 -=> about "VEHICLE LED'S" on 09-12-03  04:06.....

 RJT> You're talking about something like a resistor from base to collector?
 RJT> As simple as that?

 GM> HA! I wish!!!! But that would only make the transistor act like a
 GM> diode, more or less.

 GM> No, what I am talking about is a driver circuit that will start
 GM> putting enough base current into the transistor to saturate it over
 GM> the whole half-cycle of the input wave. Often this is a square wave
 GM> drive to the base, as being the easiest to generate.

Ok,  how would you do that?

 GM> And this circuit will also require power to run it and as a source 
 GM> for the base current, usually by stealing some from the main 
 GM> supply.

 RJT> They'd make one heck of an isolator,  from the alternator at least --
 RJT> either off or with a very low voltage drop...

 GM> True. And since the drop is lower, the heat generated is less.

Yes,  though I'd still heatsink them.

 GM> OTOH, it is a lot more complicated than a diode. And the diodes
 GM> seem to do an acceptable job in this application.

It would seem to me that at the high current levels (tens of amps) they're
wasting a bunch of power as heat,  and the voltage being dropped a bit is
also going to cause the charging rate to be just a little slower.

 RJT> I have a board here that has a bunch of power FETs on it,  gotta dig
 RJT> up some specs somewhere.  (Grabbing board)  There seem to be four
 RJT> marked K707,  and five marked K799.  They're all mounted to heatsinking

 GM> Nope, that means nothing to me. I can't keep track of these part 
 GM> names these days, and I can't afford to buy the books that would
 GM> tell me :-(

I used to be able to acquire the occasional data book here and there,  but
haven't in years now.  It's probably a good thing,  as the ones I do have
take up a bunch of space now!  :-)  This stuff is also coming out on cdrom,
 and may be downloadable as pdf files,  I haven't looked yet.

I'm thinking those part numbers are actually "2SK707" and
"2SK799",  that sound any better?

 RJT> I figured it was worth taking out of the unit it came from for the
 RJT> power FETs at least,  anyhow.

 GM> Yep, I agree. Once you find the specs of those parts, you'll find
 GM> uses for them.

 RJT> with _ten_ 12v batteries in it) and I'm guessing the board with the
 RJT> power FETs was the inverter... 

 GM> That's very likely! FETs are very good as drivers in an inverter as
 GM> they have a lot faster switchoff than bipolars and this means lower
 GM> dissipation in the switches.

 GM> IIRC those losses are due to reverse currents flowing during the
 GM> switchoff time, or was it more complicated than that.....

There were some of those parts that also seemed to have diodes across the
gate-source connection,  though not all of them.  I've seen some of those
parts that seem to have something of the sort built in,  too,  though I
can't recall just why.  Maybe to prevent breakdown at the gate?

---