> Are you sure about that switchmode of operation. I initially assumed it was
> switchmode because all my years of designing regulators for large AC
> generators were switchmode (or phase controlled rectifier based)
> designs. But the
> cursory schematics do provide me with a high degree of confidence
> that it isn't a linear regulator.
It can be hard to tell the difference just looking at the circuit. But think of
it this way, the rotor resistance is less than 2 ohms ... how much power would
the regulator dissipate if it were a linear device working into that kind of
load ? All it takes to make it oscillate is for the frequency response of the
regulator to be faster than the response of the alternator, and the alternator's
response is very slow. It's actually hard to make this kind of circuit NOT
oscillate !
> I wouldn't worry about the regulator. Any switchmode regulator will have,
> and realistically any linear regulator should have, a freewheeling diode to
> carry the rotor current when the transistor switches off (unless you
> are doing
> negative forcing but that's getting a bit esoteric).
I'm not worried about the inductive kickback from the rotor; I agree the
regulator has to be designed to handle that. BUT, while current is flowing
through that diode, the alternator is still generating power and that power has
nowhere to go except running up the system voltage.
> When you take your alternator into Auto Zone for testing do they connect it
> to a battery during the test?
No, it's a dummy load (resistor). But they don't run them open-circuit.
> I would be a highly irresponsible design of an alternator that could not be
> run open circuit like that.
>
> But we are talking about Lucas.
Not just Lucas ... all car alternator makers (AFAIK) warn against running them
open-circuit. Certainly Delco does.
Auto makers are extremely aggressive on unit price; they literally squeeze out
every last penny. (EE friend of mine swore he would never work for a car maker
again, after Ford made him completely redo a design to replace a $.09 capacitor
with a $.06 one.) If something costs money, and isn't absolutely required for
normal operation, it doesn't go in.
Randall
=== This list supported in part by The Vintage Triumph Register
=== http://www.vtr.org
|