I read the comments from Peter, Tom, and Rich. I agree that to play it
safe, one could either: avoid the configuration entirely, or move the
resonance off of the power line frequency. I also have to say that my
experience was not bad at all, I found them to be an interesting old
technology reapplied in a modern design.
-----
I have recently rediscovered saturable reactors too, another ancient EE
topic (Alexanderson, 1914, used them in his HF alternators for modulation
and control). Vacuum tubes, saturable reactors, resonant chokes, isn't
technology wonderful! Those Pentium kids don't know what they're missing.
-----
The caveats are mentioned in the comments that got posted: They bounce with
load change, hence for SSB they are not as good as a stiff C-input filter
with lots of stored energy. Damping depends on the Q of the ciruit, and the
other resonances that occur from the subsequent smoothing components that
catch the higher frequency ripple harmonics. Use of SPICE transient model
is perfect for this kind of ciruit. I used this resonant filter for
constant load FM operation to reduce the stored energy, and make the thing
so that if someone removed drive (easy to do in a wide power range
transmitter) that the plate meter wouldn't peg out and the voltage soar 30%
or more. I hated that in the old Gates transmitter I worked on once. The
FCC requirements for broadcast transmitter type notification require that
the plate voltage be metered at about 2/3 scale or close, not full scale or
< 1/2 scale. Analog meters are most accurate in the upper half of their
range, so I ran my meter high, and didn't have to worry about it pegging
when the driver was flipped off. The power supply was 'regulated' by that
filter choke.
As for operating near or on resonance, i rated thecomponents to handle the
peak voltage. And measured it ON RESONANCE. The Q of the ciruit was not
high, as the Rdc of the inductor is 10-20 Ohms.
The critical inductance to get constant current could just as easy be a
high resistance of a parallel trap at resonance. The empirical way works
too, of course every choke will swing a bit with DC bias current, so its a
guess as to which value of capacitance to use to park the resonance where
you want it with varying load currents. The way I did it was to send the
caps to the choke designer, when I had my model and prototype working, and
ask them to MAKE the choke do what I wanted, at one point.
Anyway, i want to point out that this is DANGEROUS stuff, as Tom made the
point well. If you just play around, and fire up one of these circuits,
without looking at currents, waveforms, temperture coefficients, effects of
load current, be aware that unpredictable results can occur. Components
must be rated for the high AC ripple across them. Treat this as you would
any design, think about all the worst case tolerences, failure scenarios.
Or just hook it up empirically, and be conservative, operating off
resonance. But it is not black magic, it is plain engineering, like
building RF power amplifiers. \
Maybe the Pentium jockies better stay at their keyboards for this one, out
of the lab.
John
K5PRO
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|