> Such a large unit is a bit cumbersome.
> Instead of a rod, could the filament choke be wound on a
> ferrite toroid?
> It would have less flux leakage, be more compact.
I've done that on very large amps.
> How large should the toroid be?
Depends on the core material selected. You have to look at
flux density at PEAK cathode current and the number of
bifilar turns. One thing about the open ended rod is the air
gap in the flux path is so large it is very difficult to
saturate the core.
> Since it is wound bifilar, the DC magnetic fields of the
> two windings
> would cancel. So could I use a say one-inch dia. toroid
> for a 3 A
> filament current (for a GS 35 B)?
The AC currents for the filament cancel, DC currents from
the cathode do not. It depends on the peak ampere-turns and
the core material. I think a 2" diameter 1/2 inch thick 43
or 65 material would be OK with a single GS-35. Also there
is the consideration of peak RF voltage from cathode drive
across the common impedance of the windings. This produces
an RF current that can be quite high on lower bands. It
opposes the flux generated by the cathode current on
negative RF voltage cathode swings, but dominates on
positive swings of RF cycle when the plate current reaches
zero amperes. Flux density is not symmetrical through the
entire RF cycle because of the cathode current.
Many people make a mistake with the GS-35 that adds hum to
the RF envelope. They use a bifilar filament choke and take
the cathode from one side of the winding, or they use a
center tap on the transformer. This superimposes power line
AC on the bias. A very small amount of AC drop across the
choke (or worse yet AC from the CT of the transformer) bucks
and aids the bias with every power line cycle.
The choke, to have a clean amplifier with no envelope
ripple, has to use a trifilar winding or a separate
cathode-pin choke that does not carry filament current. The
separate winding that carries no filament current has to be
the cathode DC path. It can be small gauge wire.
73 Tom
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|