> Pappenfus et al write about the length of cable adding to
> the input tuning such as to make a complete number of
> electrical half waves back to the plate of the driver
> tube: the tube acting as a constant current source. under
> those conditions, the input SWR isn't so important, except
> when using a solid state driver. I would not be surpised
> if the length of coax was chosen with this in mind.
That theory absolutely does not work.
That theory would only work if the tube acted as a constant
current source and if the impedance matching was a broadband
transformer. Unfortunately the tube does not look like a
current source, and the tank is frequency selective. The
tube is actually a non-linear time-varying resistance. The
tank a low pass filter.
Tank circuit Q smoothes the tube resistance into a mean
resistance value at the operating frequency. At harmonics
the output reactance in the tank is the primary termination.
If the PA uses a pi-L, any harmonics are terminated into a
moderate to high inductive reactance. If the network is a
pi, the load is capacitive and very low reactance.
The harmonics simply terminate in the coax and whatever the
tank output reactance is at the harmonic frequency. This
reflects back on the tube cathode (in a GG amp) as an
impedance in series with the cathode, and so cable length
can greatly affect IM performance and efficiency of the PA
if the PA has inadequate input filtering. An SWR meter will
show the harmonic energy as imperfect SWR, even though the
actual SWR is not changing!
The optimum termination at the cathode depends on several
factors but as a general rule we want an input circuit that
presents a LOW impedance to the cathode for all harmonics of
the drive frequency. This means a moderate Q low pass C-L-C
pi-network or parallel resonant tank at the cathode.
All that stuff about certain length lines and planned phase
shift is nonsense since the impedance and phase varies with
the type of equipment driving the PA and how it is
adjusted....and it is never the same on harmonics as it is
on the fundamental. The real trick is to stop the harmonics
at the tube, then the cable length doesn't matter a bit and
the PA works the same with ANY driver.
73 Tom
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|