>The problem with a grounded grid or any RF amplifier that has a non-linear
>input impedance ie non-resistive is that the exciter is not loaded with the
>same impedance during the entire RF cycle. This introduces harmonic
>generation (which is not a serious problem because it will be filtered out
>later anyway) and intermodulation distortion. If the instantaneous V-I
>curve of this impedance instead of being a straight line (resistive) can be
>described by a polynomial and that polynomial contains odd order
>components ,it will produce odd harmonics and IMD.
>
>A low pass filter has no energy storage. The idea is that a band pass
>filter with a capacitor on the output of it connected across
>the non-linear cathode circuit (or grid in some cases). The capacitor in
>conjunction with the inductor make an AC energy storage system. The amount
>of energy stored in the system is related to the Q. A low pass filter has
>a Q of less than 1 (actually it is Q= center frequency/
>bandwidth=1/2). The reason for the capacitor being the at the output is
>because it is a short term constant voltage device. It and the inductance,
>tries to maintain a constant sinusoidal AC voltage between the cathode and
>ground during several RF cycles. If the capacitive reactance is 1/3 of the
>cathode's minimum instantaneous impedance it maintains a very good
>sinusoidal wave form for a single RF cycle (what is needed) and does fairly
>good for 3 cycles. This is why high Q is not necessary.
** True for very good, however, good enough for typical
transceivers/exciters is an input capacitive reactance roughly equal to
the cathode's minimum instantaneous ohms during the negative peak in the
driving cycle -- which translates to a Q of 2 - 3.
>The exciter sees
>a constant impedance over each RF cycle also.
>If the output component of the filter connected to the cathode is an
>inductor (Inductor output L or T or Pi-L) it will try to maintain a
>constant sinusoidal RF current thru the RF cycle. This will make the IMD
>even worse.
>
>Well, there is my simple explanation.
>
Excellent discourse, Bill. Tnx
> ... ... ...
- R. L. Measures, a.k.a. Rich..., 805.386.3734, AG6K,
www.vcnet.com/measures.
end
|