Bob's statement is perfectly correct because he doesn't suggest to
add an extraneous circuit but simply redrawed the PI circuit in another
form (the most didactical one in my opinion).
A PI network can be seen as two back to back L networks matching
a common intermediate impedance between them.
The gap of the L network is that given the loads, only a solution exists,
thus the Q can't be selected but comes out just (Q is equal to the square
root of (Rp/Rs)-1).
Obviously, 2 L networks connected back to back overcome the single
L network limitation that would give no chance of Q selection.
Developing the algebrical sum of the components of the 2 L networks it
comes out the known and desired PI network configuration.
73,
Mauri I4JMY
-----Messaggio Originale-----
Da: Richard W. Ehrhorn <w4eto@rmi.net>
A: <realex@flash.net>; <amps@contesting.com>
Data invio: venerdì 13 agosto 1999 19.23
Oggetto: Re: [AMPS] Pi-Net math
>
> Hi Bob...
>
> With all due respect; your statement below appears to me to be incorrect.
The addition of an "L" network to the output side of the plate pi adds
another pole and typically increases harmonic attenuation by about 12-15 dB
at the second harmonic, somewhat more for higher orders. Achieving that much
improvement by increasing loaded "Q" of the pi network, if doable at all in
practice, typically would result in excessive circulating rf current, place
exorbitant requirements on tank components to avoid overheating, and
increase tank losses significantly.
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|