[Lowfer] The real filter deal!

[Bill Ashlock]

Filter friends,

I think I finally figured out what is happening here! The filter must loaded 
with it's design load resistance over a wide range of frequencies or the 
first L and the shunt C to ground form a series resonant circuit that 
effectively short-circuits the PA. Try this:

1) Set up a simulation, or bench circuit using the famous 63uh-18nf-63uh 
filter with a 50 ohm load and a generator with a 1 ohm source resistance. 
The frequency plot of voltages and currents will look reasonable.

2) Remove the 50 ohm load and watch the input current increase to the same 
numeric value as the source voltage at the resonant frequency of the filter, 
IE: 7A if the input is 7V!  (i = v/1ohm).

3) Re-attach the 50 ohm resistor but in series with the resistor add a 
series resonant circuit composed of an L of 5mh and a C of 140pf (This 
represents either a vertical antenna or a reflected-back loop load via the 
stepdown transformer). The current again goes out of sight. Why? The filter 
is not seeing a 50 ohm load over it's range of input frequencies.

In general, if the resonant frequency of the 63uh/18nf pair is close to the 
resonant frequency of the 5mh/140pf pair some real strange things happen to 
the voltages and currents and I'll guarentee this is what is causing all my 
bench testing problems! If the resonant frequency of the filter is changed 
to at least twice the resonant frequency of the load things begin looking 
fairly good (IE: descent filtering of the upper harmonics and normal current 
into the antenna load), but there is still a nasty peak at the resonant 
frequency of the filter that could cause some undesirable radiation if this 
falls on a harmonic of the square wave out of the PA.

So where do I go from here?

Bill






_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com