It sounds like the shunt cap on the filter end may be used for tuning out
any inductive reactance in the filter itself. The rest of the circuit, with
the series L, shunt C might be used as an impedance matching network that
would step the impedance of the filter up to a new value.
This type of arrangement, without the shunt C at the input end is a classic
way of matching a 50 Ohm doubly-balanced mixer output to a high impedance
crystal filter. Good design practice would usually have the coupling caps to
have an impedance of between 1 and 2 Ohms.
Now, if you were feeding a lower impedance load, you could look at this
network from the load end as a series L, shunt C matching network that would
match the impedance of the load to the higher impedance of the crystal
filter. In this case, the shunt C would be to tune out the imaginary
inductive term at the input of the load itself (this would be called
conjugate matching).
If you were to connect this circuit to an RF network analyzer, you would be
able to see the magnitude and the angle of these terms, but a good
simulation program such as Ansoft Designer (free) would show you what is
going on. You would need to have models of the devices in the load itself
for maximum accuracy.
That's the fun part of RF design.............the experimentation and the
learning!

Pete

"AndyS" wrote in message
oups.com...

john wilkinson wrote:
Hi,
When I was looking at an IF amp in the "Experimental mathods in RF", I
noticed a filter I have not seen before.
I will try to describe:

200 Ohm i/p impedance feeding 1 cap. The othe side of the cap feeds
another cap down to ground and an inductor. The other side of the
inductor
has a cap down to ground. The junction of the ind and cap, has another
cap
feeding a load of 100 Ohm.

When I simulate this it is a bandpass filter.

What do you call this filter, and where can I find some info about it?

Thanks,
John

Andy writes:

This is a 3 pole low pass filter. The type ( butterworth,
chebychev, bessel, etc)
is determined by the component values, as are the frequency , skirt
selectivity and impedance.
Look in the ARRL handbook for typical configurations and
values....This
filter can also be used to transform impedances, although when that is
done
you leave the "type" classification and become more just a low pass
impedance transformer.....

It ain't rocket surgery, but if you aren't into LaPlace transforms,
you need
to stick with the tables.... or the simulator.....

The cap connected to the 200 ohm output is just for coupling and
is
usually really big, but, in some cases , may be made small enough to
provide a small breakpoint for high pass but at a frequency
considerably
lower than the low pass passband.....

If you use PSPICE , or equiv, to simulate it, you can learn a great
deal...

Andy W4OAH