[Collins] 75A4 filter bridging

[Dr. Gerald N. Johnson, electrical engineer]

In the latest CRA Journal there's a brief article about another article
about bridging with mechanical filter with a small capacitor (wire
spaced a half inch from the other terminal).

This is a known technique that can enhance rejection, if the relative
phase is correct. It might be viewed as neutralization except that the
skirts of the typical Collins mechanical filter of that era are not from
leakage, they are from the Tchebyshev response curve. The Tchebyshev
response curve trades pass band ripple and stop band loss for a steep
transition slope. The more the bass band ripple accepted and the less
the stop band loss accepted the steeper the transition from pass band to
stop band. There's a notch just past that transition, then the response
THROUGH the filter rises and slowly tapers off. The bridging capacitor
cancels that response area. Probably at the cost of raising the signal
leak through in that close in notch and at greater frequency spacings.
Fortunately the IF transformers will help at greater frequency spacings.

Since the loss at the close in side responses was noted as 80 dB down,
it doesn't take much stray coupling to enhance it when the phase angle
is right, but that small C is hard to control and will be changed when
the house hold spider builds a web in that area, or just lives there. Or
the shimmer of a distant earthquake moves the wire a hair. The
adjustment likely won't survive shipping and when the added stray
coupling is stronger than the filter stop band loss, the receiver stop
band loss will be lessened, reducing performance.

The bridged filter stop band loss depends on matching phase and
amplitude on parallel paths to make cancellation, and that's always a
delicate balance. The phase has to be consistent at 180 degrees and the
amplitude has to match. It works but it is often very touchy, and is
extremely sensitive to very small variations in either element. A 1 dB
change in either path can lead to a 10 or 20 dB change in stop band
loss. It would not be a good production technique unless done in welded
wire and hermetically sealed inside the filter package.

73, Jerry, K0CQ, Technical Advisor to the CRA
-- 
Entire content copyright Dr. Gerald N. Johnson, electrical engineer.
Reproduction by permission only.