Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!uxc.cso.uiuc.edu!tank!shamash!nic.MR.NET!srcsip!gorby!ferguson
From: ferguson@gorby.SRC.Honeywell.COM (Dennis Ferguson)
Newsgroups: sci.electronics
Subject: Re: How do I make a notch filter with cable?
Summary: distributed element filters
Keywords: filter, notch, cable, short, simple, hi-Q
Message-ID: <26080@srcsip.UUCP>
Date: 21 Jul 89 19:05:49 GMT
References: <4436@merlin.usc.edu>
Sender: news@src.honeywell.COM
Reply-To: ferguson@gorby.UUCP (Dennis Ferguson)
Distribution: na
Organization: Honeywell Systems & Research Center, Camden, MN
Lines: 49

In article <4436@merlin.usc.edu> cyamamot@nunki.usc.edu (Cliff Yamamoto) writes:
>Greetings,
>
>	I am trying to make a sharp notch filter to cut out an overbearing
>paging transmitter that swamps my PRO-2004 scanner on the VHF bands.  I
>*think* I've read somewhere before about how one can cut a piece of cable
>to a certain length and short it at the end to kill the frequency of said
>wavelength.  Something like this...
>
>         |<------------feed line cable length important??--------->|
>
>         -----------------------------------------------------------
>Antenna  ------------------------------+---------------------------- Receiver
>         ----------------------------+ | +--------------------------
>                              ^      | | | \
>                              |      | | |  \ 
>         The length of this   |      | | |   \___location of BNC "T" connector
>         "stub" is cut to the |      | | |       important??
>         wavelength of the    |      | | |
>         offending frequency  |      | | |
>                              v      |_|_| <---Short center conductor to shield
>
>The pager frequency is 152.4 Mhz.  So using the ol' ARRL handbook I get a
>wavelength of 1.96 meters (length of the "stub").
>

The effective wavelength of your stub is the free space wavelength
divided by the square root of the dielectric constant of the coax.
RG58 or RG59 usually has a polyethelene dielectric with Er=2.2-2.6.
Therefore the real wavelength is approximately 1.25 meters.  A full
wavelength is not necessary, for the shorted stub one half wavelength
will do and for an open stub, one quarter wavelength will work.  The
placement of the stub with respect to the antenna and receiver depends
upon if the antenna and receiver impedances look like the standard RF
transmission line impedance of 50 ohms.  Probably the receiver does but the
antenna does not.

A different notch filter that works just as well does not require placement
at a specific distance from the antenna is a transversal filter.  Take a piece
of coax one half wavelength of the frequency you wish to notch (152MHz) and
provide two paths for the signal, one normal path and one which is parallel
but 1/2 wavelength longer.  Since the signals sum 180 degrees out of phase,
the unwanted signal is cancelled out (as well as all harmonics of the signal).

Probably the best approach is to build a twin tee notch filter.

Opps.... my notes aren't laying around... I'll email you the details.

Dennis