This duplexfilter is devellopped after study and some experiments with construtions described in ARRL-publications.
When, at our request, the filter was measured by the RDR (governement), they were surprised that such good results were possible with so "simple" and cheap construction, and without the use of elaborated equipment. Notches were over 100 dB deep. The filter did not suffer from transporting to the measuring-place.
The filter is now in service for abt. two years. It gave no problems, and needed no re-tuning when it was checked after abt. a year of service.
The filter comprises of 2x3 notches + 2x1 bandpass. The bandpasscavities can be left out. The filter will then have a little less passband-damping, but less selectivity. This is not recommended when strong in- and outband signals are present (crossmodulation / intermodulation / mirror-reception), and when it is not sure if the repeatertransmitter has very clean sidebands (sidebandnoise at 600 kHz distance).
Our STORNO CQF612 in the 2mtrs repeater PI3RTD can handle up to 5mV transmitter signal before saturation (at 600 kHz shift). With a transmitter output of abt. 20V we need more then 80 dB suppression of the transmitter carrier. The with this filter obtained 90 dB or more gives some spareroom, and less chance to have intermod products from very strong signals in and near the amateurband.
Sidebandnoise from the CQF612 at +/- 600 kHz is estimated to be abt 95 dB below carrier (BW 12 kHz). Receiver noisefloor is estimated abt. 165 dB below transmittercarrier. The needed suppression to avoid desensibilisation of the Rx due to transmitternoise: 80 dB or more.
In The Netherlands a known problem in towns is: intermod in receiver-input-stages coused by very strong signals from paging-transmitters on abt. 155 / 160 / 163.35 Mhz, due to overloading of the receivers. The best remedy is: add selectivity. Some cavity filters act only as a "Notch filter", adding little or no further extra selectivity. The filter described here is designed for repeater-use in urban areas with lots of HF, adding extra selectivity to the repeater-receiver and transmitter, becouse of an extra bandpass-cavity in each filter-half.
How the filter behaves in harsch environments can not be said. Our filter is installed in a air-conditioned room. But installing it in a closed and temperature-controlled box should also be a good solution.
Single cavities could be used as bandpassfilter or notchfilter to suppress unwanted very strong signals in a stationreceiver. For a notchfilter some own devellopment then must be done with respect to Cn or Ln, to get the notch at the desired distance of the pass-frequency. A frequency-distance (shift) of less then 600 kHz is not recommended, as the filter becomes difficult to adjust (pulling of the pass-frequency when adjusting the notch), and the deepness of the notch will be less.
The use of a single cavity as a bandpassfilter is simple. No notch is then needed, and even in-band signals can noticeble be suppressed. But the usable bandwidth is very narrow, especially with respect to VSWR (transmitter). When you should want to use it on frequently changing frequencies, you have to change the construction for the movement of the plunger. Warning: if a tuning-rod sticks out of the toplid, it could pick up signal, thereby deteriorating the selectivity of the filter!
Before building the whole filter, i suggest that you should make first one separate test-cavity. Test it as a bandpassfilter on the desired freqency, and then with Ln and Cn, as a notchfilter with a notch at exactly +/- 600 kHz distance. You will then become familiar with the production, adjustment, measurement and properties of a cavity bofore building the whole filter.