Good thoughts Milt, and I'll add a few
> While not an easy thing to find I would suggest that you most likely
> need some sort of a bandpass cavity on the receiver to protect from
> the noise that gets past the heliax notches.
>
> Remember that a notch duplexer only removes the notched
> portion of the
> TX signal on the RX side and the RX signal on the TX side, all other
> noise is passed directly to the load. Thus you only have two small
> notches, one at the RX frequency and one at the TX frequency.
> Everything else is passed.
A duplexer specification that often goes overlooked is "mid-band isolation";
that is, how much isolation there is between Tx and Rx ports mid-way between
the Tx and Rx frequencies. For notch-only duplexers, this value is often
very low, often less than 10 dB. The effect of low mid-band isolation is
that wideband noise or spurs from the transmitter can result in receiver
desense, even if there is enough isolation at the operating frequencies. In
other words, the wideband noise passes right across the duplexer at
frequencies far enough removed from the notches to cause problems.
For pass/reject or bandpass duplexers, the mid-band isolation will be
substantially higher, may be somewhere in the range of 30 to 60 dB depending
on band, offset, number of cavities, etc.
Mid-band isolation is often quoted in manufacturer's specs as a simple
scalar value, if it's given at all. Quite often they just give you
"isolation", and that's just at the Tx and Rx frequencies proper; it doesn't
tell you anything about what's happening at other frequencies. A swept
transmission response across a broad range from Tx to Rx port with the
antenna port terminated will show the true isolation you're getting.
As far as adding a pass cavity to attenuate desense caused by noise or spurs
coming from the transmitter, it would most likely be more effective if you
put it on the transmitter leg of the duplexer rather than the receiver leg.
> You probably should also look at the TX signal to check for spurs.
Micors are generally pretty clean machines, but keep in mind that lowband
repeaters were fairly rare back in the day; I don't know if duplex isolation
curves were ever published for lowband Micors (ZZU, you QRV?). For the
Mastr II you only needed about 50 dB of carrier supression and a little over
60 dB of noise supression for 100 watts at 1 MHz split.
> I also have had duplexers that look good with a tracking
> generator but
> fail under TX power.
And we've all had antenna systems that did the same. And I've had dummy
loads that did the same as well; point being, don't rule out a problem in
your test equipment...
--- Jeff WN3A
