> If the duplexers for each system are only 50 ohms at each 
> receiver and transmitter, then I should see 4 frequencies 
> where there is a 50 ohm load, and they are all different. Why 
> would I need more than some sort of phasing harness to 
> connect the two duplexers to the single transmission line, 
> assuming the BpBr duplexers have enough isolation to keep the 
> two repeaters from bothering each other?

Because your assumption that the duplexers will provide enough isolation
betwen Tx1 and Rx2 and Tx2 and Rx1 may be unrealistic.  On the receive side,
notches for a given duplexer are tuned to reject that duplexer's
transmitter, not the other transmitter.  Same for the transmit side - the
notches are tuned to notch that repeater's receive frequency, not the other
repeater's.  Am I saying it definately won't work?  No, it's far from being
a sure win.

Think about it this way.  Not only do you need adequate isolation between
your Tx and your Rx, but you also have to consider the isolation required
between your Tx and the other Rx, and between the other Tx and your Rx as
well.  Receiver-to-receiver isolation is likely not an issue unless you have
LO leakage issues.  Of course, do the math to make sure you don't have any
intermod hits either...

For the heck of it I just swept a 4-cavity duplexer (Decibel DB4076W) on the
VNA to see what the response looked like.  With the duplexer tuned for
444/449 MHz, I looked at the transmission response from the the transmitter
port to the antenna port.  The duplexer provides about 46 dB of attenuation
at the 468 MHz receive frequency, i.e. the isolation from your Tx to their
Rx would only be 46 dB + whatever insertion loss is in their duplexer
(probably about 1 dB).  And that appears to be the best case scenario.  If
you look at the other side of the notch (lower in frequency), the response
is much less, so the same duplexer used for the 463/468 box would likely
provide little (on the order of 20 dB) of attenuation of the other
repeater's transmitter noise at your receive frequency.  I can guarantee you
that having only 20 dB of noise supression will make you very sad.

The better the front end of the receiver, the less Tx carrier supression you
will need.  The cleaner the transmitter, the less noise supression you will
need.  Those rules hold true whether you're talking about one repeater, two
repeaters, or ten repeaters.  Without knowing exactly what you have, and
they have, as far as equipment, it's hard to guess how much isolation you're
really going to need from any port to any other port.

If you had two "true" bandpass duplexers at your disposal, each of which had
adequate isolation at the standard 5 MHz split to prevent desense when used
in a single-repeater application, you would likely make out OK by teeing
them together with appropriate-length cables since your Rx is close to 14
MHz away from their Tx, i.e. significantly further than the 5 MHz away your
own Tx is, and therefore, that much more attenuation could be expected due
to the bandpass characteristics of the filters.

A simple solution, if you don't want to replace duplexers, would be to use
two window filters, one tuned to pass the ham Tx and Rx, and the other tuned
to pass the commercial Tx and Rx, and tee them together with the
appropriate-length cables (which will vary based on the design of the
filters, and is something best left to be done on a network analyzer).
This, of course, assumes that the skirts on the window filters are steep
enough to make up the additional isolation required to avoid desense between
systems.

This could also be done with a 4-port combiner/multicoupler, but would
likely require 8 to 12 cavities total to accomplish (a combination of pass
and pass/reject most likely).  I've built transmit/receive combiners for a
number of ham projects, including one where I have a 442/447 ham repeater
combined with a 450/455 MHz repeater.  That one took 10 cavities and two
single-stage isolators to achieve the required isolation, as well as provide
adequate Rx filtering to avoid desense from other co-located stations.

Why not just put up a separate antenna?

                        --- Jeff WN3A

Reply via email to