Tim, It appears to me that your measurement procedure is correct - and that the results you have gotten would normally be sufficient isolation to allow desense-free duplex operation.
If you have some attenuators available - or better yet - a switchable [in 1 dB increments] attenuator - there is a test you could try. Assuming you are running your desense test with the repeater terminated in a quality 50 ohm resistive power load and feeding your generator in through an iso-T, you establish a sensitivity reference [e.g. 12 dBS] then key the repeater transmitter and readjust the generator for the same reference. The difference in readings is the amount of desense. Insert an attenuator in the receiver line between the duplexer's receive port and the receiver's antenna port. Start with about 10 dB. You should then require 10 dB more signal from your generator to achieve your initial quieting reference. Now when you key the transmitter again measure the desense. The attenuator gives you the equivalent of that much additional isolation. Your desense should be correspondingly less. With a switchable attenuator you can determine just how much more isolation you need. The results you get - whether they 'track' or aren't linear may give some clue as to the nature of the issue. WB0EMU -------------------------------------------------------------------- --- In Repeater-Builder@yahoogroups.com, Tim <tahr...@...> wrote: "...I measured each way to the common point... RX to antenna & Tx to antenna, and each one had a notch of about 102dB at the opposite frequency. With the 50 watts at the antenna port is where I see the -55dBm on the receiver port. (into the spectrum analyzer)...."