Hi Paul,
I meant that it was not complicated to figure out. For example if the receiver noise floor (sensitivity) is -120 dbm then the transmitter noise power must be known at the receive frequency (as supplied from the manufacturer or measured) and an amount of attenuation added at the transmitter via cavity filters to bring that noise power down below the -120 dbm receiver noise floor. If the transmitter generated -40 dbm of noise power on the receive frequency then you would need 80 db of attenuation in the transmitter line to limit the noise seen at the receiver to -120 dbm. That's all there is to it on that side. On the receiver side you must know at what level the receiver begins to go into compression and cause desense at the transmit frequency. You get that from the receiver manufacturer or you measure it. You then add attenuation in the receive line to attenuate the power of the transmitter carrier to a level below that point. For example, 100 watts of transmitter power is +50 dbm. If the receiver can handle only -30 dbm at the transmit frequency before desense starts then you need 80 db of filtering of the carrier in the receive line to reduce the carrier to -30 dbm at the receiver front end. Not complicated, you just need to know what the transmitter and receiver capabilities are. As a note, if you are measuring transmitter noise power with a spectrum analyzer you will probably need to place a notch filter at the input of the spectrum analyzer to reduce the transmitter carrier or the spectrum analyzer will overload and give a false reading. A 100 watt carrier is +50 dbm. If you are trying to look at noise at the -40 dbm level that is a 90 db range. Many analyzers do not have that kind of dynamic range and will overload from the carrier level. A notch filter that knocks the carrier down say 30 db then allows the analyzer to operate in a 60 db range. Some analyzers may not even be happy with that level especially those in service monitors so you will need more than 30 db of notch filtering for the carrier. Be careful with close spacing like at vhf (600 KHz) so that you don't also attenuate the noise power at the frequency you are trying to measure. 73 Gary K4FMX _____ From: Repeater-Builder@yahoogroups.com [mailto:[EMAIL PROTECTED] On Behalf Of Paul Plack Sent: Thursday, November 01, 2007 1:31 PM To: Repeater-Builder@yahoogroups.com Subject: [Repeater-Builder] Re: Calculating required T/R isolation Gary, I appreciate the comments, and agree with all but one. If transmitter sideband noise, filter skirts and dynamic range can all be represented by numbers, then they should be able to be distilled down to a formula. A complicated one, to be sure, but nothing "magic." Those isolation-vs-separation charts date back almost a half-century and were probably created with a combination of direct measurement and sliderules. Today we have Excel and NEC. With an accurate enough model, even stray coupling can be approximated. Or, at least I'm gonna try! Most repeater antennas are chosen for gain and pattern and off-the-shelf availability, with an eye toward maxmizing gain at the horizon within budgets of money and tower space. All the "rules of thumb" are based in those priorities. My thought is that on low band, gain is limited by available tower space, so I'll shift my priority into reducing tower coupling first, optimizing gain and pattern second. If I can get 110 dB of isolation, 2 dBd gain at the horizon, a reasonable radiation pattern that's well-matched between transmit and receive, and a design that's both reproduceable and mechanically practical, it's a step forward. I'm quickly concluding that I'll need to model the antennas first, test to see if the model is accurate, then look for radios to fit. I'll let you know how I make out! 73, Paul AE4KR ----- Original Message ----- There is no magic formula. The curves given by the duplexer manufacturers are "typical" for radios like Motorola and GE... ...Receiver desense is a function of how much overload the first mixer can handle. In off frequency operation as in a duplex situation it is affected by the mixer capability itself, front end filter rejection capability and the preamp if one is used... ...Broadband noise is a function of how clean the transmitter is... ...Both of these specs are provided by the radio manufacturers of better radios. They will give that info at certain frequency separations. The charts that show antenna isolation with vertical separation are for antennas exactly in line with each other. Substituting a gain antenna for a dipole usually won't degrade isolation. A gain antenna will have a deeper null in the vertical direction in most cases. Isolation figures can vary with types of installation from stray coupling. Low band is a problem as it takes more separation to get the same isolation as vhf gives you. 73 Gary K4FMX
