At 2/11/2007 08:27 PM, you wrote: > >Ponder this Bob (and anyone else). Assume the preamp input is connected > >directly to the output port of the isolator. For the sake of argument, say > >that the preamp has a NF of 0.5 dB, and 6 dB input return loss. The power > >transfer from isolator output to preamp input would be 75% for 6 dB RL; the > >other 25% of reflected power was lost/dissipated in the reject load. > >Wouldn't it be mathematically correct to say that there was effectively 1.25 > >dB of additional loss ahead of the gain stage, and therefore the noise > >figure would likewise then be 1.25 dB worse (i.e. 1.75 dB)? Yes? No? > >Maybe? > >No. The reason is that the noise figure is specified for a 50 ohm input >feed to the preamp, IOW the noise source used to test the preamp is 50 >ohms, so any power lost due to mismatch at the preamp input is already >taken into account in the measurement. > >Bob NO6B > >True that the noise figure is specified for a 50 ohm input but that is for >a non reactive 50 ohm input. Once you hook it to a cavity or antenna etc. >you no longer have a pure 50 ohm input so the noise figure may not be the same. > >73 > >Gary K4FMX
True, but the original question was if the noise figure of a preamp was degraded by the poor input match looking into the preamp itself, & the answer is no. Feeding the preamp with an impedance other than 50 ohms yields unpredictable results, & if the preamp is not unconditionally stable, it can oscillate. Chip Angle guarantees his PHEMT GaAsFET preamps to be unconditionally stable across all input impedances, IOW the entire Smith chart (see http://anglelinear.com/gaasfet/gaasfet.html). I don't know if ARR makes the same claim. Bob NO6B
