Hi Ranko, Thank you for your quick response! It might be worth considering noting this in some form so that others can benefit from your suggestion. Obviously the high SWR protection circuitry works to some degree, do you feel that the icing presents some sort of "outside the design envelope" situation? Given that the time between the onset of the icing event and the "danger time" is fairly rapid (maybe <1 minute), one possible solution is to have the amplifier (in conjunction with flex xcvr) map and store the good swr curve of attached antennas and shutdown the amp when a significant deviation is detected..that way it may be possible to be proactive without having to actually look out the window and react manually....sometimes such icing events are not always predicted. Just trying to come up with ideas that could save final transistors....73s es Thanks again de Tim (N0TB)