For a long long long while now, I have had varying reports of some sort of problem where the rxsquelch setting for chan_usbradio changes depending on whether you just did a 'radio tune rxnoise' and then saved it, or you just brouight up chan_usbradio from its saved configuration. In addition, I myself, have seen cases where the RSSI and squelch values seemed just really silly high for some reason. I believe I have found the culprit and have fixed it. It turns out that the hardware gain of the USB sound chip was different depending on whether it just had a 'tune rxnoise' or if it was set up from the config file. It turns out that because of a rounding error, it would often set its value 1 too low (compared with what the 'tune rxnoise' calculated it should have been). Believe me, when you only get a range of 1 to 16, 1 too low makes a big difference (about 1.5 db)!! This has now been fixed in the new version of chan_usbradio that will appear on the SVN this afternoon after about 3:30 PM PDT. Also, because of changes made in the squelch processing a few months ago (hysteresis was added), the squelch turn-off point is at least 60 'RSSI units' (whatever they are) higher then the signal level. Its probably a good idea to set the rxsquelch level to 80-100 higher then the level with no signal. This should both fix and explain a lot of strange things going on. I dont think this deals with some of the weirdness with 12.5 kHz receivers, but it might at least help in this area. JIM WB6NIL