Hi James,
Chuck Henderson, WB9UUS, made a couple modifications to the code... one I believe actually fixes a bug in how the RSSI is calculated (which is probably your issue), and the other changes squelch action.
A previous thread here:
If you look in the source code, you can find the specific changes:
The predominant issue here was even if the received signal was full-quieting, if it was a short transmission, it could cause a long squelch tail.
As for how the RSSI is calculated... if you look at the RSSI in Allmon, without this patch, you'll see the RSSI level jumping all over the place, especially on noisier signals. With the patch, things are much more stable.
Firmware images with and without the patch enabled are available on the repository for you to try out. They should be there for all hardware flavors:
If you are using a MicroNode RTCM, you will likely want the "SMT V1.5 Chuck" file.
Lee