Glad that the "bug" is a user bug!
Perhaps I should add some routines to check what the users input, but that would cost me valueable computing time and make the GUI a little awkward. Right now, the input is being processed as one types in the values. If I would check the values for correctness, I could only apply them once they are fully input. Not going that route... But I agree, some error checking should be done, at least to prevent the program from crashing.
Anyway: I just uploaded a new version.
It features a corrected and improved channel power meter.
This is an advanced measurement, which determines the integrated power over a selected bandwidth (normally the channel bandwidth).
I keep this on a separate tab and the spectrum must be stopped to run this measurement.
My initial tests indicate that the precision is pretty accurate (see blog).
Regards,
Vitor