I’m not questioning you, please don’t take it as a personal attack. We see too much of that.
Here’s why I’m questioning what really happened. Because we use FM the frequency of the RF signal varies with the frequency of the audio which modulates it. But the government (and courtesy) stipulate how much bandwidth we’re allowed. For FM stereo broadcast transmitters are allowed a wide bandwidth so they can broadcast a wider audio spectrum. But two way radios is only expected to reproduce speech. In fact I would expect our radios to have a filter between the microphone and the modulator to avoid creating too wide of a bandwidth. I would also expect that neither the speakers nor microphones on the inexpensive radios to be able to reproduce ultrasound.
I thought I read in their PDFs, that Ribbit attempts to send digital data by converting data into audio tones that are in the center of the spoken voice audio spectrum, between 500 Hz and 2500 Hz. Their centerpoint is 1500 Hz and they go 1000 Hz either side. 20,100 Hz is way above that, but 2100 Hz is right in there. If you were able to transmit at 20100 Hz, I would be interested in what an RF spectrum analyzer would measure.
But, I always say you can’t argue with empirical evidence, so I’m trying to understand what really happened.