I had a chance to try some testing with sideband inversion.
BFO = 11.9986 MHZ and VFO = 4.8486 MHZ this gives me a starting frequency of 7.15 MHZ(RF = BFO - VFO). Set the encoder to increase frequency in CW direction. But this to increase VFO or decreasing RF. So I switched it so RF goes up in CW direction.
I usually listen in on a net at 7.163 MHZ. To tune in the voice I had to increase RF. So I find that LSB is on the wrong side of the (non-existent) carrier.
Think of it this way: VFO - IF = RF. Does this result in sideband inversion? (hint: -IF + VFO. It it positive or negative?) Why don't you try it and see what happens. If you get nothing but garble, then the answer is "no...it does not.
If the results are clear signals, then you must decide if the LSB is being inverted, or is it truly USB. Remember that the ordinary condition by convention is that the LSB on 20m is garble or non-existent. Therefore you must hear USB inverted to LSB or true USB for intelligibility. Be aware of AM; test your results through several different QSOs.
Actually, it doesn't matter which it is -- as long as communication takes place.
