I have tried such isolator ,but for dc voltages . It works ,but high non - linearity , for audio it would mean distortions of probably percent range. Or special optocoupler needed ,which will have one led and few photodiodes or phototransistors , one of them for local feedback. Similar technology i saw once in old crt tv, all audio and video RCA jacks had such way of isolation . Because that tv had strangest technical solution used , it had no power supply, just buck converter. All other voltages were provided by hv transformer.
About usb - i think you should add ground at least , usb data lines as i know are not floating like LAN. Audio noise and clicks before and after playback is i think issue of power saving in tablet , turning off amplifier while playback is stopped.
About usb - i think you should add ground at least , usb data lines as i know are not floating like LAN. Audio noise and clicks before and after playback is i think issue of power saving in tablet , turning off amplifier while playback is stopped.
May be some minor "coloration" (distortion). Once made and tested you will need to decide by yourself if noise or distortion.
In any case it would be intersting to know if noise is modulated or added into audio path. It isn't the same: modulation, as in your small AM TX, implies nonlinearity and analog multiplication of both signals and is very difficult to get rid of them. If added, it is a linear process and is easy to fix.
Those double diode photocoplers were extensively used in analog passage of signals across a large voltage barrier in AC/AC motor control with more than 1KV and plenty of RF noises from commutation of large IGBT's across motor windings, and warranties that nonlinearity in photodiodes and led is negligible small although not null.
In any case it would be intersting to know if noise is modulated or added into audio path. It isn't the same: modulation, as in your small AM TX, implies nonlinearity and analog multiplication of both signals and is very difficult to get rid of them. If added, it is a linear process and is easy to fix.
Those double diode photocoplers were extensively used in analog passage of signals across a large voltage barrier in AC/AC motor control with more than 1KV and plenty of RF noises from commutation of large IGBT's across motor windings, and warranties that nonlinearity in photodiodes and led is negligible small although not null.