Oznog
Active Member
I have an LM386 circuit which will have headphones hooked up at the time it turns on.
The headphones are capacitively coupled which means there will be a pop as the LM386 output stage biases. Also the rest of the amp has biased circuits which could make a pop even if the output amp didn't.
It is difficult to get a low impedance series element to disconnect the headphones during this time. A dual relay is too large, pricey, and takes too much power. Since this is not a lot of current, I was wondering about just shorting the output during this period. The digital pots can disconnect the audio input during this time no problem so there's only pop current. At first I thought about shorting past the headphones with an N-channel FET to ground, but then I remembered about the intrinsic diode in MOSFETs. Putting an NMOS/PMOS back-to-back would be an undesirable level of complexity (board space, component count, cost). I thought about using a transmission gate like a 4016 either in series or to short the pop to ground but in either case the resistance is much too high.
Any other ideas? I could always do a bipolar transistor. It would mean the speakers would still see a 0.3v pulse due to the limited saturation voltage, and the other prob is the signal needs to be switched on by a microcontroller which means there's going to be a brief between power application and microcontroller execution. With the FET I could have left a 10k pullup resistor to Vdd on all the time and it wouldn't drain the battery when off, and the microcontroller only pulls the gate low after the turn-on delay expires.
Any ideas what I should do here?
The headphones are capacitively coupled which means there will be a pop as the LM386 output stage biases. Also the rest of the amp has biased circuits which could make a pop even if the output amp didn't.
It is difficult to get a low impedance series element to disconnect the headphones during this time. A dual relay is too large, pricey, and takes too much power. Since this is not a lot of current, I was wondering about just shorting the output during this period. The digital pots can disconnect the audio input during this time no problem so there's only pop current. At first I thought about shorting past the headphones with an N-channel FET to ground, but then I remembered about the intrinsic diode in MOSFETs. Putting an NMOS/PMOS back-to-back would be an undesirable level of complexity (board space, component count, cost). I thought about using a transmission gate like a 4016 either in series or to short the pop to ground but in either case the resistance is much too high.
Any other ideas? I could always do a bipolar transistor. It would mean the speakers would still see a 0.3v pulse due to the limited saturation voltage, and the other prob is the signal needs to be switched on by a microcontroller which means there's going to be a brief between power application and microcontroller execution. With the FET I could have left a 10k pullup resistor to Vdd on all the time and it wouldn't drain the battery when off, and the microcontroller only pulls the gate low after the turn-on delay expires.
Any ideas what I should do here?