Not sure that it actually needs the servo, but I was not happy with 18mV worst case DC offset at the output.
Seems to simulate pretty well.
Seems to simulate pretty well.
Is Servo really necessary?
Does not OPA164x hold the output near zero?
But It is true that the 18mV is on the high side.
It should work very well. What is the bias in output stage?
Does not OPA164x hold the output near zero?
But It is true that the 18mV is on the high side.
It should work very well. What is the bias in output stage?
OPA164x has worst case input offset of 3.5mV according to the datasheet, at 5x gain it would result in ~18mV DC offset at the output.
The standing current in the output stage is ~100mA. Large enough to drive pretty much any reasonable headphones in class A.
The standing current in the output stage is ~100mA. Large enough to drive pretty much any reasonable headphones in class A.
If you decrease the value of R2 the output offset will also decrease. Try 100nF 270k for hp filter. If the filtering is set too low, it will cause slow transients, excessive bass, and the overall sound will lack tightness and feel sluggish. The current generator can be replaced with 100R 5W resistor.
Thank you, that would be the case with a bipolar opamp, but with FET opamps the gate current of the input stage is so small (in the femto-ampere range) that the value of this resistor can be arbitrarily large without really affecting the offset.
Unfortunately, that would reduce the output transistor's ability to sink current.
I understood. But however, because of the high time constant of the RC filter (220nF and 1M in this case -> 0,7Hz 0,22sec), resulting in a very low cutoff frequency, the bass will dominate, the sound will lose detail, and the highs will be attenuated by the excessive low-end emphasis. I would suggest changing it to at least 220nF and ~150kΩ ->4.8Hz 33ms this is more than enough for headphones, but everyone has their own taste.😛