Great project with excellent docs and build info as always from @PMA.
I'd like to add one simple modification which allows mounting the power transistor directly on a grounded metal chassis without need for insulation. The mounting tab of power transistors connects to the middle pin and that is usually the Collector terminal. The ground reference point simply moves from one end of the supply to the other, with basically no change in circuit behavior.
(Transistor model and part values are only for illustration, please stick with Pavel's values for a working circuit).
This circuit is also ideal when the load resistor R2 is replaced by a current source which then is the only factor that determines power supply ripple rejection.
------:-----
And here is another little variation that directly improves on power supply ripply rejection without the need of a capacitance multiplier, both with open input and terminated input (rejection largely depends on source impedance because this impedance basically is a parasitic like the ESR of the cap to which it adds).
Further, the split output capacitor (which doubles up as supply filter capacitor) is moved outside the loop and therefore frequency response is only limited by the input capacitor. Output impedance stays constant at low frequencies. That may or may not be of relevance for the application.
What's relevant is handling of power on/off transients which are quite severe in all of these circuits. Probably the simplest solution is a DPDT mute switch, with one contact shorting the input and the other one shorting the output. That way, both capacitors charge quickly.
Also note that the second circuit absolutely requires a fully floating supply (2-prong wall-wart etc) otherwise it will send DC through your speaker when the input also gets earth-grounded :-(
I'd like to add one simple modification which allows mounting the power transistor directly on a grounded metal chassis without need for insulation. The mounting tab of power transistors connects to the middle pin and that is usually the Collector terminal. The ground reference point simply moves from one end of the supply to the other, with basically no change in circuit behavior.
(Transistor model and part values are only for illustration, please stick with Pavel's values for a working circuit).
This circuit is also ideal when the load resistor R2 is replaced by a current source which then is the only factor that determines power supply ripple rejection.
------:-----
And here is another little variation that directly improves on power supply ripply rejection without the need of a capacitance multiplier, both with open input and terminated input (rejection largely depends on source impedance because this impedance basically is a parasitic like the ESR of the cap to which it adds).
Further, the split output capacitor (which doubles up as supply filter capacitor) is moved outside the loop and therefore frequency response is only limited by the input capacitor. Output impedance stays constant at low frequencies. That may or may not be of relevance for the application.
What's relevant is handling of power on/off transients which are quite severe in all of these circuits. Probably the simplest solution is a DPDT mute switch, with one contact shorting the input and the other one shorting the output. That way, both capacitors charge quickly.
Also note that the second circuit absolutely requires a fully floating supply (2-prong wall-wart etc) otherwise it will send DC through your speaker when the input also gets earth-grounded :-(
Last edited:
Regarding FETs, I published the mosfet follower 22 years ago: https://sound-au.com/project83.htm
The goal here in this thread was to make it the simplest possible way, just for fun, but with acceptable distortion and low output impedance. Already said in post #1.
You are mentioning NP - my view is different from his. My necessary condition is low output impedance, which cannot be achieved with common source or common emitter 1 transistor circuit without GNFB. Zout >= 1ohm is unacceptable to me. DF must be 10x at least. The trick is that the follower has very strong feedback, though it is not understood and accepted by many audiophiles. What you hear with these high Zout circuits is not a harmonic distortion (as some try to suggest over and over), but frequency response modulation created by divider consisting of high Zout / complex impedance of the speaker load. That is all the "miracle" of SE tube circuits or common source amplifiers (but yes, if SE tube has H2 >10%, then it might be audible). Air compression stories and "improvement" of sound by H2 are fairy tales.