I recently built a prototype power amplifier, primarily to learn about designing amplifiers. (So it's not really meant to be a "good" amp, but it IS supposed to give me practice, and to be something realistic to measure and tweak, and from which to extrapolate to build a better amp next time.)
Another goal was to minimise cost, primarily by using up components I already had. (For instance, the LSK489 JFETs in the input stages were left over from an active oscilloscope project https://codeberg.org/gtw/active-probe and the IRF510 output MOSFETs were from an RF power amplifier https://codeberg.org/gtw/hfv-1 and I had a 24 VCT transformer from an unused power supply and I already have 2N3904/6s bought by 100s. So I didn't buy much except the IRF9510s.) The other main requirement was to refrain from more power than I need: I have sensitive speakers in a small room, so 5 W per channel gets me over 100 dB SPL peak. Let's call that about 80 dB average for classical music and 90 dB average for rock. That's already above the threshold where my signal-to-noise ratio plummets because my wife screams complaints about my music from the next room: more power doesn't help.
So... in the end I suspect the result is rather idiosyncratic, but it does work, and has certainly been instructive. Now to decide what to do next time...
The schematic:
SPICE analysis before building it:
Evidence seized from the crime scene:
Evaluating the damage:
Another goal was to minimise cost, primarily by using up components I already had. (For instance, the LSK489 JFETs in the input stages were left over from an active oscilloscope project https://codeberg.org/gtw/active-probe and the IRF510 output MOSFETs were from an RF power amplifier https://codeberg.org/gtw/hfv-1 and I had a 24 VCT transformer from an unused power supply and I already have 2N3904/6s bought by 100s. So I didn't buy much except the IRF9510s.) The other main requirement was to refrain from more power than I need: I have sensitive speakers in a small room, so 5 W per channel gets me over 100 dB SPL peak. Let's call that about 80 dB average for classical music and 90 dB average for rock. That's already above the threshold where my signal-to-noise ratio plummets because my wife screams complaints about my music from the next room: more power doesn't help.
So... in the end I suspect the result is rather idiosyncratic, but it does work, and has certainly been instructive. Now to decide what to do next time...
The schematic:
SPICE analysis before building it:
Evidence seized from the crime scene:
Evaluating the damage:
