Inspired by the Hakuin SE Class A HPA Amp of xrk971 and the KISS HPA of AKSA and Nico Ras I built a kind of mix, BJT only:
Both Hakuin and KISS amps make use of a Sziklai CFP, a fiery combination of a NPN and PNP transistor with a chance of oscillating, especially so because this time I used a faster PNP BJT Q1 as an experiment. To prevent oscillation I applied three methods at the same time:
1. A base stopper resistor R10 of 47 Ohm;
2. Both transistors are placed as close together as humanly possible (for me that is);
3. A ferrite bead at the emitter of the NPN transistor Q2.
I think I succeeded in getting the circuit to behave at the first try. While testing it with open inputs, loose cables and such there was little noise to be heard. In fact it sounds very promising just using my phone as a test source.
R9 is a multiturn pot to adjust the characteristics of this amp:
32 Ohm headphones:
R9 4k THD 0.08%
R9 9.6k THD 0.04%
80 Ohm headphones:
R9 4k THD 0.06%
R9 9.6k THD 0.002%
150 Ohm headphones:
R9 4k THD 0.06%
R9 9.6k THD 0.005%
R9 = 4k, 80 Ohm headphones:
Only R4 and R5 become hot so use 5W ones. The heatsink I recycled from an old PC power supply, it gets slightly warm.
I suspect a circuit like this is vulnerable (soundwise) to noise from a switched mode power supply. Therefor I made a common mode choke from three smaller ferrite rings I glued together, about ten windings of twisted pair from an UTP cable and some other components.
Attached the LTSpice file.
Both Hakuin and KISS amps make use of a Sziklai CFP, a fiery combination of a NPN and PNP transistor with a chance of oscillating, especially so because this time I used a faster PNP BJT Q1 as an experiment. To prevent oscillation I applied three methods at the same time:
1. A base stopper resistor R10 of 47 Ohm;
2. Both transistors are placed as close together as humanly possible (for me that is);
3. A ferrite bead at the emitter of the NPN transistor Q2.
I think I succeeded in getting the circuit to behave at the first try. While testing it with open inputs, loose cables and such there was little noise to be heard. In fact it sounds very promising just using my phone as a test source.
R9 is a multiturn pot to adjust the characteristics of this amp:
32 Ohm headphones:
R9 4k THD 0.08%
R9 9.6k THD 0.04%
80 Ohm headphones:
R9 4k THD 0.06%
R9 9.6k THD 0.002%
150 Ohm headphones:
R9 4k THD 0.06%
R9 9.6k THD 0.005%
R9 = 4k, 80 Ohm headphones:
Only R4 and R5 become hot so use 5W ones. The heatsink I recycled from an old PC power supply, it gets slightly warm.
I suspect a circuit like this is vulnerable (soundwise) to noise from a switched mode power supply. Therefor I made a common mode choke from three smaller ferrite rings I glued together, about ten windings of twisted pair from an UTP cable and some other components.
Attached the LTSpice file.
