The transistors in the current mirror, and differential amplifier will benefit from having thermal contact, to keep the temperature difference as small as possible, thereby minimising the DC output offset. I have taken the best picture I could, of this coupling for one channel, in my amplifier.
The beast has been tamed. It sounds good too, although my headphones are cheap Sennheisers. The headamp is enclosed in a Hammond enclosure, using the aluminium casing as heat sink, it gets hot. It makes for a nice way to keep the coffee warm by placing the cup on top of the amplifier.
Q6, Q8, Q15, and Q17 are the output transistors, and needs heat sinks, I have not calculated the exact size needed, but the bigger the better. I use the aluminium chassis of a Hammond 1455N2202 with the transistors mounted on the bottom, and it gets warmer than I like. Never test the amp without proper heat sinking in place, as the output section will most certainly release the precious blue smoke.
Power supplyI have used a 2X12V toroid, a rectifier bridge, and 4700uf on each supply rail. Each channel draws about 200mA.
GroundingSome attention is advised when wiring the ground, my first ground layout resulted in a nasty saw-ish 50Hz hum from the amplifier. I now connect the separate grounds from the input RCA connectors to their respective ground pads, next to the input pads. The power ground are connected to a star ground between the 4700uf power supply capacitors, as well as the ground wire from the headphone jack.
Build the last simulated version of the SE class a headamp today, on breadboard. It works as an amplifier, it is class A, it is hot, and it is oscillating. I have to read a little more about taming the beast, did try some tricks, and it helped. Also I have to try some other transistors in the output stage, I have two different brands of BD139. I was unable to measure the hFE of the ones of which I have the most, even though the other brand measures fine. Since this was a prototype, I used the strange ones, maybe they are bad to the bone.
NOTE: I found this thread at diyaudio.com read through, it's really nice stuff: Pure Class A Single End Amplifier Idea!
John Linsley Hood Class A amplifier. Distortion figures and PSRR look quite nice, but the circuit has only been simulated, I have a TDA1545 DAC to finish If I could only keep my mind on a single project. If my calculations are correct, the amp will deliver 250mW into a 32Ohms load in class A.
PS: Sorry for the extra ground symbol.
Generated on 2018-05-03 01:14:21.910717