1970s Design Indulgence

1. The output stage bias is now fixed (no longer adjustable) and is determined purely by the thermal "feedback" diodes. Two are on each output transistor (one on each transistor) which, by their own temperature coefficient, reduce the bias voltage very slightly as the output transistors get hot, which prevents them going into thermal runaway. The two additional diodes which make up the base spreader voltage are on the circuit board.

All told there are 4 diode voltage drops between the driver transistor bases, which turn on the output transistors just enough to conduct some standing current, to prevent "crossover" distortion around the zero crossing of the loudspeaker signal.

The 4 MUR120 diodes gave an insufficient base spreading voltage, and so one of the board mounted ones was changed to a 1N4148 signal diode, which has a slightly greater diode drop.

The meters, which now only serve to indicate rather than assist adjustment, will show approx. 12.

With music playing, the instantaneous bias voltage will be greater than 12, but not so great as to cause a thermal runaway increase. This means the output stage is operating just on the south side of class AB.

2. The front panel adjustment trimmer is now out of circuit and no amount of adjustment will make any difference.

3. The input stage 0V returns now go by a single route, via the chassis, to power supply 0V. Before, each took its own route back to power supply 0V, and then each went to the input chassis connection.

What was formed was a diamond shape path where current could flow to 0V by the other input stage's 0V, which is the unfortunate convention still used by some designers. In this a.c. coupled design, the asymmetrical current pulses worsened the cross channel 0V effect. The effect results in audible distortion (brightness mainly), but isn't as noticeable with a DC coupled amplifier.

The 0V path is now determined so cross channel 0V connection cannot exist (a T shape connection).