[Argus25]

I have attached a copy of Mullard's original document supporting the OC16. Not much of a power transistor by today's standards, but then it was revolutionary and I still like it.

This is the circuit I used in the radio I posted a link to in post 7.

There is a lot than can be learnt from this circuit that can be generalized to higher power versions.

In this case the quiescent emitter current is 400mA. This results in a 0.88V drop at the emitter, which is only about 6% of the supply rail...so its not really needed to have an elaborate bias arrangement to eliminate this. The AC gain is reduced a little by the NFB resistor Rf, and the stage produces a very faithful sine wave <5% distortion, which I cannot hear. You can tame the higher range frequency response with a capacitor in parallel with Rf, or make a tone control there.

As stated it requires a heatsink of 5.5 degC/W or less. ( in my radio I increased Re to 3.3R increasing the emitter drop to 1.38V, used an 820R for Rf and R2 turned out to be 68R for 0.42A emitter current). The RMS collector voltage still reaches 7V before clipping.

There are two key points though,

The Thevenin resistance of the bias circuit is less than 10R and is about 8.7R. This is essential because of the germanium transistor and its leakage properties.

Secondly, with the driver transformer and an appropriated driver device you can easily generate the appropriate drive power which is not insignificant, as you can see from the drive requirement specs.

I used a 22:8 R impedance ratio choke as the output transformer.

When the EF98 is used as the driver (handy as it runs off 12V) the required transformer has a turn's ratio of 23:1, the DCR of the secondary should be less than 1R and the primary less than 200R.

It is quite surprising how pleasant these class A stages sound, a lot like a valve radio...but devoid of hum, especially running off a 12V auto battery.

With a pair of OC16's in push pull you can get 17 to 20W with a 14V supply, which is very impressive for the day though you need a watt of drive power, which of course you could get from another OC16 in class A.

Also, one other interesting thing, using the EF98, it normally has a 10Meg Ohm G1 resistor. So you end up with an amplifier with two active devices EF98/OC16 & two inductive devices transformer/choke. With an input impedance of 10M, and output impedance of what your speaker is say 4R. It takes about 1 to 2V peak at the EF98 grid to get full power. That has to be one of the most spectacular power and impedance ratio transformations off all time, for just the two devices.