Silicon Implemented Oscillator

The oscillator section for the Autocontrol II is implemented in silicon.  I managed to find a datasheet for the old germanium transistor that was originally in one of the units that I had.  This implies that it had been serviced at some time and the original manufacture marked transistors had been replaced by something off the shelf with 2N markings.

The old transistor had an hfe of about 80 and a saturation frequency of about 60 MHz.  A bog standard 3906 is better in every possible measurement than the older transistor.

This is a common base Colpitts oscillator.  Q11 is doing the work.  R20 and R21 form the dc base bias at about 9 volts.  C12, C13 and C1 keep the rf signal off the dc bias or feedback limit networks.

The tank circuit at L6 and C11 set the frequency of operation.  L6 is in the attitude indicator, and C11 is the sum of the cable impedance, feedback capacitor in the servo, capacitor in the heading indicator, roll capacitor in the unit, and trim capacitors.  It should oscillate between about 9.5 and 10.h MHz as these various items move around.

The rf signal in the tank circuit varies around rf ground (0 volts).  Its maximum value is set by the emitter of Q11 - about nine volts.  R18 limits the gain so that the signal doesn't get too big, drive Q11 too hard, and distort the sine wave.

C3 feeds the rf signal back to the emitter.  The emitter of Q11 oscillates between the base bias (it can't go lower) and 9.6 volts above the base (it can't go higher due to the gain).  So Q11 provides a short shot of current to the tank circuit each cycle as the emitter feedback voltage just goes over 9.6 volts.

Q12 is an emitter follower that takes the 9 - 9.6 volt signal from the emitter of Q11 and drives it into the discriminator coil (L7 and C14).  This isolates the oscillator from the discriminator circuitry and provides a loud clear constant amplitude signal.  This is where the major change to the circuit is needed.  Due to the much higher gain of the silicon transistors the emitter resistor R22 is roughly doubled from the original 2k to 4.7k.  This keeps it a commonly available resistor and halves the current drive to keep the voltage across the discriminator from getting out of hand.

I have found various references that discuss converting germanium circuits to silicon and a good rule of thumb is to double the value of the emitter resistances.  This seems to be consistent with that.