Bias Compressive Regenerative Receiver
A properly functioning regenerative receiver requires the sustaining amplifier to provide compressive gain. If that is not the case then the circuit will exhibit hysteresis preventing a smooth transition from the non-oscillating to the oscillating state.
Compressive gain means that the sustaining amplifier will actually provide progressively less gain as the input amplitude increases. There are two ways to provide compressive gain. The first way is on a cycle by cycle basis. This can be done using a differential pair where the non-linear response automatically leads to gain compression. The alternative, as is seen in most semiconductor based regenerative receivers to provide gain compression over a number of cycles by creating changes in the bias point of the active device as the input amplitude increases.
Thermionic tube (valve) based regenerative circuits actually exhibit both kinds of gain compression which usually results in very smooth control of the regenerative process.
It should also be mentioned that amplifier distortion can contribute to the compressive effect.
Amplifier distortion typically increases with increasing input signal amplitude. This results in an increasing proportion of the output signal being converted to harmonics of the input signal frequency. In the context of a regenerative receiver these harmonics are rejected in the feedback loop because of the presence of a tuned LC circuit. This is clearly an additional mechanism by which compression can occur.
As both Q1 and Q2 are run at high collector currents their input impedance is rather low. This requires a large impedance transformation to protect the tuned around L1 from excessive damping. C1 and C2 provide a high ratio capacitor tap to create the necessary impedance transformation. This also greatly alleviates the effects of any change in base emitter capacitance with changes in bias voltage. Q1 and Q2 are jointly voltage feedback biased through R1 and R2. The emitter resistors R3 and R4 ensure that the collector currents of both transistors are roughly similar. R3 is not capacitor bypassed, this improves linearity and reduces the effects of flicker noise. R4 is bypassed to give Q2 higher gain. Actually as the input signal level increases the average collector current of Q2 increases as well. This results in a reduction of the bias voltage supplied to Q1 thus reducing its gain. By this means compressive amplifier gain is obtained by bias point alteration. L2 is the so called tickler coil by which positive feedback (regeneration) is obtained. L2 is wound over L1, the exact number of turns determined by experimentation. If L2 is connected the wrong way around you will get negative feedback rather than positive feedback. The collector of Q2 is connected to the base of Q3 which acts as transistor square law AM detector. Q3 is deliberately run at a very low collector current to prevent feedback from the detector stage from causing circuit instability. Also the low collector current means the base diffusion capacitance of Q3 is at about its minimum value. This results in improved detector efficiency and a higher frequency range. Q4 performs an impedance transformation allowing you to connect the circuit to most audio amplifier circuits.
You may notice there is no regeneration control. Actually the amount of regeneration is adjusted by altering the supply voltage to the circuit. For example you could adjust the supply voltage form 5V to 12V. It is a very good idea to use an adjustable voltage regulator chip such as the LM317 to power the circuit. In any event you should not use a switch mode power supply as they create too much noise. The antenna is connected to the resonant circuit around L1 by a 1pf gimmick capacitor (two pieces of insulated wire twisted together).
This circuit worked very well despite its simplicity. It proved to be very frequency stable and very sensitive.