R3, R4, C5, L2, L3 and Q2 together form the active mixer circuit. Local oscillator injection from the oscillator Q1 is via coupling capacitor C4 and the RF input from the antenna is coupled into the collector via capacitor C6. The detected audio, extracted from the emitter, is coupled to the audio amplifier U1 via capacitor C15. ** Q2 ** Q2 performs two functions in the KnightSMiTe. During receive, it operates as an active mixer serving as a product detector for signals at or near the frequency of the injection (local oscillator) circuit of Q1. This process is called direct conversion since the desired signal is detected (demodulated) at the operating frequency avoiding the process of conversion to an intermediate frequency (IF) as is done in superhetrodyne receivers. ** Q2 (as a mixer) ** The oscillator switches Q2's base emitter junction on and off causing the RF signal applied to the collector to combine in the emitter. Q2's emitter current is highly nonlinear when its base emitter junction is switched in this manner. This produces strong sum and difference components (as well as the oscillator and input RF signals) at the emitter. An approximate conversion gain of 4 dB was measured on one prototype of the KnightSMiTe. ** R3 ** R3 forms a voltage divider with R4 while providing only a trickle forward bias across the base-emitter junction of Q2 which remains biased Class-C due to the shunting effect of L2 which parallels (shunts) Q2's base-emitter junction. The quiescent current in the emitter of Q2 serves to ensure the linearity of the detected audio extracted from the emitter junction and minimizes large signal distortion. R3 is to some degree redundant in the presence of D1 which biases Q2's emitter above ground via R5 and can thus be removed without degradation in performance. Without D1 however, R3 is required and the following applies. R3 is not critical and any value between 10k and 50k should work well. Decreasing its value increases LO radiation while increasing it may lead to audio distortion. ** C5 and R4 ** C5 and R4 form an audio highpass filter and bypasses all but the difference components (audio frequencies) appearing in the emitter of Q2 to ground. What remains becomes the receive audio which is coupled to the audio amplifier U1 via C15. The RC time constant set by C5 and R4 establishes the detected audio frequency response of the receiver, passing only the difference frequency components (audio frequencies) created by the mixing (frequency multiplication) process in Q2. Any change in value of R4 would require an inversely proportional change in the value of C5 to maintain the same response. ** R4 ** R4 raises Q2's emitter voltage above ground to ensure the detected output of Q2 operating as an active mixer doesn't distort in the presence of high level input signals. The mixers output must remain linear over the receivers dynamic range. The bias on the emitter of Q2 permits large negative excursions of the detected waveform without clipping. The voltage only needs to be raised to 1/800 of the power supply voltage which represents the peak-to-peak input level where distortion in the audio amplifier (with a voltage gain of 200) begins to dominate. Q2 would operate well as an RF power amplifier with the emitter close to ground however, and local oscillator radiation would be excessive. Raising the value of R4 increases the emitter voltage placing Q2 far beyond cut off. With collector current significantly limited, LO radiation is reduced. An excessive increase in R4 will result in weak oscillator injection causing distortion on large signals due to reduced dynamic range. ** C5 ** C5 grounds the cathode of switching diode D2 (reverse biased during receive) to provide immunity to noise on the keying line. During transmit, C5 and R4 are shorted to ground and not functional with the key closed. C5 snubs any tendency of the key to arc when opening and key clicks are thus avoided. ** L2 ** L2, as a near short circuit to DC, prevents R3 and R4 from biasing Q2 into conduction. It's primary function is as an RF choke appearing as a high impedance at the operating frequency. This ensures Q2 is turned off until a signal is injected via coupling capacitor C4. The positive half of the oscillator signal causes the base-emitter junction of Q2 to conduct while the negative excursions of the signal allow Q2 to turn off. Q2 is said to be operating Class C when driven in this manner (i.e. emitter current flows over less than half of the input cycle). Class C amplifiers are highly non-linear and by heavily restricting collector current power amplification is suppressed and the device becomes a mixer. The value of L2 is not critical but needs to provide a high reactance at the operating frequency. Too low a value reduces coupling efficiency and too high a value consumes valuable board space. ** L3 ** L3, also an RF choke, serves as Q2's collector load. Q must remain high for good coupling efficiency at higher values of L3. A smaller value will reduce coupling efficiency if its reactance is low relative to the output impedance at the antenna port.
If you are interested in the Knightlites and Amateur Radio contact Paul at