April 1976
A serious deficiency of the Collins R392, an otherwise excellent general-coverage receiver, is the severe distortion produced by strong ssb signals. While turning down the rf gain control solves the problem, it is extremely inconvenient when dealing with a variety of signals of different strengths -- as when tuning across an amateur band.
The R392 was designed primarily for a-m and fsk use and uses a simple diode detector. Experiments with a product detector and outboard audio stages produced no significant improvement. Observing the signals at the i-f output connector on the front panel with an oscilloscope revealed that strong ssb signals caused considerable flat-topping in the stages ahead of the detector. When the rf gain control was backed off to the point where clipping did not occur, the detected audio sounded fine. This indicated that the problem was due to the agc system (designed for a-m signals) more so than the detector. Ssb reception may be improved using the audio-derived hang agc circuit of fig. 2.
A hang-type circuit derived from that described by Hartke1 is used. Q1 provides some amplification and high input impedance to prevent loading of the audio circuit at J615. The output of Q1 is rectified by CR1 and gated by CR5 onto C2, the agc hang capacitor. Simultaneously, CR1, CR2, CR3, and CR4, in a voltage quadrupler configuration, develop a control voltage gated to C1 via CR4. This control voltage keeps 02 pinched off as long as a signal is present and permits C2 to remain charged and maintain the negative agc voltage. Upon removal of the signal, C1 discharges through R2/R3 and -- after the hang time, when the pinch off voltage of Q2 is reached -- it conducts and quickly discharges C2, restoring the receiver gain.
Source follower Q3 acts as a high input impedance buffer between C2 and the receiver agc line. R4 is required to provide a dc return for the gate of Q3. The time constant R4C2 is so long relative to R2C1 that the presence of R4 causes no significant discharge of C2 during the hang period. R5 serves to adjust the dc offset at the agc terminal. The minus supply voltage is not critical, as long as it is stable (I happen to have an 8-volt zener on hand). Strong signals charge C2 negatively, reducing the drain current of Q3 and causing the source voltage (drop across R5) to move in a negative direction.
No major modification of the R392 is required. The agc circuitry was built on asmall piece of perforated board and mounted in the external power supply unit. Connections to the detector output and the agc line were made at test points J615 and J614 respectively, on the audio module (lower deck assem- bly). In my case, lengths of miniature shielded cable were fitted with pin plugs to match the J614 and J615 sockets and the other ends wired to unused pins on the receiver power socket, J103. The connections to the agc module in the power supply were thus made via the power connectors, again using miniature shielded cable. Pin J on the receiver power connector is unused and is directly available. Several other pins are only needed when the original matching transmitter and microphone/headset are used. Therefore, the wire to pin C was unsoldered, capped, and this pin on J103 was used for the other agc connection.
The front panel must be removed to gain access to the rear of the power connector. However, this is a relatively simple procedure following the instructions in the manual.2 With the panel off, capacitor C104, 0.47 mF, which is switched into the original age circuit by the bfo switch, should be disconnected. This improves the attack time of the new agc circuit. This is the larger of the two capacitors mounted on the inside of the front panel near the bfo on-off switch.
The required voltages are derived from the power supply unit shown in fig. 3. Simple zener diode regulators would also suffice since the current demands are only a couple of milliamps.
Adjustment is simple. First, with no input signal to Q1, source resistor R71 is selected to provide a static drain voltage of about 7 volts to ensure maximum dynamic range at this stage. Next, with the connection to J614 also open, R65 is adjusted for zero volts at the agc output terminal. The circuit is then ready to go. The added agc required no butchering of the original circuitry and has resulted in a dramatic improvement in ssb reception performance.
1J.L. Hartke, WIERJ, "Solid-State Hang AGC Circuit for SSB and CW"
A. D. Lightstone, VE3LF
from Ham Radio Magazine, July 1977, pages 88 and 89