How it Works
1. D6-D9 form the AGC Detector feeds AGC amp U2B-5 and charges time
constant capacitor C34 through R45
2. Q8/Q9 form the hang circuit and discharges C34 through R54 when Q9
is biased off.
3. The AGC switch (via connector 92, "AGC OFF/AGC TC") functions to:
a. turn off AGC by grounding U2B-5
b. set slow AGC by grounding C33 and increasing hang discharge time constant (C32//C33)
c. set fast AGC by leaving C33 unconnected and decreasing hang discharge time constant (C32)
Circuit explanation from G3TXQ:
U2A is included as an audio phase inverter - that allows *full-wave* rectification of the audio using the D6/D7 + D8/D9 diode pairs. Full-wave rectification provides a faster attack time - the Achilles' Heel of most audio-derived AGC systems.
The inclusion of R46 across C34 provides a leakage path which allows the AGC to follow a very slowly-varying signal; if you don't have it, the AGC voltage pretty much "freezes" at its highest value until the hang circuitry discharges it.
The inclusion of R54 establishes a sensible discharge time for C34 once the hang circuitry activates - it sounds very unpleasant if it's not included.
U2B provides a very high input impedance so as not to shunt the main capacitor C34; Q6 then buffers its output to provide low impedance drive to the IF stages and the S-meter.
Positive-going audio half-cyles which exceed the Vbe threshold of Q8 cause it to conduct and discharge C32/C33; absence of audio allows C32/C33 to charge via R49. Once the voltage on C32/C33 has risen high enough, Q9 is cut off and C34 discharges via R54 and D10.
Various AGC Mod Suggestions
From the Ten-Tec reflector, 7/9/98 (W1EAT)
I haven't tried this out yet on my CSII, but the changes suggested are
to make R45 470 ohms (was 3.3K) and R54 1Meg (was 220K). R45 is the 3.3K
resisitor I mentioned in my last message, and making it smaller will allow
C34 to charge faster. If R45 is too small, the AGC will overshoot and
pop worse, so this needs some experimentation. R54 is the resistor that
discharges C34 rapidly when the "hang" time is over, and making it larger
will make discharge time less rapid.
I am also going to try using a .47 mfd tantalum for C34, instead of the 1 mfd
electrolytic that is standard. There is probably some interaction between
these values, so if I change enough of them I can probably make the pop a
lot worse. Many of my experiments seem to end this way.
What I did was to increase the drive to the base of Q9, the hang discharge transistor, to
cause the hang to start to release before the fast/slow is completely
discharged - add 10K in parallel with R52 100K