I'll put it here.
This is the AGC from the Grant, same thing the cobra 148 uses. I want to make sure I am understanding how all of this works. Here it goes.
Starting at the front end, there are two control signals that can shunt incoming RF by forward biasing D12, the AGC control voltage, and the TX supply voltage (so the RF from the finals during TX don't fry the front end). A positive voltage on either of these will forward bias the D12 and cause a reduction in signal making it to the RX preamp.
That AGC voltage comes from TR10. For there to be a positive-going voltage on TR10's collector (increasing attenuation at the front end), TR10 must be turning off allowing more of the RX supply to go to D12. For that to happen, the base of TR10 must be going lower. To make the base of TR10 lower, the gate of the FET must be going down with respect to it's source. A negative-going voltage at the gate of the FET increases front-end attenuation and simultaneously alters the bias point of IF AMP TR16. In other words, making the gate of the fet more negative causes attenuation at the front end AND at the IF amp, a common technique to increase dynamic AGC range according to Lou's book.
With RX RF coming in to the anode of D10 from TR18, the positive peaks of the RF are clipped at 0.3v leaving the negative-going parts unaffected. This shifts the DC average at the anode of D10 negative with respect to ground. This negative voltage is transferred to C23 and C25 via D8 and D9 respectively, and for the moment, D5 is not conducting. This suggests to me that the initial response time of the AGC is governed solely by C23. However, the decay of the AGC action (in AM) is governed by C25. As C25 discharges through the shunt resistor next to it, that negative voltage on C25 moves back up towards 0v. When the voltage on C25 goes higher than the voltage on C23, D5 begins to conduct and starts turning off the AGC.
In sideband, D6 passes SSB voltage to D7 bringing C24 to (nearly) ground and in parallel with C25 changing the time constant for SSB AGC decay, making it take longer to deactivate the AGC function.
So, going back to the issue at hand. It would seem to me that C23 would have to attain a negative voltage in SSB TX for the AGC to be at blame for the momentary absence of RX when unkeying. With the RX IF chain supposedly off when PTT is pressed, I am struggling to see how this can be the case. Where does that negative voltage come from?
Therefore, unless my understanding of the AGC circuit is flawed or there is RF in the IF chain during TX, I do not believe we can blame any of the AGC caps. I think we need to look elsewhere.
