03-09-2021, 12:28 AM
(01-09-2021, 05:16 PM)BusyBee Wrote: Your delayed amplified AGC sounds about right Kalee.
Here it is, as at present.
I have a main IF amplifier (2-stage - the advantage of two stages is that each needs to have a gain variation of 10:1 to give an overall 100:1 range). The main IF amplifier feeds the audio demodulator in the normal way.
Prior to the audio detector, I pull off some IF voltage and feed to a side chain, as shown. First, there is a low-gain additional IF amplifier which serves as a buffer. This feeds a leaky-grid detector, the output of which is heavily filtered to remove audio. Because the leaky-grid detector is reasonably linear, the filtered output has a DC level which depends only on carrier strength and not on programme content.
The output of the leaky-grid detector goes more positive with signal strength, not what we want, so a further valve inverts it. A negative bias on the DC amplifier's grid keeps this cut off (so the anode voltage high) until the signal strength is enough for the leaky-grid output to be high enough to overcome the delay bias. Once this happens, the DC amplifier output starts to fall.
The outbid the DC amplifier is scaled and level-shifted and applied to the main IF amplifier screen-grids and suppressor grids.
The circuit is not optimised as yet - the various potential dividers could be combined, for instance - but it works!
Readers may raise their eyebrows at three power rails, however the 1.2V and 120V are standard LT and HT rails; the negative bias line is needed elsewhere (bias for 1p24b output valves) and also for a heavily-morphed infinite impedance detector, so it is already present.
