Challenge: build a microphone AGC circuit with two NPN BJTs as the only active components

Question

I have been trying and failed. Here I have placed a JFET instead of the second NPN, but it doesn't matter, I just can't fork off enough current from the output into the rectifier and smoothing capacitor.

Here are my input wave forms (green) output (blue) and what I get after the rectifier diode (red):

As you can see, there is just not enough juice from the rectifier, and with a lower R3 the output signal is being clipped.

The goal here is to put the output amplitude nicely into the 0-3.3 V range with full swing.

Can it be done? If I need more active components I can just use a MAX9814 and be done with it.

PS: The question was asked if diodes count as active components and let's say for the purpose of this challenge that no, they do not count.

If we have to use a JFET as a variable resistor, we could use the second NPN to amplify the control signal.

Answer 1

Varying the Re (R4) also shifts the DC bias point of Bjt in way the lower R4 the lower Dc voltage on collector.
So you have to adjust DC point so the Vcc limitation doesn’t clamp in widest input signal range. This is adjusted by Rc vs. Re ratio.

Answer 2

If J1 is an n-channel, depletion-mode JFET, then D1 is backwards and biased incorrectly.

JFETs can have an n-type or p-type channel. In the n-type, if the voltage applied to the gate is negative with respect to the source, the (drain-source) current will be reduced

https://en.wikipedia.org/wiki/JFET - 4th paragraph

To save me a lot of typing, here is a detailed explanation of how a FET works in this application:

https://www.interfet.com/jfet-datasheets/siliconix-an105.pdf

Among other things, there needs to be a coupling capacitor between J1 and Q1 to prevent the changing J1 impedance from altering the Q1 DC operating point.

AND, referring to the post title, a JFET is not a bipolar NPN transistor.