Sorry, I forgot that the audio amp sections were different.
I think the pop problem would not be solved with the C38 change. The way it was fixed was between the v3 and v4, where Q74 was added with the gate being driven by the T/R signal from the Nano.
What is the audio level if you remove Q1 on your schematic? If it does not impact the audio level, then the AGC is working. If it increases significantly, then the AGC adjustment is off.
To solve the pop problem I would suggest adding the diode, resistor, and connecting to pin 25 of the Nano (drives K1 through a 4.7k resistor).
As to changing the windings on t16, that would upset the impedances of the mixer and may cause bleed-through of the BFO into the later transmit stages. I doubt that will increase the signal level. The BFO is low enough in frequency that the mixer should be OK. The one mixer that is most in question in my mind is the first, RM1. That has the highest LO requirement (75MHz at the high end) and would push the Si5351a for drive level. Again, I still think that you have an audio issue, not an RF one.
I hope this helps