Note: groups.io will be down for maintenance this evening, starting at 11pm Pacific Time (6am Tuesday 8/3/2020 UTC), for up to two hours.
I have tried three (and a bit) AGC schemes with the uBitx with the following results:
1. TPA2016 from Adafruit in replacement of the standard audio chain after the potentiometer. Thank you Nik (VK4PLN) for sharing the code.
Like mentioned in another thread recently I also had low volume. To get more signal in I changed R51 to 10K and R50 to 330K. It worked to a point but to get enough "AGC effect" I had to reduce the setgain( ) to close to zero. This meant lower volume.
I tried multiple combination of parameters, but it felt more like a limiter than an AGC.
I take an FT-817 as a reference, although of course I don't expect to get to the same AGC performance, just to be able to listen to a net without adjusting the volume every time a new station comes on.
The final straw for the TPA2016 was when I discovered hash generated on the 40M band. This meant probably more shielding and filtering although I had it mounted on top of a double sided epoxy circuit board the same size as the uBitx main board and connected to ground in the four corners.
So I moved on the next solution.
2. I tried the 2n7000 AGC circuit as shown in the pre-production uBitx schematic.
I noticed that the gain control would come only on loud signals as there is a significant threshold in the circuit before the MOSFET reaches it conductive state.
I therefore replaced the half wave rectification with a voltage doubler version and changed the 1N4148a to BAT46 (Schotky) but still not enough sensitivity to the trigger the AGC on medium signals.
Since I needed low power to drive the MOSFET I used an 8 Ohms to 500 Ohms 350mW audio transformer with the low impedance side connected to the 2nd output of the TDA2822. This gives plenty of voltage even on low signals to control the MOSFET. The AGC action was good but I had distortion, especially on low signals.
I tried to decouple the drain of the MOSFET to the Vol-H input but to no avail.
I read on the internet about "adding half of the signal voltage to the gate of the FET to reduce distortion" for these type of AGC schemes, but I am not sure it applies to depletion of enhancement FETs.
So any experience using this "pre-production" AGC circuit in the uBitx would be greatly appreciated.
3. I also tried and I currently have a MAX9814 from Adafruit in place for AGC, connected strait between VOL-H and the top of the volume potentiometer with decoupling capacitors in and out, through an input and output trimmer pot for adjustment.
I also added a 390nF capacitor in parallel to the 100nF at the time constant terminal (TC) to slow the reaction down to about 1ms.
That solution too is not proving that great. The variable gain range seems too small (although the specs says 60dB dynamic range) as it quickly gets overwhelmed by large signals and introduces strong distortion, or, if adjusted lower, only comes into play on very strong signals.
It is a shame as the TC connection of the MAX9814 give a DC voltage related to the AGC action and could have been used for an S meter input.
So my questions are:
A. Is the pre-production 2N7000 solution (circuit attached below) producing distortion in your uBitx (where implemented) and if not did you modify the circuit in any way?
B. Any other good proven AGC scheme tested with uBitx?
C. Am I asking too much of an audio AGC circuit?
Thanks for reading.
All the best,
73, John (VK2ETA)