Hi Gents,
I wonder if anyone has an idea how to implement the AGC into QCX.....Tnx for hints !
Jarda OK1MKX

I wonder if anyone has an idea how to implement the AGC into QCX.....Tnx for hints !

Jarda

https://groups.io/g/QRPLabs/message/26070?p=,,,20,0,0,0::Created,,agc,20,2,0,24613271

73 Alan G4ZFQ

Generally there are three ways...

*Audio compressor in the audio stream.  The problem here is while the audio is 
kept from getting too loud if the station is strong the audio system will overload
and the use hears distortion.  There is a limit to dynamic rage of such a system.
Getting the levels and time constants correct for best possible dynamic range
is tedious.

*Use the audio to control a diode or fet based variable attenuator at the RX input.
This is before the receiver and both protects from overload and distortion
as well as keep loud signals from causing ear pain.  Biggest thing is connecting
to the receiver after the TR switching and before the receiver input.  Second is
getting the agc time constants correct. 

Manual switched RF attenuator (or variable) to lower the RF in for strong signals.
Its manual, just like volume control.

All work and both have the specific issues to implement.

I've implemented the RF AGC in a similar receiver using diodes with good result.
The system samples audio from the volume control rectifies it to DC that varies
with audio level and after amplification applies that to the electronic attenuator
(in my case using diodes).   Obviously I had to design it.   However it was easier
to do as I controlled the board layout and allowed room.   

The insertion point for such a circuit would be between Q5 and T1.

Allison

I have built and tested the digital AGC designed by Willy, described at this link.
Works quite well.
72 Fred WD9HNU

        Dr OM Jarda ,

Indeed, Digital AGC will solve you problem .

This is a project based on a digital POT and a PIC12F683 Pic µP .

Pass me your E-mail address and I will send you the full doc .

Note this project is open for hardware and software .

WD9HNU reproduced this 3weeks ago with success . Maybe you can ask him for a duplicate of the PCB , however this is also proposed in the doc , basedd on Kicad .

It's your choice .

in attachement the response curves of this project .

on5kn@...

73 Willy ON5KN

Hi, Jarda

Here is a little circuit I added to my QCX. Not shown on the schematic is a 2.2k ohm resistor that I added between the middle leg of the volume control and the board. The 2N7000 transistor is installed between the ground end of the volume control and the board connection where the volume control middle connection goes.

To set it up, connect an antenna to the radio and set the volume to max, tuned to a clear channel (no signal). Starting with the 1K control set to the ground end, slowly rotate the control clockwise until you hear the band noise drop slightly. Back up slightly so the noise is not reduced.

Using the component values listed there is no "pumping", no loud blast and then reduce. It sounds like a "big radio", with signals against a quiet background, and background noise coming up when the signal fades to a bar or two on the signal indicator.

Jim W4JED

Hi Anan,

I would be nice to post a link to the data package itself.

John

Hallo Jim W4JED

The AGC circuit looks great and seems easy to duplicate, so I'd like to try it on my QCX. I need some guidance.

- From where exactly did you take the input, the drawing shows EAR
- what type of schottky diode did you use
- The 2K2 resistor goes from the pot wiper to the board, you mean ground?

73, Frank
PA7F

Hi, Frank

"Ear" is the high side of the earphone jack. There's plenty of audio current here to make the "attack" very quick - so quick you don't hear it.

The schottky diodes are the glass "yellow with black stripe" ones that sell for 50 for $1.58 on Ebay - I believe they are BAT85. They work good for SWR meters, AM crystal radios, anywhere you need a small, low forward voltage diode. The first 2 are a voltage doubler and the third isolates the gate of the 2N7000 from the rest of the circuitry. That way the charge on the gate can bleed off in a controlled fashion, but can be recharged quickly if necessary.

To install the 2K2 resistor you disconnect the volume control center lug, bend it away from the board, and connect the resistor between the lug and the now vacant hole on the board. The 2N7000 (or VN10L, or whatever you have handy as long as it conducts at less than 5 volts on the gate) goes between the ground end of the volume control and the hole where the resistor is, where the middle lug on the volume control used to be.

The 1 K trimmer pot sets the bias on the diodes and the fet so everything is just barely out of conduction at no signal. As soon as something gets to the earphone the diodes and the fet conduct and put the brakes on.

Have fun not "riding" the volume control.

Jim W4JED

Hi,

I modified the circuit of James W4JED a little bit, because I've special needs:

* When doing SOTA, noise is very low and signals may be very weak, and with my earphones I often need to set volume control to maximum. With the 2K2 resistor of James' circuit in series with the vol control wiper, I would loose some volume, what I want to avoid.
* When having volume control not set to max, sidetone loudness it just right, but with volume control set to max, it is much to loud, even with AGC.

I changed the original circuit by putting the MOSFET _in series_ to the volume control wiper and added two diodes in order to limit the loudness of the sidetone, see the attached schematic.

Personally, my ears are very satisfied with this now, I'm glad to have it.

Remarks:
* The resistor marked with * must be adjusted like James describes it. Of course, you can put a potentiometer there.
* The Diodes D101 and D102 must be of equal type, otherwise the sidetone will create a DC-voltage, which will result in a popping sound.
* Both diodes prevents rectification of the sidetone by the body diode of the MOSFET, which might result in adverse effects.
* For D101 and D102, BAT 41 was to loud for me, BAT 43 was just right (look at Vf @ 10 mA in the datasheet)
* The sidetone is like a very strong signal for the AGC, thus making QSK quite deaf.

Kind Regards,
Axel
DF1ET

Hi Axel,

do I understand it correctly that the QSK I now got used to and I appreciate, is becoming pretty useless if the signals are not as strong as the sidetone ?

73's
Martin
DK3UW

Sorry to not understand the last message but what is link between QSK and AGC? Is this a new topic?

In QCX, AGC is strongly related with QSK. Sidetone goes the same way in the circuit as the received signal. And sidetone behaves like a very strong signal, hence AGC makes the RX quite deaf for maybe 100 ms.

Martin, I use QCX only for SOTA activations, and up to now, I did only one with AGC (will do more in the next 2 weeks, hopefully). My impression was, that QSK is not useless, but worse with AGC. But I need more experience from my upcoming usage.

If QSK became to bad, we can make the decay time of the AGC much faster, for example by replacing the 3.3 MOhm resistor by 330 KOhm.

Kind Regards,
Axel, DF1ET

My experiences with QSK with AGC now are (speed: 15 wpm): Between words, AGC's decay and receiver's sensitivity is enough to here the received signal. Between letters, I had once the impression to hear a received signal, but would say, generally sensitivity is to low between letters.

I intend to either increase AGC speed or to change the way sidetone is fed into the audio stage.

73
Axel, DF1ET

The problem with AGC on any radio is keeping TX operations from
pumping up the RX agc since most derieve AGC control voltage
from RX audio.

Two things;
Minimize the side tone level.  As tx sidetone will also drive audio agc.

Insure any rx to tx transfer does nto cause pops or clicks as that
also show up int he audio and will drive most audio agc to become
active (or over active).

Clamp the AGC to some level that represents a no signal receive
during transmit.  That allows the AGC to only work on receive.

If those things are properly done then agc on RX will not harm
the QSK capability.  Didn't say it was easy.

Allison

Hi Allison,
Thank you for your advice! I think, this will be very helpful. I'm curious about what the result will be.

73 Axel DF1ET

QSK with AGC is much better now. I would say, it's really good now (and AGC, too). Main problem was not the high level of the sidetone, but an RF induced voltage into the AGC circuit. An additional 100-nF-capacitor and 100-Ohm-Resistor prevents this, please see the attached schematic.

My impression is, that inserting Q100 at a slightly different point additionally improved QSK a little bit: Now I cut the track to C22 and inserted Q100 there. Hence, Q7 is now connected to the source of Q100. Before, I disconnected the wiper of the volume control (which is easier), and inserted Q100 there, which means that Q7's was connected to the drain of Q100.

73
Axel, DF1ET

Maybe, I should add: At 15 wpm (my speed), I can even here signals between the dits of, say, letter "h".

Axel

            Dear Jarslav & Alan ,

Yes , indeed I developed  an audio AGC for the QCX .

Just send me a mail tot hion5kn@gmail and I will send you the files , schematics , board layout and software file .

You have to own a PIC 12F683 , digital potmeter MCP41010 and to know how to upload a program to a PIC 12F683 with adapted programmer . You surely will find someone in your club to guide you .

73 Willy

ON5KN