Re: RFI from uBitx TX after AGC and other mods installed

Lawrence Galea

First of all glad that you had good progress.
I hadn't looked at the μbitx board but at the circuit diagram and there is no capacitor showing on pin 7 of the TDA2822.
My thinking was and is that the Q1 base - emitter junction is acting as a detector and feeding back the audio to the AF gain control through the input capacitor C1.
So you can try a higher value and see whether all the RFI will go away.
Regarding filtering at the voltage regulators, I always put the capacitors I suggested on both input and output as close as possible to the regulator, if possible across the leads on the pcb and also a 10 to 22μF capacitor on both input and output and never had a problem with RF on the regulators.
Furthermore, when a big voltage drop is required such as from 12V to 5V I usually put in a 7809 or 7808 before the 7805, splitting the heat between them.
They are cheap so better regulation (within their capacity) and less heat in each component and better component life.

As a note, we used to use different capacitor values when we used leaded components because due to the lead inductance they would not decouple efficiently at some frequencies.
We also used certain capacitor values with a particular leads length where the capacity with the leads length formed a series resonant circuit to efficiently decouple a particular frequency especially in screen grid circuits.

On Sun, Jul 1, 2018 at 9:02 PM, Rick Price <rickprice48@...> wrote:
John, Don Lawrence, others
Interesting Outcome! Shielding was no help, replacing pcb with perfboard agc helped a little, but back to pcb. bypass's on 78L05 were .68 on input and .1 on output still no help and removed 78Ll05. Did not try 1nf,10nf,100nf in parallel on 78L05 because I had already removed it.  Because pin 7 of U1 (TDA2822) is already bypassed with C78 did not do. Don suggested to filter the 5v supply. I get the 5 volts from the radino pin 3 (green wire) and use it for both AGC and mike preamp. Used a 470uf and .01 on the 5V no improvement.  Don suggested not to try the emitter-base capacitor because of forming a blocking oscillator. It looked to me that because the emitter was already grounded that it probably just bypass any unwanted signal to ground and not oscillate, connecting to the collector might have been another story.. What the heck nothing ventured nothing gained. Tried the .01 and turned on expecting possible oscillations of some sort. AGC off transmitting nothing as before, AGC on transmitting at full volume only occasionally get a very small occasional hint of audio coming through the speaker no oscillations of any kind. Before as soon as audio was present from the mike it started coming through and sometimes went into oscillations.  Hardly ever run at full volume so it looks like your theory and suggestion Lawrence may work. ARC still appears to work as before but will need a strong signal to verify.  Still have to experiment with a couple of different values to see what's best. Hope this spells relief with the agc on for the few of this it affects. Thanks to all for the input and suggestions.

From: [] On Behalf Of Lawrence Galea
Sent: Sunday, July 01, 2018 12:28 PM

Subject: Re: [BITX20] RFI from uBitx TX after AGC and other mods installed

Also try putting a 100n directly across Q1 base-emitter on the AGC board.
It will cut off some of the higher AF as it would be in series with C1 to earth across the Af gain control.  
However, this would not be of any serious consequences.
If you don't want as much AF HF cut off you can try a smaller capacitor, say 10n.
I am thinking that Q1 base-emitter junction is acting as a detector and the AF is being coupled back to the AF gain control since the level is varying with the AF Gain control.
One could also add say a 10n capacitor directly on pin 7 of the TDA2822 or across the AF Gain control as an rf decoupling.
Can you try and report?

On Sun, Jul 1, 2018 at 5:33 PM, Lawrence Galea <9h1avlaw@...> wrote:
Did you put 100n, 10n, 1n filter capacitors  all in parallel on both input and output directly across the 78L05 leads?

On Sun, Jul 1, 2018 at 2:31 PM, Rick Price <rickprice48@...> wrote:
John, Don, others,
Yesterday I attached a 78L05 (TO220) directly to the back of the AGC board that is mounted to the RF gain control on the front panel. Bypassed all power leads and ran a separate power lead to the 12V regulator that powers my uBitx. In other words the AGC board had its own isolated(?) power source. Stll heard the audio when transmitting.  Beginning to think that radiated power is being absorbed by the AGC board and then recitifed and then coupled into the audio system. I also have the pop click board installed and not sure if this was a problem before it was installed. My AGC board is one I fabricated using Don's technique of cutting pcb with an x-acto knife.  I also have a AGC board made on perf board using through hole parts.  May sub that in and see if there are any changes. I eventually plan on installing the JackAl mod when it becomes available so this problem should go away for me but its beginning to become a challenge to fix.

From: [] On Behalf Of John KC9OJV
Sent: Sunday, July 01, 2018 8:05 AM
Subject: Re: [BITX20] RFI from uBitx TX after AGC and other mods installed

I spent part of yesterday applying ferrites to this and that to little success. I also tried wrapping the 5v line around the twisted pair coming from the volume pot and that did some good. This morning I was thinking about fixing a 7805 to the metal base of my case, drilling a hole in the ground plane and routing a wire through that hole to feed the AGC board. It could be driven off the 12v relay (K1?). Has anyone else had any ideas/success?


Join to automatically receive all group messages.