How to add automatic frequency control to varicap voltage?

Question

I've built a TV tuner radio using a tuner with voltage synthesis tuning and a CD2003 FM demodulator. A similar schematic can be found here. The reception quality is very good, but it is difficult to tune in a station even if I use a multiturn potentiometer. The frequency tends to drift away. So I've added an AFC (automatic frequency control) function.

^{simulate this circuit – Schematic created using CircuitLab}

This works great. Frequency is direct proportional to tuning voltage. When tuned on a station, if frequency drifts, audio signal level drops and modifies the varicap voltage.

However, I believe that the DC tuning voltage may affect my demodulator IC. Here are the voltages measured on Test Point:

                   No signal                Tuned on a station
Without AFC          1.10 V                      1.50 V
With AFC             1.98 V                      2.25 V

To tune the entire FM band I need a tuning voltage of 10 to 14 V (the tuner covers 45 to 160 MHz, I only need 87 to 108 MHz). The voltage with AFC was measured when tuning voltage was closest to 14 V (maximum).

This is the schematic (without power supply and audio amplifier):

This is the tuner pinout:

This page shows how to add AFC to a LC oscillator. But things are different in case of a varicap oscillator.

Is there any chance this voltage imposed by AFC will affect my demodulator? If yes, how should I do it correctly?

Answer 1

This is how I understand AFC: -

Typically, an AFC circuit works by examining the IF frequency for being exactly 10.7 MHz or 455 kHz (AM). The IF frequency is the result of mixing the local oscillator with the incoming received RF and if the local oscillator drifts off the IF frequency changes and the FM demodulator will suffer and you'll hear audio distortion.

Because the local oscillator isn't fixed (it has to be at least as wide in range as the incoming RF) it is prone to drift however, if the IF output is not (say) 10.7 MHZ, this can be detected by a phase comparator that has a fixed 10.7 MHz oscillator reference.

The phase comparator will not have a dc output - it will rise and fall with the modulation but its average value will represent the "error" of the local oscillator from 10.7 MHz. This error is low pass filtered (sub 5Hz e.g.) and amplified and fed to the varactor device(s) that control the local oscillator.

Hence, once an RF station is close to being acquired (by course tuning), the AFC kind of takes over and nudges it into near-perfect "lock".

I really can't see how your AFC is intended to work.