LM386 Wien Bridge Audio Oscillator without a bulb

Here is my first cut at the LM386 Wien osc using diodes instead of a bulb to stabilize the waveform:

the resistor in the middle of the page without a value was 51k ohms, about 10x the 4.7k ohms

Here is a scope picture:

Visibly distorted although certainly listenable.  The circuit is very stable, the 10k ohm pot can almost function as a volume control.  There is no big bounce as the pot is adjusted as was found with the bulb stabilized Wien mentioned yesterday.

I removed the diodes and ran the oscillator open loop.  The result was touchy but visibly undistorted:

it looks nice but the problem is that it won't stay this way

So there are some items that remain:

0) can the diode stabilization be changed to reduce distortion ?  I know the op-amp Wien Bridge oscillators stabilized with diodes have less distortion. see item 2

1) use a different stabilization circuit with less distortion than the diodes but is more stable than a light bulb (probably a JFET or MOSFET with a diode detector ?).  Diodes are less complicated, available and cheap though.

2) "blueprint" the circuit values, these were kinda just pulled out of thin air (a 10 turn pot makes things nicely adjustable).  Reconsidering the circuit, the diode feedback resistor value needs to be increased ?

3) is frequency adjustment possible ?  With the 10x difference in the resistor values, a dual pot is out.  Possibly a dual pot could be used with 10x difference in cap values ?

4) some kind of smooth keying circuit to allow using the circuit as a CPO.

5) some kind of volume control - a series pot on the speaker would be OK or an L-pad but is there another way ?

It's always fun to dig in and try these things out !

BTW, here is the original National Semiconductor circuit from their Audio Handbook:

 

It pays to find the original sources, the correct formula for frequency is shown, alas the lamps have become scarce and expensive. 


Best Regards,
Chuck, WB9KZY
http://wb9kzy.com/ham.htm