yeah but implementing automatic gain control with J-FET is too complicated for this application and not necessary but could be a solution.
The voltage gain from the non-inverting input to op-amp output should be exactly +3.00 to keep the amplitude of the oscillation within the linear region of the amplifier. With less gain the oscillation will cease and with more gain the output will saturate. An incandescent lamp (which is represented by the variable resistor R4) is used as feedback element to to maintain the +3.00 gain.
I believe I almost figured it out! my mistake was having the value of the variable resistor, R4, as a linear function of time. After reading the book more carefully, I noticed that it should be a function of the output voltage, not time. In real life, the temperature of the incandescent lamp will increase with increased applied voltage and therefore increase of resistance. I changed the value of the variable resistor from a linear function of time (R = 1+1000*time) to a quadratic function of the output voltage (R=1+100*V(out)**2).
Now finally, I have a sinusoidal output signal but not at the desired frequency. I calculated the resistors and capacitors to be 80K ohm and 10nF for 2KHz and the output signal is measured at 160Hz.
I wonder if the problem could be related to a phase shift? The idea is to make a feedback amplifier with 180 degrees phase shift at the desired frequency. Any ideas?
I have attached my latest simulation results for reference.