Wein bridge oscillator again...

The lousy old LM324 and LM358 opamps were the first low power opamps. Their design saves power by operating the output transistors with no quiescent current (class-B) which causes crossover distortion.

Ordinary opamps use class-AB for the output transistors. Class-AB produces a small quiescent current in the output transistors to eliminate crossover distortion.

You cannot get inside the circuit of the LM324 or LM358 opamps to add quiescent current for class-AB but you can add a pull-up or pull-down resistor to their output to operate one of the output transistors in class-A for no crossover distortion if the current is high enough.

But the LM324 and LM358 also save power by having a very poor high frequency response. They have trouble above only 2kHz. Most opamps work fine up to 100kHz (except a 741 opamp has trouble abover only 9kHz).

Hi All,

I have learned a lot from you all even with this simple topic. Thanks to all electronic experts here in this forum for the generously share your knowledge. Hopefully I can be like one of you guys in future.

Anyway, back to the topic, I found this sine wave generator very stable and easy to tune (up to 5KHz even with LM324) compare to wien bridge and others. But I do not what type of this oscillator is (phase shift / square to sine ?). I probe both of the op amp outputs:
*At high frequency (~4KHz), U1A output more like a square wave, U1B output is a sine wave.
*At low frequency, both op amp output look like sine wave.

And how to simulate this circuit ? I attached the transient simulation setup and the waveform. The output waveform Vpeak-peak is so huge. What have I done wrong ? I do set initial condition for node 5 and node 2 to 1V.

Thanks in advance for the help.
Best wishes,
--CCChang

Hi

Sorry. Uploaded wrong picture for the waveform. Here is the simulation output waveform.


thanks.

Seriously, a real OP model can't generate output voltages outside the supply voltage range. If it's different in your simulation,
you most likely used an ideal OP model.

ccchang said:

Anyway, back to the topic, I found this sine wave generator very stable and easy to tune (up to 5KHz even with LM324) compare to wien bridge and others. But I do not what type of this oscillator is (phase shift / square to sine ?).
..................
Best wishes,
--CCChang

Click to expand...

Hi CCChang,
i think, you have picked up a rather interesting circuit with good properties for your purpose.

(1) the circuit you have shown is one of the classical "double-integrator oscillators".
It consists of two integrating circuits (one inverting, one non-inverting) connected in a closed loop. However, as mentioned earlier, don`t use LM324 (and watch the comment from FvM).

(2) However, if you need to tune the circuit, I recommend to use another topology consisting of two classical inverting MILLER-integrator stages and one additional opamp inverter stage (-RB/RA).
That means: Cancel R3 and C3. Instead connect the ouput from U1A to the inverter input (resistors RA) and the inverter output to R1. Leave R2 fixed.

(3) Tuning: Now you can change frequency Fo with RA or RB without changing the amplitude of the oscillating signal.

(4) 2*Pi*Fo=1/sqrt(T1*T2*RA/RB).
You can choose T1=R1C1=T2=R2C2. In this case, both integrator stages come to their limits at the same output level (soft clipping).

(5) But it is strongly recommended to make T1 larger than T2 (factor 2 is OK). In this case, only the output from the integrator with T2 (higher gain) comes to the limit (soft clipping), but the other one remains at a smaller amplitude with very good quality. Tuning Fo with RA or RB will not change amplitudes.

Correction:The last sentence is true only for additional amplitude stabilization (2 Diodes, antiparallel).

(6) Finally a question which has not been answered yet in the litereatur:
Why will be there a "soft clipping" only? There are many frequencies below Fo for which the loop phase is 360 deg. and the loop gain much larger than unity.
But don't worry, it works as described.