Simulation of Bubba oscillator

[hfc1993911hfc]

I am try to generate a 50-60 Hz sine wave using the bubba oacillator. I have my in Multisim but i dont get the desired result.The signal was just limitless amplified and i don't know why. My schematic is:
and this is my result of oscillator at begning. but soon ,it becomes like this：

 

[zorro]

Hi,

It seems that your model for the OA is "too ideal". Apparently clipping is not included.
On the other hand, it seems to me that the left terminal of R10 is misplaced. It should be connected to note 7 instead of node 6.
Regards

Z

 

[hfc1993911hfc]

Thank you. I've rectified it. Here it is: But the result of oscillator becomes like this:.Apparently it is not a sine wave. And by the way,how should i do to take clipping into consideration?

 

[zorro]

Now it fails to start oscillation.
Loop gain is greater than 1 at the frequency of 360º phase shift, so it should oscillate.
I don't know how this simulation is done. Maybe it is into an unstable equilibrium point.
Can you introduce a perturbation, for example a small change in VDD a short time after T=0?

It's not important, but why did you change R13 fron 27k to 31k?

Z

 

[kam1787]

**broken link removed**
https://www.changpuak.ch/electronics/Oscillators.php
**broken link removed**

 

[hfc1993911hfc]

In real-life, there is always a perturbation in the air,so it doesn't need to introduce one,but I don't know how it works when in simulating mode. The change of R13 doesn't matter.but I'll correct it then.

- - - Updated - - -

Thanks for your material,and can you help me with my problem ?

- - - Updated - - -

Thanks for your material,and can you help me with my problem ?

 

[FvM]

The oscillator doesn't start because the initial transient solution finds a perfect steady state. The most simple way to start the oscillator is to skip the initial transient solution, it's a SPICE option. Otherwise you need to apply a "kick-start" pulse.

 

[LvW]

And by the way,how should i do to take clipping into consideration?

Each harmonic oscillator needs a lop gain of EXACT unity. This never can be achieved (tolerances).
More than that, for a safe start the loop gain must be somewhat larger than unity at t=0 (switch-on).
Thus the amplitude rises until it is clipped.
Solution: To avoid clipping a kind of non-linearity must be introduced which automatically reduces the gain for rising amplitudes (two antiparallel diodes, FET control loop,...).

 

[FvM]

Solution: To avoid clipping a kind of non-linearity must be introduced which automatically reduces the gain for rising amplitudes (two antiparallel diodes, FET control loop,...).

For standard applications that can accept a few percent harmonic distortion, a Bubba oscillator works well without dedicated gain control if the output signal is taken from the last OP buffer output (U1D in your schematic). By adjusting the loop gain to a value slightly above the required minimal value (ideally 4), the clipping in the amplifier stage U1A and respective harmonic distortions will be low.

 

[LvW]

Yes - correct. By doing this we exploit the lowpass filtering capabilities of the circuit.
That is the most important property of the BUBBA oscillators - however, we pay a price: 4 opamps.
There are other oscillator concepts with one single opamp only, which, however, need such a non-linear device.

 

[hfc1993911hfc]

I want to know if i need to apply a 'kick -start'pulse in real desgin. Actually,I have made a PCB of it but it appears that the magnitude of the sine wave is very low ,which lies in uV class. And i don't know what to do..Can you give me some advice?

The oscillator doesn't start because the initial transient solution finds a perfect steady state. The most simple way to start the oscillator is to skip the initial transient solution, it's a SPICE option. Otherwise you need to apply a "kick-start" pulse.

 

[FvM]

In a correctly dimensioned circuit, the oscillation will start automatically. Something is probably wrong with your design.