# What's wrong with this Phase Shift Oscillator?

I am trying to design a phase shift oscillator, using a CE Amplifier. I have a Vc of approximately 6V, gain of approximately 40(>29, as required). I aim for an oscillation frequency of 1.5 kHz.

But whenever I feedback the CR network, no oscillations take place. Can anyone tell me what is wrong with the circuit?

simulate this circuit – Schematic created using CircuitLab

• Can't read value of Rc on schematic. But if it's too low, try increasing gain by reducing AC value of Re with a capacitor across it – user_1818839 Apr 10 '13 at 8:31

## 2 Answers

Works for me, if you make a few tweaks to the circuit:

Adding R7 and C4 let you get more gain out of the circuit, while keeping the DC operating point approximately the same. Your circuit wasn't oscilating because the collector of Q1 is "loaded down" by the parallel impedance of the phase shift network, therefore the actual gain of the stage was significantly less than the standard $\frac{R_c}{R_e}$ formula would indicate. Most analyses of the phase shift oscillator assume that the output of the gain stage is low impedance and isn't affected by the phase shift network, and also that the phase shift stages themselves don't load each other.

Sometimes when simulating an oscillator circuit it's also necessary to start the supply voltage at 0 volts, to "kick start" the oscillation, as I've done here.

• Nice to see you got it working. Personally I have always found this type of oscillator to be way to fiddly to get working properly. – Michael Karas Apr 10 '13 at 14:30
• @MichaelKaras I agree. Using a quad op amp you can build a buffered phase shift oscillator, or a state-variable oscillator that will make lovely low-distortion sine waves. An LM324 will work fine at low kHz and is only about 10 cents. – Bitrex Apr 11 '13 at 9:34
• Indeed. And if you only need a square wave you can make it with a single comparator, one cap and 4 or 5 resistors. Capable of a large range of frequencies. – Michael Karas Apr 11 '13 at 10:43

The following is what I'd ask myself in order to prove that the circuit is capable of working. I would need an oscilloscope and a signal generator with variable frequency capability and sinewave output. If you haven't got these then maybe try simulating the circuit. If you haven't got a simulator then maybe someone has a better idea?

If you disconnect C3 from Q1 and connect a signal generator into C3, at what frequency does the waveform on the base of Q1 become 180º phase shifted to the waveform from the signal generator. Also if the signal generator in inputting 1Vp-p, what is the signal on the base when 180º phase shift takes place?

Is the 180º shifted frequency what you expect it to be?

Is the signal going into the base what you expect it to be?

Is the attenuation produced by the phase shift network less than your gain?

If any of the the above questions is a "NO" then you have to figure out why. That's what I'd do.