Why doesn't this phase shift oscillator work?

I have to achieve 100 MHz oscillation frequency. To startup the oscillation, V2 is modeled as a step voltage source with 10V final value. But I am not getting any oscillation when I try to simulate this on a simulation software. What is the problem in this design. Please, I really need help with this.

• For 100 MHz try 10pf and compute z(f) Aug 24, 2020 at 8:01
• @TonyStewartSunnyskyguyEE75, did you mean I should replace 1pF capacitor with 10pF? I cannot do that, because i have to design for 1pF and 100MHz. These are the only requirements. Aug 24, 2020 at 8:04
• No C5 time constant is almost 1 second, when things dont work , checking the voltages for each node will tell you whats wrong. Aug 24, 2020 at 8:23
• Also the source impedance to load attenuation must be lower than gain for unity gain at 180deg Aug 24, 2020 at 12:17

Break the loop, insert a signal (Vin) as shown below and plot the AC response of the circuit from 10 MHz to 1 GHz and see what happens: -

I suspect that there will not be enough gain when the phase response is 0 degrees OR, try running the simulation for a longer time-period because sometimes, oscillators need some time to begin noticeably oscillating i.e. they do not start instantly. If you have the option, run the elongated transient response not from the estimated operating point but right from t = 0 i.e. there should be some provision in your tool to stop the tool calculating the DC bias points and "unhelpfully" beginning the transient point when those transients have settled down. On MC12 (my tool) it is shown here: -

I never run oscillators with that option ticked even though my tool annoyingly defaults to it. Maybe it's the same in your sim.

• Then wait a second for the bias to normalise...;) Aug 24, 2020 at 12:16
• It would be faster to fix it in Falstad, but my dog wants to take me for a walk/run Aug 24, 2020 at 12:21

I've never had much success with that high pass configuration of the RC phase shift network in a phase shift oscillator. You could try the low pass configuration as below in which the drain resistor (amplifier's output resistance) forms the first R in the phase shift network.

Note how the amplifier saturates. This reduces the loop gain to 1 once the oscillator is running.

The simulator I've used won't simulate up to the sort of frequencies you're after but you could try with your simulator and reduce the C values in the phase shift network.