Electronic – phase-shift oscillator design decisions

oscillatorphase shift

I've been learning about phase-shift oscillators and put together this design:

phase-shift sch

I built it and it seems to work pretty well, you can see the exact values I used, and design here: [1]. But the design I used seems different from a lot of those I've seen online. Most designs put the capacitors in series for example:

phase-shift wikipedia

I found that harder to understand which is why I went for my design. Is there any reason this design is more common? Is it because my design will result is more losses over the in series resistors?

[1] https://41j.com/blog/2014/12/phase-shift-oscillator/

Best Answer

The oscillation criterion (Barkhausen) requires a loop gain of unity: Unity magnitude (at least) and a phase shift of 360deg (resp. 0 deg). Because an inverting amplifier contributes -180deg the remaining 180deg must be created by a third order filter. This can be a 3rd order RC high pass (2nd circuit, +180deg at w=wo) or lowpass (first circuit with -180deg at w=wo).

EDIT: In most cases, the opamp based high-pass circuit is simpler to design because the required gain can be realized rather exact (-Rfb/R1). This is important because the total loop gain should be only slightly larger than unity. Otherwise there will be heavy signal clipping - unless you are using a suitable nonlinearity (e.g. diodes) for soft limiting (amplitude control). Such an exact design is harder to achieve in circuit 1 due to transistor parameters (tolerances). On the other hand, each highpass-based design is more sensitive to high-frequency noise.