[Radio Wrangler]

It isn't a wein bridge oscillator. The title is correct, it really is called a 'phase shift oscillator' which has always been a poor name because just about all oscillators have their frequency set by the phase shift of some network.

The valve gives 180 degrees of phase shift as it is an inverting amplifier.

A single series-R-shunt-C section will give 45 degrees of delay (lag) at its 3dB attenuation frequency, and will approach 90 degrees as the frequency gets very large. Unfortunately the attenuation also becomes huge.

The circuit needs to make 180 degrees of lag in the RC network. One RC stage can't do it. Two stages will just get there but at infinite frequency, with infinite loss, so that's not going to work. Three (or more) stages will get to 180 degrees total while there is still some signal getting through.

As printed with a single variable resistor, this circuit will only tune over a small range, so if used to drive a tape recorder capstan motor it'll give enough pitch shift to fine-tune things but won't turn the machine into a wide speed range job.

With any oscillator, the design job isn't finished with making a frequency control circuit. The problem arises of stabilising the amplitude too. Too little gain around the loop and oscillation stops or never starts in the first place. Too much gain and the amplitude grows until something bangs into its end-stops. Unfortunately, the gain required is a single exact value so you can't in theory make a circut work without some active self-adjustment of gain.

The wien bridge has feedback controlling the gain of its amplifier, so it can't use amplifier non-linearity as gentle limiting. The feedback network has to change with level, and so it uses a thermistor or else a detector and an FET as a variable resistor.

The phase shift oscillator doesn't fix the non-linearity of the valve, it uses it. The gain is set a bit on the high side, and as level builds the overall gain, averaged over a full cycle, falls and the level finds a stable equilibrium.

The grid 2 voltage of a pentode will reduce the gain if the voltage falls, and a hard driven stage will start to take more grid 2 current. That resistor hasn't got a value written on it for a reason. I suspect there is some text in the accompanying article about fiddling with its value to get the oscillator level to a chosen value.

So pick a value to start with, and be prepared to change it until it runs with a reasonable voltage and the waveform looks reasonably sinusoidal on a scope.

You can make a wide range phase shift oscillator with a triple ganged pot, and switched capacitors, but getting the gain right for a good level becomes difficult over much of a frequency range.

David