Hi Adam,
The approximation for Vout is:
Vout=3*Vd*R1/(2*R1-R2)
but there are restrictions. I think Jony was the first to mention this equation.
R3 is not in the equation because it is not a significant gain factor. It's just there to make the diode curves a little softer so we get less distortion due to any clipping action. It is in parallel to the diodes so the diodes do most of the control and R3 just helps the diodes do it in a less abrupt way. If on the other hand we had cared to include the diode equation itself in the formula for Vout, then we would certainly have to include R3 in the analysis because we would need it to show exactly how the diode curve changes because of it's value. In the approximation however the diode (either one) is viewed as a voltage source, not a resistance, so we dont have to consider R3's value because it has no effect on the voltage source. Of course R3 can not be too low though, or it will swamp the diode action and we will no longer have any gain control.
The equation is developed from knowing the transfer function for the positive feedback loop and the negative feedback loop. When the circuit oscillates it oscillates at w=1/RC and that means at the non inverting terminal we get a voltage that is Vout/3 and it is in phase with Vout as well. For a perfect circuit, the negative feedback loop would produce Vout/3 at the inverting terminal and this is the basic idea behind the equations, except for the fact that we have diodes (a voltage source) in the negative feedback loop as well. The diode voltage subtracts from the output and that in combination with a tiny offset in the op amp itself produces a non zero output voltage that rises with time until the diodes start to conduct.
The restriction is that R2 can not be too large or there wont be enough feedback and the output will keep rising until the op amp output saturates, and that will cause intense distortion. So R2 has to be kept high enough to allow oscillation to start but still low enough to prevent the output from going too high. This brings us to the schematic which does not look right because R2 looks too large. It may have to be decreased to maybe 15k or 17k or something like that, but you can use the equation to investigate this further.
BTW the diode approximate voltage drop for this kind of circuit may be closer to 0.5v than to the usual 0.7v sometimes associated with a regular si diode.
Please let us know how you make out with it.