I don't use EWB or whatever that is, but in general, you need to open the loop to do a Bode plot on an oscillator. I would break the loop between the U1 output and R5. Put your AC source between R5 and GND, and measure the output of U1, also referenced to GND. You should have slightly more than unity gain and zero phase shift at the calculated oscillation frequency if the circuit is going to oscillate.
When you do an AC simulation, which is how you get a Bode plot, all nonlinear components (diodes, in this case) are linearized to have constant parameters (R and C), independent of the instantaneous voltage across them. My point is that the diodes here will not affect the gain, which is OK, because they are just there to provide soft limiting so that the output oscillations don't saturate the amplifier.