I made a mistake in deriving the formula for RF power: this should be beta * R_ETM * P_eff * sin (theta_asym) * sin (fringe) * cos (2pi*f1*t).
Correct estimation of the modulation index is beta = 0.12