Thanks, guys,
by observer, i mean to reconstruct the flux information from the measured stator currents and the knowledge of machine parameters. it's "calculated/estimated" flux, which is used in Indirect Field-oriented Control of IM motor.
Talking about coil, I do have one wound around a tooth on the stator, while the v(t) = diff(flux)/dt at realtime, the fomular v=v = 4.44*pi*f*N*phi may be the steady state relationship.
Sounds the coil idea might be the only choice, Thanks a lot.