A new method is described for determining error field strengths using the dynamics of a rotating m = 2, n = 1 resistive MHD tearing mode as observed on TEXT-U. A set of external saddle coils induces a magnetic field whose m = 2, n = 1 component is either in phase or out of phase spatially with the intrinsic m = 2, n = 1 error field. This experimentally applied field, combined with the pre-existing error field, decelerates the plasma rotation by coupling to the tearing mode and induces a locked mode. The time evolution of the average rotation frequency and the Fourier amplitudes of a Mirnov signal observing the tearing mode are predicted analytically. This analytic result compares favourably with experimental data, which are analysed for the frequency and Fourier amplitudes using variable frequency complex demodulation. The error field is measured at the q = 2 surface to be 0.45 ±0.09 G by noting when in the evolution the applied field cancels the error field. This technique is unique because it measures the field error in the equilibrium of interest and is extremely sensitive (the measured error is 2 × 10-5 B0); because it is a dynamic measurement, it could conceivably be deployed in real time to prevent locked mode disruptions.
Nonlinear error-field penetration in low density ohmically heated tokamak plasmas