Search

The first TH1506B prototype showing problems of overheating and spurious oscillations, a new modified gyrotron was built. During the tests, the extwo peaks, which was never predicted by simulations. Various low evel tests were performed on the mode converter with different shapes for the launcher but without real improvement. Besides measurements, the use of a new software Surf3D[l] showed that the problem mainly comes from the 3rd mirror whose curvature is too high and not well taken nto account by the calculation. This analysis software is based on ntegral equations and the complete 3D modelling alowed to determine a new profile for the 3rd mirror. An aluminium model of a new mirror was manufactured and thorough low level tests made at FZK showed that there was no double peak.The next step would consist in building a gyrotron based on this new design, to confirm the simulation and to valdate it for long pulses.

Sawtooth control is a potentially important method for avoiding MHD instabilities triggered by sawtooth crashes, such as neoclassical tearing modes in a reactor‐size tokamak. The aim of the present study is to investigate the effect of localised current drive on Sawtooth activity. In recent experiments in the Tore Supra tokamak, Electron Cyclotron Current Drive has been observed to strongly affect the Sawtooth period, even at modest ECCD power levels. For instance, with 8 MW of ICRH long, fast ion stabilised, sawteeth were shortened from 0.6 s to 0.3 s by application of 0.6 MW of co‐ECCD. In these experiments it is estimated that only 2–3% of the total current was driven by ECCD, but this was sufficient to significantly affect the sawtooth period provided the driven current was localized accurately with respect to the q = 1 surface