Impact of WFB on response matrix

When calculating the response matrices we look at the response to delta function perturbations in each field at each theta grid point. Typically the response damps quite strongly away from the point of perturbation (depending on bakdif) such that significant response is only found within a small theta range. However, the default and passing wfb treatments effectively treat the wfb as periodic over the extended domain. If we consider perturbing a field near one end of the extended domain, the response will be near zero at the other end of the domain except for the wfb which will have order 1 response. Essentially, the treatment of the wfb as periodic on the extended grid results in both ends of the parallel domain being strongly coupled together in the response and field calculations. This occurs no matter what nperiod or nx is used, which appears unphysical. Choosing a trapped wfb treatment avoids this issue.

Comments (2)