Robust feedback control with large parameter uncertainties in a high precision stage with large sample mass variation is presented. Instead of expressing the nominal plant and its uncertainty by parametric models, they are expressed using the identified frequency response data. The open loop transfer function is mapped onto a Nyquist plot, to assess robust stability. Next, the complementary sensitivity function is derived from this data to describe tracking performance, which is verified by experiments. The advantage is that the parameter estimation step is omitted, without compromising robust stability and performance.@en

Correct tuning of feedback gains is important for AFMs to cope with uncertainties of the system dynamics coming from a large range of different samples, cantilevers and scan parameters. For state of the art AFMs gains have to be adjusted manually by the operator. The typical operator such as biologist, physicist or material scientist may not have detailed knowledge about control engineering. In order to increase usability and acceptance an easy and intuitive approach is needed. The method presented in this paper is based on the commonly applied manual tuning strategy for AFM-imaging and copies the behavior of an experienced user. By spectral analysis ringing of the feedback loop is detected when feedback gains are increased beyond the stability margins. Increasing gains is stopped when an 1/f stop criteria is reached. The algorithm is successfully tested in simulation and practical AFM topography measurements with different cantilever - sample combinations, demonstrating that auto-tuning can be applied to achieve imaging performance close to ideal settings.@en