Effect of alignment and tolerances on reverse raytrace calibration

There are several sources of error in interferometry to consider when testing surfaces in a non-null configuration. A model of the interferometer is typically used to calibrate these errors, but the model differs from the actual interferometer due to the alignment and tolerance of individual components. Reverse raytrace calibration using a model that differs from the real system corrects some errors but introduces others. Reverse optimization using measurements from known test configurations or configuration changes can produce a model that better reflects the real system. This paper addresses the tolerances required to obtain calibration precision from reverse ray tracing. The sources of error can be separated in a way that allows the amount of correction to be compared to the generated errors from misalignment. These errors can be expressed in a generic way that can be applied to any arbitrary interferometer architecture or test surface shape. The simulation results of a standard interferometer with standard tolerances shows that errors corrected by reverse ray tracing can be on the same order as the errors generated by reverse ray tracing an incorrect model. The efficacy of the calibration method resides in correction of other errors such as distortion and ray intercept coordinate error. These corrections aremuch larger than misalignment errors for surfaces with large departures. This method can be used to determine the level of interferometer component alignment required to accurately measure large departure surfaces with reverse ray tracing.