Modelling high resolution Echelle spectrographs for calibrations: Hanle Echelle spectrograph, a case study

TL;DR

This paper introduces a simplified, accurate modelling scheme for high resolution Echelle spectrographs that predicts spectral line positions with high precision, enabling better control and calibration without complex ray tracing software.

Contribution

The authors develop a novel, easy-to-implement modelling approach for spectrographs that achieves high accuracy and can be used for real-time instrument control and calibration.

Findings

Model predictions match commercial ray tracing software results.

Calibration with Thorium Argon lamps achieves 0.08 pixel accuracy.

Monte Carlo simulations show robustness against photon noise.

Abstract

We present a modelling scheme that predicts the centroids of spectral line features for a high resolution Echelle spectrograph to a high accuracy. Towards this, a computing scheme is used, whereby any astronomical spectrograph can be modelled and controlled without recourse to a ray tracing program. The computations are based on paraxial ray trace and exact corrections added for certain surface types and Buchdahl aberration coefficients for complex modules. The resultant chain of paraxial ray traces and corrections for all relevant components is used to calculate the location of any spectral line on the detector under all normal operating conditions with a high degree of certainty. This will allow a semi-autonomous control using simple in-house, programming modules. The scheme is simple enough to be implemented even in a spreadsheet or in any scripting language. Such a model along with an optimization routine can represent the real time behaviour of the instrument. We present here a case study for Hanle Echelle Spectrograph. We show that our results match well with a popular commercial ray tracing software. The model is further optimized using Thorium Argon calibration lamp exposures taken during the preliminary alignment of the instrument. The model predictions matched the calibration frames at a level of 0.08 pixel. Monte Carlo simulations were performed to show the photon noise effect on the model predictions.