Welcome to the E2E GitHub repo: a set of performance analysis simulations for the HARMONI End-to-End Model, an optical model of the complete system including all subsystems: Focal Plane Relay System, Preoptics, Integral Field Unit, Spectrograph, Detectors 🔭. This repo contains Python code that uses the Zemax Optics Studio (ZOS) API to analyse the HARMONI E2E files.
What are the End-to-End Models? The E2E models are optical models in Zemax Optics Studio that represent the complete HARMONI instrument. They are created by joining together optical models of each subsystem provided by different teams within the HARMONI consortium all across Europe 🇬🇧 🇪🇸 🇫🇷 🇩🇪. One of the main roles of the E2E models is to demonstrate that the performance of HARMONI will meet the requirements. Given the complexity of the HARMONI instrument and the myriad of possible configurations, this task cannot be done manually with a handful of Zemax analyses. This is where this repo comes in.
What does the E2E repo do? At the University of Oxford we are in charge of putting together the E2E models and analysing their performance; and this is where I come in. The E2E repo contains a bunch of Python code to automate this task, including custom performance analyses designed to match the peculiarities of HARMONI, as well as post-processing routines to plot and summarise the instrument's performance.
Why is this needed? An E2E model file represents a single instance of HARMONI for a specific choice of system parameters. But HARMONI is a workhorse instrument designed to work over a wide parameters space and has:
- Spaxel Scales: 4 independent choices of 4x4, 10x10, 20x20, 60x30 [milliarcseconds].
- Spectral Bands: 11 different gratings covering multiple bands VIS, IZJ, IZ, Z_HIGH, J, H, K, HK, H_HIGH, K_LONG, K_SHORT.
- Integral Field Unit Channel: 4 pairs of independent IFU channels AB, CD, EF, GH, each leading to a different spectrograph + detector.
- Adaptive Optics modes: 4 choices of Adapative Optics NOAO, SCAO, LTAO, HCAO (High Contrast).
This means that just for the nominal design, we can create 704 separate E2E files. Consequently, we need a toolbox capable of running a set of performance analyses across many files. Moreover, if we look at Monte Carlo analyses, where each subsystem provides not just its nominal file, but around a hundred random instances, the possibilities are almost endless.
We have implemented a range of custom performance analyses that combine Zemax operands and Python algorithms to calculate the metrics we need to demonstrate that HARMONI fulfils the requirements:
- Raytracing (general): to validate the field definitions along the optical path across all spaxel scales
- RMS Wavefront Error: can be calculated at any arbitrary surface with custom sampling
- FWHM PSF: geometric PSF based on raytracing + diffraction effects + detector effects
- Ensquared Energy: geometric value in spatial direction
- Spot Diagrams: see example below at the Detector Plane
As an example, we can calculate the RMS Wavefront Error map at the Detector Planes for all 4 IFU channels for a given set of spaxel scale, spectral band, system configuration, etc. The y axis represents the spectral direction with varying wavelengths, while the x axis is populated by the different slits. These maps allow us to diagnose whether something is wrong with the E2E models, such as vignetting at intermediate surfaces.
To summarise the performance across all spectral bands, we use violin plots like this. They show the distribution of RMS WFE values (from the Detector Maps) along with useful information like the median, the first and third quartiles, etc. This allows us to quickly assess how much margin is left by comparing the distribution to the nominal requirement (red line).
One of the more interesting things about the HARMONI instrument is not just showing how the nominal design performs, but rather how the as-built instrument can be expected to perform. For that, we need Monte Carlo instances of each subsystem, representing realistic examples with manufacturing errors, misalignments, surface irregularity and other effects. For that purpose, we are currently adapting the E2E package to deal with these Monte Carlo End-to-End models and explore how the performance changes when we include these effects, and understand the range of performance variation that we can expect when HARMONI is finally assembled.