Why are freeform telescopes less alignment sensitive than a traditional unobscured TMA?

Abstract

As freeform optical systems emerge as interesting and innovative solutions for imaging in 3D packages there is an assumption they are going to be more sensitive particularly at assembly. While it is true that the clocking of the component becomes a relatively weak new tolerance, for the most effective new class of freeform systems the alignment sensitivity is actually lower in most cases than for a comparable traditional unobscured three mirror anastigmatic (TMA) telescope. Traditional unobscured TMA telescopes, whose designs emerged in the mid-70s and which begin to appear as hardware in the literature in the early 90s, are based on using increasingly offset apertures with otherwise coaxial rotationally symmetric mirrors. The mirrors (typically 3 to correct spherical, coma, and astigmatism) have evolved to contain more high order terms as the designs are pushed to more compact and wider field packages – the NIRCAM camera for the JWST is an excellent example of this [1]. As the higher order terms are added, the mirrors become increasingly sensitive to decenters and tilts. An emerging class of freeform telescopes that provide wider field of view and/or faster f/numbers than the traditional TMA are based on a strategy where the surface shape remains a low order Zernike-type surface even in compact, unobscured packages. This optical design strategy results in an optical form that is not only higher performance but simultaneously less sensitive to alignment.