In a proposed modification of the basic design of a Michelson interferometer, the beam aperture would be split into a number (N ≧ 2) of smaller areas and the optics arranged so that the beam in each area would generate a distinct segment of the overall desired interferogram. Each segment of the interferogram would correspond to a different segment of the overall optical-path difference (OPD) needed to obtain the desired spectral resolution. As explained below, the advantage of the proposed modification is that the size and mass of a Michelson interferometer could be reduced considerably.

In a traditional Michelson interferometer, it is necessary to move one mirror a distance x/2 to obtain an OPD of x. If the beam is split into N smaller beams, each generating a different part of the interferogram, then the same overall OPD is obtained by moving the mirror a much smaller distance — only x/2N. The moving-mirror arm is a major contributor to the volume, mass, and mechanical complexity of a Michelson interferometer; reducing the overall required mirror travel would make it possible to use smaller, simpler mirror-support structures and mirror-translation mechanisms.

According to tradition, the design of an interferometer must provide for a single beam area, wherein the beam is modulated by a single set of interferometer mirrors. Traditional thinking on this subject has been dominated by the assumption that continuity of the interferogram must be maintained. The proposed modification would violate the traditional requirement of continuity. The violation is justified by the realization that the requirement of continuity is not based on fundamental physics; provided that one can find a way to splice the segments together seamlessly into one interferogram, there is no physical reason to refrain from use of a segmented aperture.

The N segments of an interferogram would be collected in one mirror stroke and would be digitized. The requirement for seamless splicing of the segments translates to a requirement to sample all of the segments on the same computational grid and to match their amplitudes. For this purpose, it would likely be necessary to oversample each segment on its own grid, then digitally resample it to place it on the same grid as that of the adjacent segment. The segments would be adjusted accordingly by comparing small regions of overlap between them and compensating for differences in sampling, signal amplitude, and offset. After adjustment, the segments could be spliced together to form the full interferogram, which could then be Fourier-transformed in the normal way.

Of course, segmentation of the interferometer would entail some engineering compromises. The decrease in overall size, mass, and mechanical complexity would be partly counteracted by a corresponding increase in the number of photodetectors and associated signal-processing circuits, by the need to prevent optical cross-talk between segments, and by an increase in the complexity of computations needed to process the signal-chain outputs. Fortunately, most of these considerations could be addressed by use of modern electronic signal-processing and computing equipment. Recent advances in electronics have yielded great increases in performance and decreases in size and power consumption.

This work was done by Michael Eastwood and Crofton Farmer of Caltech for NASA's Jet Propulsion Laboratory. For further information, access the Technical Support Package (TSP) free on-line at under the Physical Sciences category. NPO-20346

This Brief includes a Technical Support Package (TSP).
Segmented-Aperature Michelson Interferometer

(reference NPO-20346) is currently available for download from the TSP library.

Don't have an account? Sign up here.

Electronics Tech Briefs Magazine

This article first appeared in the April, 1999 issue of Electronics Tech Briefs Magazine.

Read more articles from the archives here.