Design
DEUCE is a dual-channel, extreme ultraviolet spectrograph operating in the 700-1150A bandpass. The scientific goal of the payload is to observe hot stars like Epsilon and Beta CMa in a flux-calibrated fashion for the first time across the 700-900A bandpass.
The DEUCE design consists of a grazing incidence Wolter II telescope, a normal-incidence, holographically ruled diffraction grating, and a large 200x200mm microchannel plate detector, oriented in a Rowland circle (the grating, detector, and focus are all mounted on the same theoretical circle, which matches the focal length of the grating). DEUCE fits inside 22" sounding rocket skins, and is approximately 10 feet long, weighing ~500 lbs.
The DEUCE design, as viewed in Solidworks. The path of a light ray is shown in red. Payload diameter is 22".
DEUCE has two channels, created by its two slits: a 'high' resolution (R=1850) thin slit, and a low resolution (R=180) large slit. These slits are necessitated by the fact that the natural stellar spectrum of a B star brightens dramatically above the Lyman Limit at 912A. If we had a single slit, we would either not be able to observe the fainter <900A but comfortably register the >912 flux, or have the right count rate on the <912 flux but have the >912 flux blow the detector out of the water. This dynamic range issue has proven problematic to past missions trying to make this measurement, and is why the dual-channel design of DEUCE is crucial.
The slits as seen in a laboratory spectrum
The slitjaw for the flight, showing high and low resolution slits
A model spectrum and data of Epsilon CMa, showing both the 730-1150A flux gap and the dramatic brightening of the stellar flux as you rise above the 912A Lyman Limit, which necessitates the dual-channel design of DEUCE.
A laboratory spectrum, showing the FUV and EUV spectra registered by the detector when the slits are illuminated by UV light. The FUV, large slit portion is blocked off by a physical plate in the detector to prevent the extremely bright stellar FUV flux from overwhelming the detector electronics and washing out the EUV in flight.
Payload Design
Payload design is a mix of mechanical and optical engineering, with a mind for the science requirements, the budget, physical reality, and what will make all three play nice. For instance, we already had a previously flown grazing-incidence telescope, which would be one of the most expensive components to build or obtain. So the optical design was based around the telescope already in our possession. The experiment design took into account estimates of the in-flight count rate coming from Epsilon CMa in both the high and low resolution channels, using existing stellar models as a basis. DEUCE was designed to be at least twice as good as the "bare minimum" it needed to get the science done; this is a good practice that allows for wiggle room and the inevitable non-idealities of real life.
I came in once the majority of the rocket had already been designed, including all of the optical design. However, by the time of launch I owned the design, and made whatever changes I felt were needed to improve the payload or make it work as required. For instance, I realized that the base design was planned in such a way that the main cavity wouldn't actually pump down-- it had lightweighting holes that circumvented the main O-rings on the spectrograph canister. I redesigned the bulkhead accordingly, calculated the O-ring groove depths and tolerances, had it machined, and have launched that working design twice. During build up and testing, I was on solidworks daily, checking parts, making changes, and verifying anything we needed for that particular test or construction. Solidworks is like LEGO for adults, and it was a lot of fun picking it up and and making changes to the payload design based on our science criteria or as our conception of what we wanted progressed.