3.1.

Calibration Cap Features and Optical Simulations

The calibration cap assembly (Fig. 2) is mounted on top of the spectrograph tank, which is located under the siderostat mirror as seen in Fig. 8. Three main components make up the body of the calibration assembly—the adapter plate, housing, and cover. The adapter plate mates with the field lens holder on top of the spectrograph tank. The housing is secured to the adapter plate through screws on its bottom flange. The middle of the housing body contains fiber adapter plates for the eight optical fiber ends to mount to its surface. The top of the housing holds the 90-mm diameter optical shutter with a raised edge to prevent stray light from entering the spectrograph. The shutter is placed onto positioning pins in the housing and locked into place by the cover, which has a central cutout to allow light into the spectrograph tank. The calibration cap design maximizes the amount of space between the calibration cap and siderostat to prevent mechanical interference during the most extreme tip and tilt configurations of the spectrograph tank during focusing, field alignment, and observing.

The calibration cap is designed to provide partially diffuse illumination onto the focal plane from the calibration lamps when its shutter is closed. When the shutter is open and calibration lamps are off, the incoming light from the primary parabolic mirror enters the spectrograph through the calibration cap cover aperture. The cone angle of the aperture accommodates the incoming focused beam from the primary mirror after passing through the siderostat mirror cutout. The shutter is designed to fail open such that science observations can still be made in the event of a shutter failure. While the calibration cap shutter is closed, the light produced by the $D_2$ or Zn lamp within the calibration box is transferred through optical fibers, illuminating the shutter blades. The custom fiber bundle has two common ends that split into a fan out of eight fibers. The eight fiber ends are mounted perpendicularly on the conical surface of the calibration cap housing, directing light toward the shutter blades at a distance of $\sim 60\mathrm{mm}$ away. A combination of the distance between fiber ends from the shutter and the angle (21.2 deg) between the conical surface relative to the shutter surface normal was optimized to provide light beams that overlap on the shutter to create a reflected field of light onto the focal plane.

Two separate simulations were done as follows: (1) to determine the proper shutter aperture size to prevent vignetting of incoming light from the primary mirror above (Fig. 5) and (2) to approximate the distribution profile of light reflected off the shutter blades and into the spectrograph tank (Fig. 6). BSDF (bidirectional scattering distribution function) data was collected by ScatterWorks from black Teflon shutter sample blades. The mechanical model and BSDF data were used to construct a light-intensity distribution plot at the spectrograph tank entrance. Black Teflon-coated shutter blades were used in the first iteration of the calibration cap design, but were changed to ${\mathrm{AlMgF}}_2$ (ZM)-coated shutter blades to increase the the intensity of light reflected into the spectrograph. The intensity plot profile slope for ZM blades is expected to differ since ZM creates more specular than diffuse reflections. It is also expected that reflections from ZM would create a much greater overall intensity throughout the spectrograph entrance area compared to black Teflon. Although the resulting ZM shutter blade irradiance profile is not able to be confirmed without the ZM-coated blade BSDF data, we were able to confirm that it provided sufficient UV light for calibration processes while testing the integrated payload system. Throughput testing showed sufficient light intensity at the spectrograph detector reflected from the ZM-coated shutter blades for calibration activities as seen in Fig. 10.

Fig. 5

Zemax optical simulation footprint diagram plot showing that the minimum calibration cap aperture diameter of 90 mm (black circle) is required to prevent vignetting of edge rays entering the spectrograph tank. Light rays in this plot are contributions from all optical components of the instrument, where field positions are with respect to the center of the 90 mm aperture with unit circle coordinates. The 5.93% loss of rays through the aperture is at the primary mirror, through the siderostat cutout.

Fig. 6

Simulation showing the intensity distribution of reflected light from the shutter blades. The calibration cap assembly mechanical model and black Teflon shutter blade coating BSDF data were used to create a plot that approximates the light intensity profile entering the spectrograph tank field lens holder aperture area after reflecting off the shutter blades above. Left: Calibration cap assembly 3D CAD model. Labeled are the components crucial for the optical simulation. Right: Simulated LightTools irradiance plot showing the distribution of light entering the spectrograph tank. The light intensity scale is normalized by the peak value.

3.2.

Calibration Box Features and Optical Simulations

The calibration box contains the deuterium and zinc lamps. It also contains all the optics that direct light into the fibers and all the electronics needed to power and control calibration lamps, lamp shutters, cooling fans, and pressure and temperature sensors. Light from each lamp is transferred from the top of the box into the calibration cap with custom 2-to-8 fan-out optical fibers made by Hellma. The low polarization fibers are designed for 190 to 1200 nm, each of the eight fibers has a $300\mu \mathrm{m}$ diameter core and each of the two common ends outputs a spot diameter size of 1.1 mm. The calibration box is hermetically sealed pre-flight, maintaining near-atmospheric pressure during flight. This ensures a stable environment for electronics operation for the mission duration. Continuous temperature monitoring and controls within the system allows the operator to turn off the calibration system if temperatures exceed $\sim {50}^{°}\mathrm{C}$.

All mounting structures, lamps, optics, and electronics are modeled in SolidWorks and fit within the cylindrical enclosure (9.5 in diameter × 20 in tall). This mechanical model confirms that each lamp is aligned with the fold mirror, the two biconvex lenses that focus light through each shutter, the fused silica window, and finally into the fiber collimating lenses mounted on plates clamped onto rods with 2 mm graduation marks. Lamps, fold mirrors, biconvex lenses, and fiber collimating lenses are movable along the optical axis to optimize light throughput into the fibers. All components are mounted on a main central plate that attaches to the calibration box flange cover with a reinforced bracket except for the lamp shutters and biconvex lenses, which are mounted in the shutter mounting box shown in Figs. 3 and 4.

A Zemax optical simulation is shown in Fig. 7 with calculated nominal positions for deuterium and zinc lamps relative to their fold mirrors, lenses, and fiber collimating lenses. The positions of the optics differ for each lamp in the calibration box assembly due to lamp mounting and size differences. In addition, the optical components have about $\pm 5\mathrm{mm}$ margin of movement along the optical axis since both calibration lamps produce a wide beam of diffuse light.

Fig. 7

Zemax simulation showing an optimization of the two lamp channel (deuterium and zinc) optical components.

3.3.

Assembly and Payload Integration of the Calibration System

The fully integrated new calibration system shown in Fig. 8 was used to take initial calibration images (Fig. 10). Before payload integration, all custom-designed and machined components, electronics, and optics were mounted on the mounting plate for an initial mechanical fit check as seen in Fig. 9. Then, each subsystem was electronically tested (see Appendix A for electronic subsystem diagrams in Figs. 13Fig. 14Fig. 15Fig. 16–17). The calibration box and cap were first tested through an Arduino Uno Rev3 microcontroller board programmed to control the lamp relays, shutters, and temperature and pressure sensor readouts. This simulated controls from the guider onboard computer (GOBC) on the gondola, and an external 28 VDC power supply was used to simulate the gondola power supply.

Fig. 8

Left: Labeled photograph of the calibration box on the gondola and calibration cap mounted on the spectrograph tank. Right: Calibration cap with cover off with the shutter placed in the housing.

Fig. 9

Top left and right: Labeled photograph of the calibration box with mounted electromechanical and optical components placed on the test stand. Bottom: Calibration box flange cover showing electrical and optical outputs.

Once all subsystems were functioning through Arduino controls, optical components were placed in the calibration box according to positions calculated in the Fig. 7 Zemax model on the test stand as discussed in the previous subsection (Sec. 3.2). The calibration box flange and mounting plate assembly were placed back in the enclosure, and a black non-reflective shroud was placed around the calibration box window and fiber optics to block out ambient light during spectral measurements. The spectral throughput of the zinc and deuterium lamps was tested with an Ocean Insight HR4000 spectrometer that is sensitive from 200 nm to visible wavelengths. Spectral throughput was first measured at the top of the calibration box at the fiber optic collimating lens with a single fiber to confirm the presence of the deuterium continuum around 200 nm and the 214 nm zinc emission line. Then, throughput from each of the eight fiber legs in the 2-to-8 custom fiber bundles was tested. Both lamps showed emission from the deuterium continuum and zinc emission lines around FIREBall wavelengths of 200 to 208 nm. The minimum flux requirements of the zinc 214 nm emission line and deuterium at 206 nm are $\sim {10}^{-12}$ and $\sim {10}^{-14}\mathrm{erg}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}{\mathrm{arcsec}}^{-2}{\AA }^{-1}$, respectively, at the spectrograph detector to achieve detector gain verification and slit positioning calibrations described in Sec. 4. Both lamps achieve the necessary criteria to continue installation of the calibration system into the gondola. Our calibrations require adequate illumination for 60-s in-flight exposures, essential for point source function, spectral, and spatial calibrations. It is crucial to emphasize that while the calibration lamps provide the emission for calibration purposes, they are not intended to be used as flux calibration sources, flux calibration will instead rely on sky sources. The calibration lamps’ primary function is to provide reference signals for instrument calibration and validation processes in Sec. 4.

After testing the calibration box throughput and electronics independently from the gondola, it was considered ready for integration with the FIREBall guider onboard computer (GOBC). The calibration system shutters, lamps, and sensor readouts are all commanded through the GOBC during flight. The GOBC contains a PCI-1706U I/O card, allowing for in-flight communication with the calibration system by sending and receiving data. Each calibration subsystem (see Appendix A) that requires commands such as opening and closing shutters, turning lamps on and off, and temperature and pressure readouts are wired to digital or analog pins on the I/O card and then programmed to command each subsystem at the assigned pins. All functions are controlled by the GOBC over the flight duration through direct commands by the operator on the ground and are not part of an observing sequence that is autonomous.

With the calibration cap mounted on the spectrograph tank and the cap shutter closed (Fig. 8), the first images were taken of the zinc and deuterium lamp light reflected off the shutter blades and through the spectrograph tank. Initial images showed clear deuterium continuum and zinc emission lines that can be used for in-flight calibration processes discussed in the following sections.

4.1.

Pre-flight Calibration and Initial Calibration Results

Following all scientific observations in a nominal flight, the final flight phase, calibration mode, would consist of taking dark frames at the same exposure times as the science observations and calibration lamp exposures (either Zn or $D_2$) at detector flight gain levels while the gondola doors and calibration cap shutter are closed. Calibration lamp images and dark frames are taken with the same science masks as those used during science observations to ensure calibration flux through the same spectrograph configuration. The data can then be used to provide spectrograph wavelength calibrations and measure spectrograph alignment and slit-mask positioning relative to the UV detector.

Although flight calibration data were not collected due to early flight termination, initial calibration lamp images, shown in Fig. 10, were taken during ground instrument calibrations. Both gain characterization and verification were performed on the ground before flight. During characterization, it was crucial to block out all ambient light external to the calibration sources (deuterium and zinc lamps) that could faintly illuminate the instrument’s focal plane. In electron-multiplying (EM) mode, the deuterium UV continuum was used to statistically quantify the signal above the readout noise threshold. Subsequently, the zinc lamp helps identify specific emission lines for gain verification. Utilizing the deuterium continuum data, only the pixels above the readout noise threshold were selected to recreate a zinc lamp emission pixel distribution across the detector. Upon successful completion of the EMCCD amplification gain verification, the recreated zinc spectral image matches an actual zinc emission line image. EMCCD gain verification would once again occur in flight after science observations following the same procedure. Figure 11 shows an example of a bias- and cosmic-ray-subtracted calibration image with an EM gain of $\approx 700\hspace{0.17em}\mathrm{e}\text{-}/\mathrm{e}\text{-}$ and a read noise of $\approx 35$ DN (digital number). The image has a mean background level of $\approx 134$ DN for a 60-s exposure (the same exposure time used for flight), and the spectral lines from the calibration lamp have intensities of $\approx 1000$ DN, reduced in the displayed averaged spectrum due to the extended $y$-extent of the region of interest used. The emission line width is calculated to determine the spectral resolution at the given wavelength, and the measured centroid position informs the instrument alignment. Ground calibration data showed a spectral resolution $⪆1400$, and the slit centroid positions were measured to within $\pm 10\mu \mathrm{m}$, consistent with the requirements outlined in Table 1.

Fig. 10

MOS images from FIREBall-2 pre-flight calibration. Left: Continuum (deuterium) spectra for one of the nominal science masks (top) and for the long-slit mask (designed for calibration targets or quasi-stellar objects (QSOs)) (bottom). Right: The same masks are illuminated using emission lines (zinc) lamps.

Fig. 11

Example of a bias- and cosmic-ray-subtracted calibration zinc lamp detector image showing emission lines used for EMCCD calibration. Above: Calibration zinc lamp image. Black horizontal lines are cosmic rays that are subtracted out. Below: Zinc lamp spectrum plotted with data from the detector area above within the green box.