Next: FUSE Instrument Technical Appendix Previous: In-Flight Instrument Performance Table of Contents: Table of Contents
The FUV detectors on FUSE were exquisitely sensitive and had a significant dynamic range. However, very bright FUV targets could actually damage the detectors. A bright target observing policy was implemented to protect the detectors whereby an upper limit to the expected target flux of F(λ) = 1×10-10 ergs cm-2 s-1 A-1 anywhere in the FUSE range was imposed. In addition there were hardware and software limits set to protect the detectors. The IDS monitored the total count rate reported for each detector segment, and would reduce the high voltage to SAA level for a segment if the count rate exceeded 45,000 c/s (corresponding to a flux level of approximately 5×10-10 ergs s-1 cm-2 -1 across the FUSE bandpass). The detector DPU software monitored the MCP current and would likewise reduce the HV to SAA level if the current exceeded preset thresholds.
As a policy, the project was conservative in observing bright targets in order to conserve sensitivity. Only a finite amount of charge could be extracted from each detector pixel; thus, the critical parameter for preserving detector lifetime sensitivity is the number of extracted counts, not the count rate (flux). Sensitivity suffers at about 107 e-/pix. Raising the HV could compensate partially, but there were limits to the amount the high voltage can be increased without either incurring detector shutdowns as well as due to limitations of the power supply.
Emission line sources were a particular concern because of their potential to create local regions of reduced sensitivity, or gain sag at critical locations. Consequently, an equivalent flux limit per unit wavelength for narrow emission lines was also applied in planning to prevent local impacts to the detectors.
The scientific user community always expressed significant interest in observing targets that were above the bright limit, for a number of reasons. The FUSE project spent considerable resources, especially during late 2002 2004, on developing and testing various methods of easing the bright limit. Some of the easier adjustments, for targets up to 5 times the nominal brightness limit, were put into operations, allowing new regions of observational parameter space to be opened. Other techniques, such as utilizing lowered high voltage levels or defocus techniques, were developed and tested but never put into operational usage due to technical difficulties and/or lack of resources. The outline below provides some details of what was done and why. At the FUSE Observers Advisory Committee meeting in November 2004, it was decided to terminate efforts on the more aggressive techniques that would have been required to implement observations of targets more than 5 times the nominal brightness limit.
The common thread in all of the attempted techniques was to limit the flux reaching the detectors, such as by observing only in the SiC channels. However, in most of these strategies there was the potential that pointing uncertainty during a target acquisition, or by loss of pointing control when in two-wheel or one-wheel mode, would result in the full flux of the source entering a spectrograph through one of the LWRS apertures in the LiF channels. To safeguard against this possibility, the normal HV management was augmented for bright object observations:
1) The HV was left at SAA level until after the target acquisition was complete
2) IDS rules would be activated to monitor the pointing error reported by the ACS; these rules would lower the HV to SAA level while the pointing error exceeded 5 arcseconds or if the guide stars were lost and a reacquisition was triggered
3) The HV was lowered at the end of every visibility period prior to breaking track on the guide stars.
This enhanced HV management was only possible because the enhanced interface between the IDS and ACS developed for gyroless operations provided the necessary information to the scripts running in the IDS.
No special data reduction is required for any of the implemented bright-object observing techniques.
At the time this mode was first implemented, the throughput of the SiC channels was roughly one-fifth that of the pre-launch LiF throughput used to set the bright flux limit. As a result, the flux limit could be set five times higher for sources observed only in the SiC channels. This was accomplished operationally by observing a source only in the MDRS or HIRS aperture and by moving the LiF FPAs to the high end of their range of motion. This technique could not be employed for LWRS observations, as the FPAs could not be moved far enough to ensure that the source would miss the LiF LWRS apertures. The main drawback to this technique was that no data would be obtained longward of 1100. This became the primary technique for observing objects brighter than F(λ) = 1×10-10 ergs cm-2 s-1 A-1 .
This bright object observing configuration was typically restricted to a single LiF channel - the LiF guide channel, which was LiF1 prior to July 12, 2005 and LiF2 thereafter. The non-guide channel image was typically offset in the y-direction by ~40 arcseconds to ensure that this spectrum was not imaged on the detector. No steps would be taken to avoid the SiC channels, but failure to obtain SiC data would not trigger a re-observation. This mode was planned for only a handful of targets throughout the mission, that happened to be close to the bright limit at LiF channel wavelengths and for which the science programs did not require SiC channel data.
In certain borderline circumstances, the flux from a source would be below the bright object limit for one detector segment and above it in another. One example was Sirius-B, which exceeded the bright limit longward of 1100 . This star was observed both in SiC-only mode, and subsequently in the LiF1A and LiF2B segments. The high voltage was left at SAA level for segments 1B and 2A in the latter observation. This technique was used in a handful of observations during the mission.
Two additional techniques for observations of bright objects were developed and tested. A substantial effort was invested by MP/SciOps staff to safely plan, execute, and assess these tests. The lowered HV technique was not found to be useful and was ultimately discarded. The defocus technique produced usable data of bright objects, but because of the level of effort required on the part of the staff to execute this procedure in a way that maintained the health and safety of the instrument, it was never put into operational use. However, because the test data from these programs resides in the FUSE archive, descriptions are provided here for completeness.
The basic concept of this technique was to defocus one or more telescopes, thus enlarging the size of a stellar image at the focal plane(s). By using the MDRS or HIRS aperture(s) to pick off only a portion of the total light, the flux reaching the detectors would be reduced and targets brighter than the nominal limit could be observed.
It was originally hoped that this technique would enabled observation of targets up to 50 times the brightness limit, but the actual attentuation gained was closer to a factor of 15. The defocus of the primary mirror led to a significant increase in the focal ratio of the beam entering the spectrograph, primarily in the dispersion direction and only slightly in the cross-dispersion direction. This had two impacts on the data: shadows cast by the grid wires that were completely washed out under full illumination became distinct, and the spectral resolution improved. The effects of the grid wire shadows could be removed by FP-splits, which were performed in many of the tests.
Tests were performed that offset the mirrors in both directions and by varying amounts. For safety reasons, only the LiF2 and SiC2 channels were defocussed for these tests, and the high voltage on detector 1 was left at SAA level.
Testing demonstrated this approach to be viable scientifically, but very difficult operationally. The FPA and mirror positions had to be manually determined, scripted, and commanded and the nominal position for the next observation determined. Small inaccuracies in the relatively large mechanism motions also necessitated additional channel co-alignment procedures, once after the mirrors were defocussed, and again after they were returned to their nominal positions. These operational difficulties and human resource issues made usage impractical. Consequently, this technique was not implemented as a standard procedure.
Test data obtained using this technique are archived under Program IDs S520 and S523.
The concept behind this technique was to decrease the high voltage on one or more detector segments temporarily, in order to observe overbright targets (35-100x bright limit). Of course, changing the high voltage changes the behavior and calibration of the detectors significantly. It was thought that perhaps one or two standardized lower setting might be calibrated for use operationally. Tests and analysis indicated only limited utility of this technique. The behavior of the detectors was very sensitive the HV setting and was not repeatable at the required level. More importantly, reducing the HV to the point where most of the events fell below the detection threshold resulted in very poor imaging performance by the detectors. The resulting spectra were unusable over most of the active area of the detectors. At the conclusion of testing, the lowered HV technique was not considered operationally viable.
Test data obtained using this technique are archived under Program ID S525. In most cases, the star observed at reduced voltage was also observed at normal voltage for comparison purposes. Users should not assume that all observations with Program ID S525 are at reduced voltage: the HV keywords in the FITS headers should be examined to determine the configuration for each observation in this program.
The concept of this technique was to place extremely overbright targets near (but outside) a science aperture and integrate long enough to observe the target via the extended wings of the PSF: the scattered light. The resulting flux would fill the spectrograph apertures, so the spectral resolution would be that of an extended source rather than a point source. This technique was tested with observations in the program S521.
This technique, while feasible in principle, was abandoned as a result of several factors: the attenuation factor for the transmitted flux was extremely sensitive to the exact offset of the star from the apertures, this offset varied over an orbit by several arcseconds as a result of the image motion anomaly in all but the guide channel, and the overall pointing itself was potentially variable by many arcseconds as a result of poor controllability with only one or two reaction wheels. The number of stars bright enough for this technique to be applicable was small, and the potential for risk of damage to the detectors was significant, so this technique was not implemented.
Next: FUSE Instrument Technical Appendix Previous: In-Flight Instrument Performance Table of Contents: Table of Contents