GALEX OBSERVATIONAL CONSTRAINTS and SAFETY
The high-sensitivity GALEX detectors dictate strict brightness limits for target fields. The GALEX satellite operates in a largely autonomous way, with observing plans typically uploaded weekly. These properties constrain the observations GALEX is able to do. The major constraints are summarized below.
Pointing centers must be separated from bright stars by:
Please keep these limits in mind when you design your observation program. Also, see helpful hints at: http://galexgi.gsfc.nasa.gov/tools/chkbstar_whatif.html
Non-linear responses when observing bright fields:
There are two sources of photometric non-linearity in the GALEX instrument: global dead time resulting from the finite time required for the electronics to assemble photon lists, and local sensitivity reduction resulting from the MCP-limited current supply to small regions around a bright sources (gain sag). (The microchannel-plate, or MCP, electron-multiplier array is the central component of the GALEX detectors.)
Global dead time refers to the fraction of detected events lost due to the finite processing speed of the electronics. It increases monotonically with global input count rate. It is easily measured using an on-board pulser, which electronically stimulates each detector anode with a steady, low rate stream of electronic pulses that are imaged off the field of view. Since the real rate of these pulses is accurately known, the measured rate is used by the pipeline to scale the effective exposure and thus correct the global dead time for all sources in the field simultaneously. This correction is typically about 10% in NUV and negligible in FUV, however it can become quite significant (~30%) for the brightest fields.
The actual fraction of observing time lost is 1-(1/(1+Td*R)), where R is the corrected count rate, which is what is seen in GALEX data products, and Td is a constant, 5.52e-6 sec, which is the (non-paralyzable) deadtime for each photon event. The deadtime correction returns the proper fluxes, but, of course, does not alter the reduced SNR that results from the loss of observing time. To mitigate the effect of the deadtime on SNR, one has to increase the exposure time by a factor of 1+Td*R.
Local dead time, or "gain sag" results from the limited ability of MCPs to provide current to a locally-intense region of illumination. It is difficult to correct with high accuracy, partly because it requires many observations to calibrate sufficiently and partly because it is a function of MCP gain (which varies over the detector). Local dead time affects both the measured count rate of individual bright sources and the shapes of those sources. We have used standard stars to estimate the local dead time in each band, as shown in the count-rate linearity figure. (These observations are already corrected for global dead time effects, so the dead time shown in the figure is zero up to the point where the MCPs begin to suffer from local dead time effects.) The gain sag causes some photon events to fall below electronics pulse-height thresholds. These gain sag effects recover immediately after the bright source is removed, except in the case of very long, bright-source exposures. [The GALEX mission planning team intends to avoid this sort of observation, to avoid permanent damage to the detector photocathodes.] The NUV detector is suffers less from gain sag than the FUV, because it is proximity-focused and thus presents a larger image (with lower count density) to the MCP.