At the start of the scan, the scan mirror is positioned near its ''center'' position, where all three detector arrays view the sky; data are collected from all three arrays. The scan mirror is moved in a one sided ramp with ''flyback'' around this position; during the ramp motion, the scan motion of the telescope is compensated by the motion of the scan mirror to freeze the field on the arrays. This operating mode is also called ''freeze frame scanning.'' The flyback occurs at intervals of 1 and 3 pixels of the 160 µm array alternately (i.e., ~16´´ and 48´´) at the slow and medium scan rates, and every 4 pixels worth of slew at the fast rate (i.e., ~64´´). When the telescope scan direction is reversed, the parity of the scan mirror ramp/flyback is reversed. The pixel sizes for the stressed Ge:Ga (160 µm) and Si:As (24 µm) arrays provide good sampling of the telescope PSF. Improved sampling on a fractional pixel level in the scan direction is obtained on the 70 µm array as the source transits, because the frames are taken at non-integer shifts at its pixel scale. In the cross-scan direction, a slight difference between the slew direction and the scan direction also provides fractional pixel sampling of the PSF.
Figure 3.18: MIPS operation at 160 µm in scan map mode. In normal operation, every point along the scan track is observed once; redundant coverage requires rescan. In fast scan mode, the sky coverage is in bars, one pixel wide (16.0´´) with every other bar not observed. Full sky coverage requires rescan to place the pixels in the uncovered areas on the rescan.
Frequent stimulator flashes are required to calibrate data from the two germanium focal planes (see sections 4.1.1 and 4.1.2). To allow scans that take longer than the time between germanium stimulator flashes (~2 minutes), the motions of the scan mirror and telescope are coordinated. The scan mirror gradually ''gets ahead'' of the telescope, so that the individual frames lead the telescope by increasing amounts as the telescope scan progresses. This lead grows by linearly offsetting from the zero point of the scan mirror motion. When the time arrives for a stimulator flash, the scan mirror resumes its action at the zero point, and the time for one exposure has been made available for the calibration without loss of sky coverage. This process also guarantees that the array is pointed at a previously imaged portion of the sky during the stim flash, allowing for accurate subtraction of the sky signal. This pattern of operation is illustrated in Figure 3.16. The resulting coverage on the sky is shown in Figure 3.17 and Figure 3.18.
Note that the first several frames of a scan leg necessarily have extrapolated stim backgrounds (see section 7.2.3 for more information). In general, but especially in the case of fast scans over bright objects, be sure to have enough background measurements around your source that extrapolated stims are not an issue.