Files you think you should have, but don't (or extra files you have but think you shouldn't).
Because all three of the arrays are ''on'' whenever MIPS is on, even if an observation requests just 24 micron data, data are obtained in the other arrays. This makes intuitive sense in the case of scan maps, where valid data are clearly simultaneously obtained in all three arrays. However, what may surprise you is that even for 24 µm photometry mode observations, there are data obtained at both 70 and 160 µm. Note that these other two arrays image slightly different parts of the sky compared to the 24 µm array; they create in the end unfilled maps because the telescope and scan mirror motions are optimized for the 24 µm array. In this specific case, the stimulator frequency is not set properly to obtain useful 70 and 160 µm data. Here, the 24 µm data are the ''prime'' data set, and the Ge data are the ''non-prime'' data. For 70 or 160 micron photometry sequences, the 24 micron data are perfectly fine to use, but they still create an unfilled map. The implications of this data taking strategy are summarized in Table 6.3 below.
The FITS header keyword PRIMEARR tells you whether data in a given file were obtained with the prime array; PRIMEARR=1 for prime data, 2 for data that are not prime, and 3 for data that are not valid at all (e.g., Ge data during 24 µm photometry). Mosaics are only produced for PRIMEARR=1 Ge data, and BCDs are not even produced for PRIMEARR=3 data. Moreover, in certain scan mirror positions, it is possible to obtain PRIMEARR=3 data for 24 micron data; in this case, the 24 micron array is in the dark. Note also that non-prime 24 micron data may not be correctly flat-fielded given the pick-off mirror contamination; see section 7.1.1.
The Archive will serve you non-prime data - it provides BCDs for PRIMEARR=1 or 2, but not 3. It only provides mosaics for PRIMEARR=1, however. If you seem to have much more photometry mode data than you expect, this may be the reason. If you recreate your mosaic and seem to have non-contiguous sky coverage, this may also be the reason - you have combined PRIMEARR=1 and 2 data. If you look at your photometry mosaic and find that the data appear to have missed the target, you are probably looking at your PRIMEARR=2 data rather than your PRIMEARR=1 data; your target will be centered in the PRIMEARR=1 data.
Because MIPS BCDs are high quality, most people should not need the raw data, but it is currently provided (for PRIMEARR=1, 2, or 3) through the Archive interface.
Table 6.3: Prime and non-prime data.
Notes on the non-prime data
All three arrays
No non-prime data
24 µm photometry
Ge data has stims without proper background, covers unfilled portion of sky. PRIMEARR=1 for 24; PRIMEARR=3 for Ge.
70 µm photometry
24 and 160 µm data covers unfilled portion of sky, but should produce useful serendipitous data. Mosaics may be produced but may look strange due to the amount of ''dead space.'' PRIMEARR=1 for 70; PRIMEARR=2 for 24 and 160.
160 µm photometry
24 and 70 µm data covers unfilled portion of sky, but should produce useful serendipitous data. Mosaics may be produced but may look strange due to the amount of ''dead space.'' PRIMEARR=1 for 160; PRIMEARR=2 for 24 and 70.