APOGEE Observing

When the observatory is ready to observe a plate, the observatory staff plug optical fibers into the holes drilled into the plates, and map which fibers correspond to which holes (and therefore which objects) by shining light through each fiber from the slit ends. These mapping data are incorporated into one of the Header and Data Units (HDUs) of the apPlate file described below.

Observers mount cartridges containing the plugged plates on the telescope.   A series of 500s exposures are made for each plate. For most APOGEE plates, on most nights, 8 exposures are obtained on any given night, although this number can vary based on available time, observing conditions, etc.

The resolution of the APOGEE spectra in comparison to the pixel size of the APOGEE detector leads to the property that the spectrum from a single exposure slightly undersamples the resolution at the short wavelength end. To avoid the challenges of working with undersampled data, APOGEE spectra are taken in pairs, with the detectors shifted slightly (by a distance of one half of a detector pixel) between the two exposures of the pair; we refer to this shifting as dithering, and each “dither pair” of exposures includes observations at these offset dither positions. A standard 8-exposure APOGEE observing sequence consists of exposures at the two differ dither positions (A and B) and typically taken in the pattern ABBA ABBA. The data reduction requires exposures in dither pairs, so that any unpaired exposure is discarded.

APOGEE’s infrared detectors have the capability to be read “non-destructively”, so that the amount of charge per pixel can be detected without affecting that charge. This permits the levels on the detectors to be measured during the expsosure. While readout noise can be significant for a single read of the detectors, the ability to read them multiple times during an exposure allows for a reduction in the net readout noise in the final exposure. For APOGEE, the detectors are read in an “up-the-ramp” mode where the detectors are read every 10.7 seconds. Thus, a single exposure generally consists of 47 readouts over an exposure time of 500 seconds. Because of the multiple readouts, the raw APOGEE data for an exposure are actually in a “data cube”, with two of the dimensions representing the location on the detectors and the third dimension representing the time sequence.