The IRAC observations are the fundamental "backbone" dataset. The field is observed with the IRAC mid-infrared instrument on Spitzer as part of the routine calibration suite. However, the cadence is non-uniform as is the pointing, so the following may be helpful. Early in the mission, these observations occurred at every instrument turn-on and turn-off (the three Spitzer instruments being turned on and off in roughly two week instrument campaigns), as well as every three days. After several years this was scaled back to just the instrument power events. In the warm mission (after the cryogen ran out in year 5), the instrument was left permanently on. As a result, the calibration suite was simply executed roughly every two weeks. However, without cryogen the two long wavelength detectors (5.8 and 8µm) no longer functioned, so only 3.6 and 4.5µm data are available for 2/3 of the mission lifetime.
Furthermore, all four detectors are not co-aligned. The 3.6 and 4.5µm detector view the same field of view simultaneously, as do the 4.5 and 8µm detectors. However, the two fields of view are separated by roughly a half-detector width. The dark observations were executed as 3x3 mosaics. The figure at right illustrates the coverage pattern for the two FOVs (in red and blue) for simultaneous observations during one calibration campaign.
Furthermore, as the spacecraft orbits the sun with an orbital period of roughly one year, the rotation angle of the instrument rotates with respect to the sky at the rate of roughly one degree per day. As a result, the relative location of the two FOVs is constantly changing. In effect, they rotate around the point between them, which is tied roughly to the center of the Dark Field. Slight adjustments are made during the scheduling process in order to avoid several bright stars that lie near the edge of the field.

This link is a movie of the actual 3.6µm data as it is taken over the course of the mission, illustrating the observation pattern.