Acquisition onto the slit (comprising telescope slew, target acquisition and instrument reconfiguration) takes 15 minutes or less for bright targets (point source H < 15 mag) and ~20min for faint ones (point source H > 15 mag). In principle, "blind" acquisitions using a reference source (for H > 18.5 mag targets) take the same time as a normal acquisition, however as with all acquisitions, actual time depends on the complexity and the quality of the Phase II preparation (e.g., finding charts, helpful notes). The program is always charged the actual time used (including errors in the OT definition, but not including telescope-related problems).

Overheads during an observing sequence are usually dominated by telescope motions (nominally about 10 seconds/motion, for small offsets) and by array readout. The readout overhead is equal to the minimum exposure time given in the OT, and is incurred per coadd. The lower noise readout modes have longer overheads because multiple reads are involved, but since the exposure times are usually longer, the fractional overhead is still low. The time to write the file is of order 2 seconds and is usually consumed in the ~10 seconds for telescope motion. For most observations you can assume an overhead (after acquisition) of 10%, provided you set things up rationally. M band observations are an exception to this: assume an overhead of 40% because of the short exposures. (If you insist on working at R=1700 with the short camera in M band, assume 110% overhead – which is why it's not recommended.)

Readout overhead for different readmodes is given on the science detector page. These overheads are NOT currently included in the OT total time calculation. For most observations these are likely to be negligible, with the exception of very short exposure times, such as one uses in the M-band, or very long sequences.  For an 0.5 sec observation, the total time per exposure is ~0.7seconds, nearly a 50% overhead.  The readout overhead is the same for each coadd, however increasing coadds reduces the overhead writing to disk, as well as the total number of files. For very short exposures one should generally use coadds to accumulate the data in one file until it is time to dither on the sky. For long sequences and/or very short exposures, the PIs should calculate the readout overhead and take it into account in their total time allocation.