Observing Overheads

Overheads to be used when calculating time requirements in PATT or Service proposals.

The table below gives rough estimates of the overheads one can expect with the various observing modes and instruments at UKIRT. Overheads include time to offset, readout times, times to change the instrument setup, and time to acquire calibrations (e.g. flats, arcs and standards in spectroscopy mode). Although these are largely based on experience, they remain only "estimates": in spectroscopy mode, for example, a faint target can take 2 minutes or 10 minutes to acquire, depending on the quality of the finding chart, the accuracy of the coordinates, and the target itself (clustered, nebulous, etc.).

Overheads associated with WFCAM observing can be 50%-100%. In other words, an MSB that acquires 5 minutes of total integration time on a target can take 7.5-10 minutes to execute. Obviously, longer exposure times, fewer filter changes and fewer telescope offsets between targets or the four positions in a tile all improve efficiency (note that microsteps and "jitters" are done by tilting the secondary, rather than by moving the whole telescope, and are therefore very quick).

When preparing telescope proposals, use the numbers below in conjunction with the UKIRT ITC.

Example:

• You want to observe 10 tile-sized galaxy fIelds in H with WFCAM to a depth of ~20 mag.
• The ITC indicates that, under nominal conditions (seeing 0.9 arcsec, airmass 1.2, etc.), to get a signal-to-noise ratio of 5 you will need a total exposure time of 240 secs at each pointing. For four pointings in each tile, and 10 target fields, you will thus need 240 sec x 4 x 10 = 160 mins.
• You therefore plan on observing each position in the tile with 1 coadd x 10 sec exposures, 2x2 microstepping and a 6pt jitter. This gives you the 240 secs you need per pointing.
• The closest example in the WFCAM table below is number 3, which has an overhead of 50%. In the technical section of your proposal, you should therefore request 160 mins + 50% = 240 mins (strictly speaking, 250 mins if you add 10 mins for focusing - see the WFCAM notes).

General Notes:
Observing sequence - sequence of microsteps and/or jitter offsets (defined in the ukirt-ot with the offset iterator) and filter changes, if any. If a sequence is repeated the efficiency should remain roughly the same.
Exposure time x Coadds - the exposure time and number of coadds for each exposure.
Total Integration Time - total time spent gathering photons, per pointing on the sky.
Execution time - the time needed to execute the entire MSB, e.g. slew to standard (if not WFCAM), observe calibrations (if necessary), slew to target, acquire and observe target, etc. In spectroscopy mode, this includes the time typically devoted to imaging acquisition with UIST, and peakup with CGS4.
Overheads - (Execution time - Total Integration time)/(Execution Time), expressed as a percentage.

WFCAM-Specific Notes:
(1) To observe a WFCAM "tile", the microstep/jitter sequence (a "paw" print) must be repeated four times, and four separate guide stars must be acquired.
(2) A filter change and 1-second flush at the start of an observation, if included in an MSB, will typically take a further 30 seconds.
(3) A single dark taken at the start of an observation will add ~1 min to the total execution time (for exposure times ~ 5-30 sec).
(4) On average, for every 20 minutes observed you will be charged an additional 1 minute to cover focusing.