Each observation that was made with Herschel implied certain overheads. These are detailed in the time estimation breakdown and were charged against the observation. The onus was thus on the observer to make observations as efficient as possible, so that precious observing time and thus irreplaceable helium were not wasted on unnecessary overheads.
Herschel took a certain amount of time to slew between targets. The median slew time was found in the early phases of routine observing, as expected, to be of the order of three minutes (although this depended critically on the density of targets in the sky, which differed for different instruments and also the level of population of the database -- having many observations to choose from always made scheduling more efficient because slews could be minimised), thus it was decided that all unconstrained observations were to be charged 180s as observatory overhead for slewing the telescope (for constrained observations a 600s slew overhead was applied -- see Section 6.7.4, “Constrained observations”). Given that this was found to be a fair assessment of the mean slew time required for AORs throughout the mission, this value was maintained and was not adjusted from adopted the pre-launch value.
When making maps there are certain overheads that were implicit in the process. These varied according to the exact observing mode being used and could have a major influence on observing efficiency.
In a raster map the telescope must make a slew, stop and wait for the pointing to be stabilised. Due to the satellite's large moment of inertia the process of acceleration, deceleration and stabilisation adds a significant dead time (of the order of 5s) to the measurement in each position. This value was optimised in the light of in-flight experience and remained stable for most of the mission.
Scan maps have generally been more efficient and added less overhead to an observation than a raster map, although for a scan map, the calculation of the overhead used a complex formula because several variables are involved. In this case the overhead is the acceleration at the start of a scan and the deceleration at the end of the scan, which will have varied according to the length of the scan itself (for short scan legs, as in mini-scanmaps, the telescope will spend a much larger fraction of the time accelerating and decelerating). The telescope then made a small slew to the start position for the return scan.
For large scan maps, square maps were considerably more efficient to carry out than long, narrow strip maps.
Although mini-scan maps look inefficient due to these overheads when compared to point source photometry, in fact they go significantly deeper in the same total time and produced much more accurate photometry (a mini-scan map typically took less than half the time of an equivalent point-source photometry integration). However, mini-scan maps were only suitable for point sources and small targets, as the area that they cover to the maximum depth is only about 1 arcminute in diameter.
Each observation required an internal calibration against black body sources maintained at rigidly controlled temperature. These measurements were essential to the health and success of all observations and were thus charged against the observation. The calibration time was typically in the range 30-300s according to the AOT used.
If the calibration time was less than the slew overhead, it was not charged to the user as an overhead, as the calibration was carried out in its entireity during the slew; when this calibration time exceded the slew overhead that has been applied, the excess was charged as an overhead to the astronomer. Obviously, if two observations were concatenated and no slew was involved, the whole of the calibration block had to be charged against the observation; for this reason, in some circumstances there may still have been a small overhead on concatenated observations, even when no slew overhead is applied.
Constrained observations (see Section 6.5, “Constraints on observations”) limited the telescope scheduling and limited the observing efficiency, producing what were effectively hidden overheads (e.g. the telescope was forced to slew to a point on the sky that would not be picked otherwise, making the overall telescope scheduling less efficient), thus a flat rate of 600s was be charged on all constrained observations, in addition to other observational overheads that might be applicable, as discussed previously.
If a constrained observation was concatenated, the 600s overhead was applied only to the first observation. However, for constraints on other linked observations (follow-ons, group-withins, etc -- see, for example, Section 6.5.4.2, “Follow-on observations”), a 600s overhead was charged on every non-concatenated observation.
For a fuller definition of what constitutes a constrained observation that would be charged a 600s overhead, please see the (Policies and procedures) document.