XMM-Newton Users Handbook

3.5.3.5 Practical use of OM. Some examples

The choice of a mode or configuration for OM should be driven by a compromise between the time resolution needed and the spatial coverage required for a given pointing. Time discontinuities are due to the overheads (see XRPS User's manual for a detailed description) needed to download the data and configure the instrument for the next exposure.

The full frame modes cover the whole FOV, but they have large overheads. If time resolution is needed, then windowed modes provide almost continuous coverage. The maximum time resolution is achieved in FAST mode (up to 0.5 sec). The default image configuration, or IMAGE+FAST, allows the user to cover 92 % of the FOV in five successive exposures, obtaining in each of them a small window (less than 2$\times$2 arcmin) centred on the pointing attitude and eventually a FAST mode window ( $10''.5\times10''.5$) centred at the same coordinates. If one needs time coverage on a larger area, one can make several short (> 800 sec) Science User Defined Mode exposures (image window up to $7'.6\times7'.6$, if binned).

Field spectroscopy is better achieved using full frame low resolution, while if one wants to monitor spectral variations of the target, then there is a pre-defined window configuration for each of the grisms which allows to obtain only its spectrum.

Example 1: Introducing an exposure time of 1 ks in the EPIC or RGS image mode (which consists of 5 sub-exposures) gives a total exposure time of 5 ks. There is an overhead of 1.3 ks for the first exposure of observations in windows modes (including the star tracker acquisition time). To this number an overhead of $\sim$1.6 ks must be added to the composite exposure. Therefore, in order to be able to make two exposures, e.g. with two different filters, at least 14.5 ks of total observing time are needed (1.3 ks + [2$\times$(1.6ks + 5.0 ks)] = 14.5 ks). The overhead times are added automatically when entering the data in the RPS tool.

Conversely, if 10 ks are available, only one filter can be used with this mode. This is because if this time were equally distributed between two filters, each exposure would have an exposure time of $\sim$660 seconds, which is below the minimum allowed (cf. Tab. 16).

If Full Frame Low Resolution Mode is used, in 10 ksec we can obtain two exposures of 1.6 ks each, in two different filters: 1.0 ks + [2$\times$(2.9ks + 1.6 ks)] = 10 ks (the FF Low Res. exposure overhead is 2.9 ks). The overhead of the first exposure of the observation for this mode being $\sim$1.0 ks.

Otherwise, one could define a single window (Science User Def mode; with an overhead of 300 s) covering the central 5'$\times$5' at high resolution, or $7.6\times7.6$ arcmin$^2$ at low resolution (2$\times$2 binning), and one could make six 1200 s exposures (e.g. one with each filter): 1300 s + 6 $\times$ (1200s+300s) = 10300 s.

Example 2: A 24 ks X-ray observation allows four EPIC/RGS Image mode OM exposures: 1300 s + [1600 s+(800 s$\times$5)] $\times$ 4 = 23700 s. In this case the preferable filter choice is: U, B, UVW1 and UVM2, unless science needs suggest otherwise.

Example 3: the overhead for each exposure in Full Frame Low Resolution Mode is 2900 s. Within 20 ks of available observation time, at most 3 exposures of $\sim$3400 s each can be accommodated: [1000 s + 3$\times$(3400 s+2900 s)] $\simeq$ 19900 s.

Example 4: Users interested in U, B, V photometry should choose the filters in the order V, U and B. Within a total observing time of $\sim$10 ks, the EPIC/RGS image mode could not be used. However, the Science User Def mode would allow the usage of the three filters. In order to reach similar limiting magnitudes, one observation with the V filter of 3.8 ks and one each of 2 ks with the U and B filter (in that order) can be chosen (cf. § 3.5.5.1) [1300 s + (3800s+300s) +  2 $\times$ (2000s+300s) = 10000 s]

Example 5: If a user just wants to complement a 15 ks X-ray observation with data in a wider wavelength range, then three exposures of 1.8 ks each, in Full Frame Low Resolution [1000 s + 3$\times$(1800 s+2900 s) = 15000 s] using filters U, UVW1 and UVM2, will do the job.