3 Observing with XMM-OM

3.1 Specifying windows

The full FOV of XMM-OM is a square 17 $\times$ 17 arcmin, covering the central portion of the X-ray FOV. Within this field the observer can define a number of data collection windows around targets or fields of interest. Up to five different Science Windows can be defined with the restriction that their boundaries may not overlap. However, one window can be completely contained within another.

Because of constraints on the telemetry rate available, it is not possible to transmit the full data on every photon that XMM-OM detects. Instead a choice has to be made between image coverage and time resolution. Thus two types of Science Window can be defined, referred to as Image Mode and Fast Mode. A maximum of two of the five available science windows can be Fast Mode.

Image Mode emphasizes spatial coverage at the expense of timing information. Images can be taken at the full sampling of the instrument or binned by a factor of 2 or 4, to yield a resolution element on the sky of approximately 0.5, 1.0 or 2.0 arcsec (a factor of four finer for the Magnifier). The maximum total size of the Science Windows is determined by the memory available in the DPU. A single Image Mode window binned by a factor of 2 $\times$ 2 can be up to 976 $\times$ 960 detector pixels, which results in a 488 $\times$ 480 binned pixel image being stored in the DPU. At full sampling (with no binning) the window can be up to 652 $\times$ 652 pixels. Any drift in the pointing direction of the spacecraft is corrected in the image by tracking guide stars (Sect. 5.5).

Figure 2: Setup of XMM-OM imaging default mode. There are 5 exposures, each one being made up of two image mode windows: win 0 and win 1. Win 0 is the same for each exposure and is unbinned (0.5 arcsec). Win 1 for each exposure is binned 2 $\times$ 2 pixels and the position changes for each exposure. For exposure 1, win 1 is a square containing the central half of the FOV, with win 0 entirely inside it. For the other exposures, win 1 is arranged to fill in the rest of the FOV

Fast Mode emphasizes timing information at the expense of spatial coverage. The maximum total number of pixels that can be specified for a Fast Mode window is 512. Thus the maximum size of an approximately square window would be 22 $\times$ 23 pixels (= 506 total). Note that there is no binning within a Fast Mode window. The pixel locations of individual photons within the window are recorded and assigned a time tag, which has a user-specified integration time of between 100 ms and the tracking frame duration (10-20 s). No tracking correction is applied to Fast Mode data. This can be applied on the ground, from the drift history supplied by XMM-OM.

To simplify observation set-up, two standard observing sequences of five exposures have been created that together cover the whole XMM-OM FOV at one arcsec sampling while at the same time monitoring a central target at full spatial sampling (0.5 arcsec). In the first variant, each of the five exposures contains an unbinned Image Mode window centred on the prime instrument's boresight (the position of the main target), and a second Image Mode window, binned by 2 $\times$ 2 pixels, that is defined in each of the set of five exposures so as to form a mosaic of the entire field (see Fig. 2). The second variant is exactly the same as the first except that a Fast Mode window is added around the prime instrument's boresight position.

The length of an XMM-OM Image Mode only exposure can be set in the range 800-5000 s. However it should be noted that there is an approximately 300 s overhead associated with each individual exposure. The maximum length of an exposure that contains a single Fast Mode window is 4400 s, or 2200 s if there are two Fast Mode windows.

At the time of writing a further mode is being commissioned which allows the full field to be imaged at 1 arcsec sampling in one go, at the expense of tracking information and correction. This is made possible by the impressive stability of the XMM-Newton spacecraft compared to pre-launch expectation.

Window coordinates can be specified either in detector pixels, or in sky coordinates. To facilitate the latter, the XMM-OM performs a short V-band observation at the start of each pointing. The DPU compares the image with the positions of uploaded field stars to calibrate the absolute pointing of the OM.

3.2 Filter selection

The XMM-OM filter wheel rotates in one direction only and, to conserve the total number of wheel rotations over the expected lifetime of XMM-Newton, the number of filter wheel rotations per pointing is limited to one (unless there are very strong scientific arguments for more). Thus filter observations have to be executed in a particular order during a given target pointing. The filter elements are listed in the order they occur in the filter wheel in Table 2. The instrument is slewed with the blocked filter in place, and thereafter a field acquisition exposure is performed in V.

   

Table 1: OM filters, with the V-magnitude brightness limits for the main optical and UV filters, for a range of stellar types, assuming that they are not in the centre of the FOV. (Note the order of the filter elements in this table is not the order in which they should be used; see Table 2)

	B0	A0	F0	G0	K0	M0
V	7.71	7.68	7.65	7.65	7.63	7.59
B	9.38	9.18	8.83	8.58	8.29	7.68
U	9.79	8.34	7.88	7.57	6.66	4.50
UVW1	9.49	7.55	6.53	5.98	3.89	1.50
UVM2	8.94	6.82	4.53	2.70	-0.23	-2.63
UVW2	8.76	6.55	3.83	1.86	-0.63	-1.80
Mag	10.03	9.68	9.41	9.27	9.13	8.91
White	11.58	10.28	9.72	9.50	9.22	8.93

The same telescope focus setting is used for all the filters except for the Magnifier (see Sect. 5.6.1), where the optimum focus is different (the image quality is the most sensitive to focus position when using the Magnifier). The XMM-OM instrument is optimised for the detection of faint sources. If the source count-rate is too high the response of the detector is non-linear. This ``coincidence loss'' occurs when the probability of more than one photon splash being detected on a given CCD pixel within the same CCD readout frame becomes significant. Coincidence loss is discussed in more detail in Sect. 5.2. If a source is predicted to exceed the coincidence threshold for a given filter, then a different filter with lower throughput can be selected. Alternatively a grism can be selected which disperses the available light over many pixels.

The XMM-OM detectors can also be permanently damaged by exposure to a source that is too bright, reducing both the quantum efficiency of the photocathode and the gain of the channelplates. This is a cumulative effect dependent on the total number of photons seen over the lifetime of the instrument at a particular location on the detector. The deterioration is therefore more severe for longer observations of a bright source. For this reason limits are imposed on the maximum brightness of stars in the FOV (see Table 1) and apply to any star in the FOV irrespective of whether it is within a science window or not. Even more stringent limits are applied to the central region of the detector that will usually contain the target of interest. In the event that there is a star in (or near to) the FOV that violates the brightness constraints, a different filter, which has lower photon throughput, can be selected. Also, if the bandpass is appropriate, the Magnifier can be used to exclude bright stars further than a few arcmin from the field centre.

The grisms (one optimised for the UV, the other for the optical) form a dispersed first order image on the detector, together with a zeroth order image that is displaced in the dispersion direction. The counts in the zeroth order image of field stars determines the brightness limits used for observing with the grisms.