Planning NAOMI Observations

NAOMI observers will need to:

• Provide a guide star for each target
• Provide PSF-calibration stars, if required
• Predict the AO correction NAOMI is likely to achieve
• Predict signal-to-noise for the proposed observations
• Decide on the appropriate integration time per exposure, number of exposures coaveraged to give one image, number of images at each dither point, number of dither points and size of dither - see the INGRID page for advice.
• Calculate approximate observing overheads

The guide star can be the target itself, if the latter is sufficiently bright and sufficiently pointlike (< 1.5 arcsec, say, to be smaller than the sub-aperture guide boxes on the wavefront sensor). NAOMI has successfully guided on the nuclei of galaxies. For predictions of AO correction as a function of guide-star mag, see the NAOMI performance page. There should be no stars of similar magnitude within ~ 5 arcsec of the guide star, to avoid confusing the wavefront sensor. The H-band image below is of a guide star R = 11 which turned out to be a 0.5-arcsec double. The wavefront sensor guided on this star with no problems, delivering FWHM 0.16 arcsec.

Use Roy Ostensen's guide-star finder to search for possible guide stars close to each of a list of candidates. Star counts, as functions of apparent mag and galactic latitude, can be found in Allen's 'Astrophysical Quantities'.

For tip-tilt mode (guiding only, i.e. no AO correction), a guide star V < 19.5 within 0.9-1.5 arcmin of the target depending on its position angle with respect to the target, and the telescope sky position angle, is required. The reason for the range in useful guide star separations is that the "patrol field" within which the pickoff probe can move is sharply curtailed in certain orientations. The situation is depicted in the following image, which corresponds to telescope sky position angle zero. The small green and red circles in the centre of the image indicate the nominal science target acquisition position on OASIS, and the nominal on-axis position of the pickoff probe, respectively.



Therefore, with the sky position angle of the telescope at zero, the separation of useful guide stars over the entire south-east quadrant from the target is reduced from 1.5-arcmin, e.g. to only ~0.9-arcmin directly SE of the target. In such cases the guide object can still be acquired by appropriately changing the sky position angle of the telescope; increasing sky position angle corresponds to a clockwise rotation of the cardinal directions in this image, so that for sky position angle 90-degrees North is up and East is left. Conversely, when a specific telescope sky position angle is requested, a guide object with target separation greater than ~0.9-arcmin, but less than 1.5-arcmin, may nonetheless be located beyond the limits of the patrol field at the desired sky position angle, and so cannot be acquired. As an illustrative example, a guide star at, say, 65-arcsec due South of its target cannot be acquired in a telescope sky position angle of, say, 45-degrees. This image shows the boundaries of the patrol field in pixels of the AG3 acquisition camera, for which the pixel scale is 0.407 arcsec/pixel.

Finally, the outer parts of the patrol field are vignetted, and due to this guide stars which fall in the outer parts should ideally be brighter than the V-limit of 19.5 mentioned above, to ensure reliable guiding. The level of vignetting in the periphery of the patrol field has not been well characterised, but in practice experience shows it's possible to guide on stars brighter than V~18.5 as far as ~80-arcsec from their target, even with thin cirrus present.

The AO-corrected PSF (point-spread function) typically comprises a near-gaussian diffraction-limited core plus a faint blobby pattern extending over a disk of similar radius to the uncorrected seeing. The form of the PSF varies strongly with radius from the guide star. It also varies with position angle from the guide star, with the major axis usually pointing towards the guide star. The PSF at the science target (if not coincident with the guide star), can be estimated from the PSFs of nearby objects, but the density of stars in the field is rarely high enough for this.

A better approach is to obtain separate exposures of star pairs of similar separation and position angle to the guide star and science target. The separations should be similar within ~ 5 arcec, and the PA within ~ 30 deg. The mag of one of the stars should be within ~ 0.2 mag that of the guide star (this constraint can be relaxed for V < 10). These PSF-calibration exposures should be made every ~ 15 mins, close (~ degs) on the sky to the target. The dither pattern should be similar to that used for the science observations, and the exposures should last at least 1 minute at each dither point.

Use Roy Ostensen's PSF star finder and PSF star-pair finder to locate PSF stars and pairs near specified positions on the sky.

Observing overheads with adaptive-optics systems are higher than for normal observing, and these overheads should be included in the request for time. For acquisition of the guide star by the wave-front sensor, observers should allow 5 - 10 minutes per target. We believe, as of Jun 2003, that there is no settle time for the AO system, after locking onto the guide star (the PSF measured from the first exposure after closing the loop is similar to that from later observations). For dithered observations, allow ~ 7 sec per dither point (this is for movement of the telescope, corresponding movement of the guide-star pickoff probe, and closure of the AO loop), i.e. ~ 50 sec for a 5-point dither. INGRID itself takes ~ 1.5 sec to read out. E.g. if a coaverage of 40 1-sec exposures is required at a given point (to avoid saturation of the INGRID chip at ~ 25000 counts/pixel), the overhead is ~ 60 sec, i.e. 150%. Additional time should be allowed for calibration of the point-spread function (PSF), if required. The PSF delivered by an AO system is a function of radius from the guide star, but can be estimated by observing star pairs of similar separation (to the guide star and target) at frequent intervals during the night (~ every half hour). For extended targets, offset sky exposures will also be required (as for normal IR observing).

For a straightforward programme, and when long PSF-comparison exposures are not required, the observing efficiency (fraction of night spent exposing on science targets) can be high, e.g. 60% on the night 2003 Sep 10 (NAOMI/INGRID observations made by visiting observer Maria Rosa Zapatero). For non-AO spectroscopic observing, the observing efficiency at 4-m telescopes is typically ~ 70%.

Exposures can be made with the AO loop open to check the natural seeing, or if AO correction is not required.