AAO Newsletter October 1996 - Page 4

http://msowww.anu.edu.au/~dhj/ttf.html.) Lower resolution imaging (15-75Å bandpass) can be carried out over the full R and I bands, although the performance tails off below 6300Å. Our testing of TTF both on and off the telescope throughout the year now sees the instrument fully calibrated and well understood. In conjunction with charge shuffling and frequency switching, the TTF offers several new ways of observing with TAURUS-2 as demonstrated below.

Astrometry with SuperCOSMOS

Introduction
With the impending completion of second epoch surveys of the northern and southern skies by Schmidt telescopes, there will exist a vast amount of astronomical information in the form of photographic plates. Data are available in blue, red and near infrared wavebands, and at two epochs separated by ~ 20 to 40 yr. For positional astronomy the information content of photographic plate surveys is unique—the UK and Palomar 48 inch Schmidt telescopes have plate scale ~ 67 arcsec mm^-1, so a sub-1 micron precision capability combined with epoch differences of up to 40 years result in relative proper motion accuracy's of the order of milliarcseconds per year.

The SuperCOSMOS microdensitometer
SuperCOSMOS is the latest in a line of microdensitometers designed and built at the Royal Observatory, Edinburgh—it is the successor to the powerful COSMOS machine. The design philosophy of the system was that the precision of data output will be limited as far as possible only by the photographic emulsion being scanned, i.e. the measurement process itself should not degrade the accuracy of any quantity being measured. The full system will be described in subsequent papers; however we give a brief description here of relevant features. A Tesa-Leitz PMM 654 air bearing granite table in a class 100 clean room at operating temperature 20.0 ± 0.05° C forms the core of the xy measurement hardware; a 2048 pixel linear CCD and 500 kHz 16-bit analogue-to-digital converter digitise a 14 inch Schmidt plate at 10µm resolution in ~ 2 hr, producing 2 Gbyte of data. The fixed geometry of CCD imaging gives higher positional accuracy than the 'flying spot' system used in the predecessor COSMOS machine. The pixel map is produced as a series of lane scans; currently the lane width is 1280 pixels (12.8 mm). The pixel map is then thresholded and analysed using similar image analysis software to the COSMOS system. An output file is generated containing 32 image parameters such as co-ordinates, sizes, magnitudes, orientations and shape (the origin of the acronym 'COSMOS') for each image detected. Both the pixel map and image parameter files from the image analysis mode (IAM) are routinely archived and copies sent to users.

Definitions
We define two terms pertaining to the positional accuracy of the measuring table itself: the absolute accuracy is the accuracy to which a position can be measured with respect to a standard length. Good absolute accuracy is needed when measuring plates with different field centres (e.g. POSS I/II) or when tying measurements in to a standard, global astrometric reference frame; the repeatability is the accuracy to which measured positions can be repeated on successive measurements separated by long or short timescales. Good repeatability is important when measuring relative positions on plates with the same field centres. In addition, we note that in any procedure which compares two measurements in order to determine absolute accuracy or repeatability, the standard errors on differences in position will be square root of 2 times larger than the error in determining a position in a single measure. The installation and acceptance tests of the Tesa-Leitz machine were made using gauge