Coming Soon--The CCD Mosaic Imager

(from KPNO, NOAO Newsletter No. 46, June 1996)

For several years now, work has proceeded at NOAO on developing a large CCD mosaic for wide-field optical imaging. Input from users' and advisory committees, comments from the community, and the interests of the NOAO scientific staff have all contributed to making this a high priority for the optical instrumentation program. The mosaic is now nearing completion. Commissioning tests are planned for this summer and fall. If all goes well, the instrument will be offered for shared-risk use in the spring 1997 semester. This article describes the instrument and gives some information on its expected performance. A detailed description of its capabilities will appear in the Newsletter before it is released for visitor use.

The Mosaic dewar with 2 Arcon controllers atached.

The chosen format of the Mosaic Imager features eight 2048 x 4096 15 um pixel CCDs arranged as an 8192 x 8192 pixel detector. The resulting mosaic is a square about 5 inches on an edge. The gaps between CCDs are kept to about 0.5 mm. We are commissioning the Mosaic Imager with unthinned Loral CCDs. These chips have some flaws which compromise their performance, but they will be suitable for proving the instrument and for certain types of science observations. Their principal defect is the result of a foundry processing error which requires that they be run no colder than -60 degrees C, resulting in a dark current of 100-200 e hr sup -1 pixel sup -1. The read noise is about 10 e- rms, but charge injection in three of the chips results in an effective read noise of about 30 e- rms. The sensitivity of these chips will peak in the red, where they will have about 40% quantum efficiency. Their response is not UV enhanced. We are planning to replace these chips with science grade CCDs as soon as acceptable (thinned, blue sensitive, flat, and with good cosmetics) chips are available to us.

The Mosaic Imager will be available initially on both the Mayall 4-m telescope (at the prime focus) and the KPNO 0.9-m telescope. After one year of scientific operation at KPNO it will be shipped to CTIO and will be available on the Blanco 4-m telescope for one year. The field of view of the Mosaic Imager is about 38' on an edge at the 4-m telescopes and about one degree square at the f/7.5 focus of the 0.9-m telescope. See the accompanying article about the new correctors to be used with the Mosaic Imager at the KPNO telescopes.

The Dewar is a larger version of the standard KPNO universal Dewar. It holds 6.5 liters of LN2 and should have a hold time of about 30 hours. It is radiatively coupled to the chip mount in order to ensure uniform temperature over the entire array.

The filter track, running in a loop over the top of the Dewar, holds 14 filters. A pair of pneumatic cylinders positions the track to any filter in a few seconds. The filters are 5.8 inches square, and inserts will be available to accommodate smaller square and round filters with some vignetting. All filters will be parfocal. Initially, at least 5 filters, B, V, R, I, and Halpha, will be provided. Additional filters will be acquired as funds permit.

The Dewar base plate with two of the eight 2K x 4K CCDs mounted.

The shutter consists of two sliding metal curtains controlled by pneumatic cylinders. Tests indicate that their motion is precise and repeatable, and exposures as short as one second should produce reproducible results.

Along two chords of the Mosaic focal plane, small prisms pick off fields for guiding. These fields, 2.5' x 4' at the 4-m PF, are optically relayed to two independent intensified CCD cameras. The camera video signals are fed to "leaky guiders". The TV fields are imaged through separate small filters in the filter track.

The eight CCDs are read out through four ArCon CCD controllers. These controllers currently run between 50 and 100 kilopixels/s per CCD, yielding an anticipated read time of approximately 83 seconds. However, a design modification is in progress which will allow them to run at twice that speed. This modification is expected to be in place by the start of shared-risk observing in the spring semester of 1997.

Each time the Mosaic Imager is read out, 128 megabytes of data are generated. It is not our intent to send visitors home with many gigabytes of unreduced data; we are putting together a data system which will run a pipeline reduction of each night's observations. This reduction will include bias and dark removal, flat fielding, geometric rectification, and combining of multiple frames. The reduced data (and raw data if desired) can then be written on Exabyte tapes for transport home.

The next few months will represent the culmination of work on the Mosaic Imager subsystems. The entire system will be integrated and tested during three engineering runs on the KPNO 0.9-m in May, June and July. Then in August and September, the first tests of the Mosaic at the KPNO 4-m with the new Corrector and atmospheric dispersion compensator will take place.

We are now entering the time frame during which it would likely be profitable for users to think about the scientific opportunities enabled by the CCD Mosaic Imager's capabilities. Watch the next Newsletter for a Mosaic update. In addition, any of the undersigned may be contacted for additional information.

We are grateful to the large number of people at NOAO who have contributed to this project over the years.

Todd Boroson tboroson@noao.edu
Taft Armandroff tarmandroff@noao.edu
Alistair Walker awalker@noao.edu
George Jacoby gjacoby@noao.edu
Rich Reed rreed@noao.edu