A Rough Cost Estimate

Most aspects of the SWIFT project can be costed with reasonable accuracy based on existing facilities. Since we assume that no technological breakthroughs are achieved, the estimates below are conservative.

1.

$50M - The fore-optic and enclosure: a relatively simple 8.4-m telescope that includes primary and secondary, mirror cells, telescope structure, control system, dome, and building. The cost basis derives from Gemini, Magellan, LBT, and the 3DT.

2.

$4M - the 1.1-m corrector which includes the ADC, cells, and control system. The cost basis derives from the cost of the recently installed CTIO corrector (0.49-m), increased by diameter cubed.

3.

$16M - the 4 spectrographs at $4M each, including the collimator, camera, and 3 gratings for resolutions of 500, 5000 and 20,000. The cost basis assumes that the similarly specified but somewhat smaller 4-m version will be $3M plus detectors.

4.

$12M - Optical detector systems, consisting of 4 mosaics at $3M each, including per mosaic, 32 CCDs (2Kx4K), 8 controllers, a dewar, window, and shutter. The cost basis is from the 2 NOAO CCD Mosaic imagers.

5.

$20M - IR detectors consisting of 4 mosaics at $4M each including per channel, 16 2Kx2K arrays, plus controllers. The cost basis is the current rate for a 2Kx2K array at $0.25M.

6.

$1M - Instrument computers: 8 are required for data acquisition at $10K each, including storage for DVD or other media to archive the data. Since the integrations are fairly long at faint flux levels, the data rate is expected to be 50 Gbytes per night, which is not very high by tomorrow's standards. The cost basis is from existing Mosaic computers, and the continuing decline in the cost of computer hardware.

7.

$6M - Software - includes data acquisiton, pipeline reduction and archiving. The cost is based on the other large scale software projects at NOAO (the IRAF project integrated over its entire 15 year lifetime totals $5M in 1999 dollars). Much of the software, including some database and mining tools, could build on the legacy from the SDSS and the STScI archive. Further, we expect that external, diverse sources of funding will be available for the development of sophisticated ``expert'' software for higher level analyses, and we intend to team with external organizations in integrating the software into an accessible analysis package. The cost estimate reflects this expectation.

Thus, the total estimated construction cost for the SWIFT system -- telescope, instrument, and software pipeline -- is $109M. Operations for an integrated spectroscopic facility are relatively inexpensive: we estimate approximately 15 FTEs will be needed to support SWIFT. Except for the spectrograph costs, the other items are fairly secure and conservative. Obviously, technology breakthroughs may reduce this cost significantly and can be expected in the areas of the telescope, detectors, and software.