Two cameras ("green") are required for the [OIII] coverage, 4950 - 5070 Å and one ("red") for the Halpha coverage, 6488 - 6645 Å.
We are in principle interested in commercial suppliers for both, but especially for the red CCD. The green CCDS have to be state-of-the-art and are therefore likely to be prohibitively expensive for us, so we would use observatory units.
Specs are: 15 micron pixels, 2048 x 2048. This gives a sampling scale in the spatial (non-dispersed) direction of 3.30 pxl/arcsec (3.00 pxl/arcsec in the dispersed direction). Corresponding field sizes are then 10.33 and 11.42 arcmin, respectively.
Larger pixels would lead to poorer sampling. E.g. 24 um pixels would be sub-nyquist at better than 0.7 arc seeing.
Of course the pixel size can be smaller (but not larger) if the total number of pixels is also increased to cover the same area. EEV chips with 13.5 um pixels would reduce the field if the number of pixels in any direction is less than 2300. So a 2k x 4k chip would reduce the field size to 10.33 and 10.28 arcmin respectively if arranged appropriately.
The quantum efficiency should be as high as possible in the range, preferably at least 80%.
The maximum integration time, with low (4 electron) read noise and negligible dark current, should be at least 20 min.
Quote: we received a quote (Fairlight NL, 10/98) for a closed-cycle cooling head capable of housing the "standard Andor CCD" chips. Including the controller card but no chip this was just under GBP 36,000 ($A 54k). estimate for a 42-40 chip was GBP 48,000 (73k).
See also the LPO detector page.
The quality is not quite so critical as this is an imaging camera operating along side the main spectroscopic instrument.
Quote: an informal estimate from Apogee for a 14-bit camera based on a cooled Thomson 7899 2048x2048 (14 um) chip, best cosmetic grade, is $A 34k.
