previous contents
Previous: How To Get the Latest Copy of This Manual

What's New?

Faster Read-outs and 16-bits!: Major Changes for Fall 1996

New 16-bit A-to-D converters have now been installed, and fully implemented, in all of our CCD controllers. This means that ``digital saturation" now occurs at 65,535 rather than at 32,767, and we've gained a significant improvement in dynamic range and/or read-noise sampling. IRAF handles these 16-bit words as ``unsigned integers" (type ``u"); once you have processed the images they will be 32-bit ``real's." As long as you don't change the defaults, IRAF tape writing (WFITS) and processing will handle this data-type fine, and this change should be transparent---other than the fact that you will have twice the dynamic range for a given read-noise sampling than before this change.

The default gains have been revised to reflect this improvement (see Sec. 3.1.1); sticking with the default gains will generally make you happy.

CCD Default Gain electrons 0.1% linearity 1.0% linearity

(electrons/ADU) @65K ADU (electrons) (electrons)
Cryocam 0.8 52,000 150,000 160,000
F3KB 2.3 149,500 170,000 190,000
GoldCam 1.4 91,000 80,000 80,000
S2KA 2.5 162,500 145,000 160,000
S2KB 2.8 182,000 210,000 230,000
T5HC 3.2 208,000 420,000 450,000
T2KA 3.6 234,000 180,000 190,000
T2KB 3.1 201,500 220,000 250,000
T2KC 1.7 110,500 190,000 200,000
T1KA 3.0 195,500 250,000 250,000

The electronics has also been readjusted to minimize the effects of overshooting for these new default gain settings for the chips used for direct imaging; users who are concerned with the effects of very saturated stars on their imaging data should stick with these default gains to minimize the effect. Spectroscopists should be unaffected by this problem.

Because these A-to-D converters are also faster than the old ones, the read-down times of all of our chips are now shorter. For every 2048 x 2048 image you take, you will realize a savings of some 16 seconds. While this may not sound like much, it translates to an hour or more per night spread among the various Kitt Peak telescopes; over the course of a year, we estimate that this should save about 5-10 nights worth of telescope time in read-out time alone!


previous contents
Previous: How To Get the Latest Copy of This Manual
Updated: 26Aug1996