Previous Article
Next Article
Table of Contents

CCD News (1Jun93)
(from CTIO, NOAO Newsletter No. 34, 1 June 1993)

CCDs on 0.9-m Telescope:
The demands for our single Tek 2048 CCD on the various telescopes are
sufficiently high that it is proving increasingly difficult to make it
available on the 0.9-m, our smallest imaging telescope. In addition, we are
trying to reduce the number of CCD changes at the 0.9-m and to use only ARCON-
operated CCDs, which should improve reliability and decrease the burden on
support personnel.

We are examining various options. One would be to permanently schedule a
Tektronix 1024 at the 0.9-m.  These CCDs have poor UV response, which would
mean that most UV work would have to be moved to the 1.5-m, and wide-field
programs would suffer (in the past five months the Tek 2048 has been scheduled
for 44 percent of the nights at the 0.9-m).  A second option is to install the
STIS Tek 2048 in an ARCON dewar and make this (and the Tek 1024) the only CCDs
available for the 0.9-m. The STIS Tek 2048 is a coated, front-illuminated CCD.
It has three times the area of the Tek 1024, will do UBVRI, but has peak QE of
45 percent rather than the 75 percent of the thinned Teks. This CCD and a
Tek 1024 thus would complement each other very nicely.

We will have a clearer idea of what we can and cannot support by late July;
watch the next issue for details.

The Short Career of Reticon #3:
In Newsletter No. 33 we reported the performance of Reticon #3 CCD, which at
that time had just replaced Reticon #2 in an Air Schmidt dewar, for use
with the three 4-m spectrographs. Reticon #3 was measured to have very high QE,
particularly in the blue where it was found to be a factor of two better
than Reticon #2. Unfortunately, this is no longer the case. There appears to
have been some deterioration in the ITO (Indium Tin Oxide) bias flashgate, and
the QE-pinned condition is no longer achievable. We have thus re-installed
Reticon #2 and returned Reticon #3 to the manufacturer for evaluation. Many
people will have, of course, assumed the availability of Reticon #3 when
calculating exposure times on their Semester II observing time applications. An
appraisal of exposure times is part of the technical evaluation of all
proposals, and this information is available to the Time Assignment Committee.
In some cases, the Tek 1024 plus Folded Schmidt camera is now a better option.
As in previous semesters, those people who are successful in gaining time on
the 4-m spectrographs will be advised of the particular camera/detector
assigned to their project along with the letter advising the time allotment.

Tek 2048 System Non-Linearity:
Peter Stetson found that his photometric data, obtained last December at the
1.5-m with the Tek 2048 CCD, shows evidence for non-linearity of the CCD system,
at the level of almost 2 percent per magnitude. Photometric users have already
been advised; the only other use of this CCD has been with the Echelle
spectrograph long camera where a non-linearity of this size is unlikely to be
of great significance. The non-linearity is best corrected as the first data
processing step; anyone with Tek 2048 data taken over the past year should
contact the undersigned if they are uncertain of how to proceed.  To a good
approximation the raw data can be linearized by the relation

                   I = I(raw)* [A + B*I(raw)/32767]

where A=1 and B=-0.03 (until 11 March 1993).  From 11 March until 6 April 1993,
B=-0.023. Since the response of this CCD is relatively flat, to a very good
approximation processed frames can also be corrected by the same formula, after
adding the bias level to each datum. We use the (local) IRAF task IRLINCOR,
which can be obtained on request.  Alternatively, for photometry, a correction
can be applied to the derived magnitudes of stars. The CCD FET bias voltages
have now been re-optimized, and the Tek 2048 is now linear to about 0.1 percent.

New CCDs:
We have received a thinned Loral 800 x 1200 CCD from the Steward
Observatory-NOAO collaboration.  This CCD is AR coated and has impressive QE,
peaking over 90 percent and exceeding 50 percent from 3500-9300 A. This CCD
will replace the EEV CCD at the 1.5-m spectrograph, but due to pressure
from other projects this will not take place for another 3-4 months. We also
anticipate receiving a thinned Loral 3072 x 1024 CCD. This will be optimized
for the blue and will go in a modified Air Schmidt dewar, for use with the 4-m
spectrographs.  This is a quite complicated installation and will use an ARCON
controller.

Alistair Walker

Previous Article
Next Article
Table of Contents