Next: The Faint Object
Up: Spectrographs
Previous: The RGO Spectrograph
This spectrograph was commissioned in June 1988, and allows the spectra of
objects as faint as V = 16-17 to be observed with a resolving power
(
) of 60,000 to 115,000 depending
on the detector used.
The spectrograph uses the f/36 coudé focus and is operated entirely
under
computer control. Either the IPCS or a CCD can be used as the detector, though
in practise the TEK CCD now offers superior performance in all but a few
very specialised applications.
Figure 5.2 shows the layout of UCLES, and a full description of the system
is
given in AAO UM 25.2 : The UCL Echelle Spectrographs. (See the
Documentation area of this Manual or the
Documentation area of the AAO WWW
pages)

Figure
5.2: UCLES optical path
- The camera is a folded Schmidt design which incorporates a special
field
flattening and correcting lens. It has a 70cm focal length and a field
coverage of 38.5 x 18.8mm.
- Two gratings are available, one with 31.6 and the other with 79 lines/mm,
both
giving the same dispersion. The main difference between them is that the
free
spectral range (FSR) and order separation available with the 79 lines/mm
grating are 2.5 times those of the 31.6 lines/mm grating. The 31.6 lines/mm
grating thus allows full coverage of the free spectral range of an order
with
the IPCS (standard format) for
wavelengths shortward of 7000Å but has the orders pushed
closely together, with a minimum order separation of only
3.5" (at around
7500Å). With the 79 lines/mm grating, only a fraction of the free spectral
range can be recorded at any time but the minimum order separation is
~9"
(at 7800Å). Thus, in general the 31.6 echelle is to be preferred for
observations of bright objects requiring continuous spectral coverage, while
the 79 echelle is more suitable whenever accurate sky subtraction is an
important consideration.
- The spectrograph has a quartz-halogen continuum source, a deuterium lamp
for
blue flat fields and a thorium-argon hollow-cathode lamp for wavelength
calibration (for an atlas of the arc lamp, see AAO UM 28 and 34: UCLES
spectrum of the thorium-argon hollow-cathode lamp (I and II)).
- The optical performance of the spectrograph is remarkably good, giving
images
with FWHM <
20um over the full field of the
IPCS. For this reason it
is possible to obtain a (non-standard) high resolution mode by reducing the
IPCS pixel size to 10um. This setup
yields a resolving power of 115,000.
The main disadvantages are the small slit size required
(0.35") and the
decrease in spectral coverage, since the detector length is effectively reduced
by 2/3.
- Because of the large format normally used with the IPCS (2048 x 256),
non-linearity effects become evident at counting rates as low as 0.2-0.4
Hz.
When observing bright objects it is important that ND filters be inserted
to
reduce the count rates below this level.
- One of the motivations for building the echelle spectrograph at coudé
is
the inherent stability of this focal station. Repeated one minute CCD exposures
of alpha Centauri taken shortly after
the commissioning period showed that
random shifts over a period of one hour are no worse than 0.005 pixels (peak
to
peak), corresponding to velocity shifts of only about 15m/s. Intended
improvements in the temperature stabilization of the main coudé east
room
may further improve this figure.
- The configuring of the spectrograph is also extremely reproducible. The
spectrograph can be reset to a specified wavelength to an accuracy of better
than 0.3 IPCS pixels. This reduces significantly the frequency with which
exposures of the ThAr calibration lamp need to be recorded. For most
applications one such exposure every few hours (per wavelength region observed)
should be adequate.
- An image rotator is available in front of the slit to compensate for
the
field rotation at the coudé focus. It can be set to track a specific
position on the sky, or to align along the direction of atmospheric
dispersion (parallactic angle).
Next: The Faint Object
Up: Spectrographs
Previous: The RGO Spectrograph
This Page Last updated: Feb 21, 1996, by Chris Tinney.
Note
that the discussion above is out of date in that it assumes the IPCS as a
detector.