ALFOSC: Setting the proper rotator angle to minimise slit losses
due to differential atmospheric refraction
When doing long-slit spectroscopy of point sources, determining
the the proper rotator angle which aligns the slit along the
atmospheric dispersion direction is trivial with an Alt-Az telescope.
Simply put, the slit should always be oriented perpendicularly to the
horizon, which also happens to be perpendicular to the altitude axis
of the telescope.
For ALFOSC slits that image along the rows of CCD#8 (horizontally),
this can be achieved in normal operation of the
telescope, with the rotator in automatic mode by issuing the TCS
(that's a zero, not an "oh").
As of autumn 2006, we have a new telescope presetting command,
which will put the rotator tracking at a predefined (parallactic)
angle. To define this angle use
at the start of the night (this angle stays until a new one is given).
Then on the TCS interface use CATALOGUE function key 6 'Preset w par
ang' instead of the usual 4 'Preset to this'.
Alternatively, on the Sequencer use
The telescope will then go to the selected object and
put the rotator to the defined parallactic angle.
Note that this sets the slit appropriately at that instant only; the
rotator continues to rotate while tracking/guiding and thus the slit
drifts away from the parallactic angle with time.
For vertical slits in ALFOSC the parallactic angle corresponds to
rotator positions of +90 degrees, i.e., use
instrument-parallactic-angle 90 before presetting with
6 'Preset w par ang'.
When swapping from ALFOSC spectroscopy to ALFOSC imaging, dont
forget to reset the rotator to field-rotation +90 if you want
North up in the images.
Why do this?
The atmosphere is a dispersive medium, and the greater the zenith
angle, the greater the effect will be. Thus, at high airmasses, a
stellar image will be spread out into a small spectrum and this spread
can be significant compared to the seeing disk size. Under these
conditions, if the slit is not aligned along the dispersion direction,
light of some wavelength ranges will be lost through obstruction by
the slit jaws while other wavelengths will pass into the instrument
unhindered. Typically, this will manifest itself as a varying
continuum in the observed source, with one end of the spectrum
steadily losing signal as the object descends towards the horizon.
The specific wavelengths lost will primarily depend on the relative
colours of the object used for guiding and the target object -- for
example, a relatively red guide star will cause the telescope to
follow the red part of the target object and blue light will be lost.
The obvious solution to this is to orient the slit with the
parallactic angle. That is, the slit is oriented so that it is always
perpendicular to the horizon, since this is the direction in which
starlight is dispersed. The above procedure will acheive this and
minimise losses due to differential atmospheric refraction.
Calculated Differential Atmospheric Refraction (DAR),
normalized at 5000 Å, for typical conditions at the NOT (T = 10 deg,
R.H. = 30 %, and Pressure = 770 mbar), using a routine written by E. Marchetti,
ESO. The graph shows values for 5 different airmasses (1.1, 1.2, 1.3, 1.5,
For more details see tabulated values in wavelength steps of 500 Å for
airmasses from 1.0 to 4.9 in steps of 0.05 in the ascii files linked here:
Taken from the 1994 ING Observers' guide: parallactic angle as a
function of telescope position.