Calibration

The calibration unit is mounted at the broken Cassegrain focus and contains one slide with lamps and two filter slides. The unit is operated from the AF2 control system.
As the lamps are on a slide, only one lamp can be used at a time.

The default set of lamps mounted in the WYFFOS calibration unit (and their position number) are:

Quartz-Tungsten-Halogen	1
Neon	2
Mercury	3
Helium	4

Also available are:

Thorium-Argon

Cadmium

Zinc

Any of these lamps can be mounted in the calibration unit if requested by the observers, who need to let their support astronomer know at least two weeks in advance.

The filters currently mounted in the filter slides A and B are:

Filter slide A:

ND 0.5

ND 1.0

ND 2.0

ND 3.0

Filter slide B:

GG395

These filters will be used for attenuating or blocking the light coming from the calibration lamps.

The system works by shining the light onto the tertiary mirror which is then relayed to the fibres at prime. Consequently, light from the arc lamps can be superimposed on the starlight. Hence, primary mirrors petals should be closed when taking arc or flat frames.

The Neon and Helium lamps cover the entire low and medium resolution spectral range (370 - 950 nm). Usually the Neon lamp is used for calibrating the red end of the spectra and the Helium for the bluer part.

Grating	Neon plots	Helium plots
R158B
R158R
R300B		R300B
R316R	7500A zoom BLUE RED
R600B		600B 5100A
R600R
R1200B	1200B 4900A	1200B 4900A
R1200R
H2400B

The Hg lamp is mainly used for calibrating in the blue with high resolution grating, e.g. H2400B, R1200B.
Cd and Zn arc lamps are recommended when observing in the bluest echelle orders (orders 6 and 7).

The ThAr hollow-cathode arc lamp, with its rich line spectrum, is useful for those science programs that aim for high radial-velocity precision in the red.

The flat field exposures are used mainly for correcting the pixel-to-pixel variation and for tracing the fibre spectra on the detector.

The flat frames can be either QTH lamp frames or twilight sky exposures. The QTH flats have the advantage of a smooth spectrum that can be easily normalized with a low-order function. However, due to the difference in the optical path with respect to the science exposure, this is not optimum for flat-fielding correction, e.g. the FWHM in the spatial direction is narrower than in the science spectra. Sky flats are better for this purpose, but have the disadvantage of showing some absorption lines, which make normalization of the spectra difficult, a high order function being needed.

For tracing the fibres, both QTH lamp and sky flats are equally good. However, the QTH lamp has a very red spectrum, showing low flux at the blue end, and this can be a problem when using a low resolution grating, as the red part can be well traced while there are not enough counts in the blue part to find the trace. In these cases the sky flats are recommended.


Sky subtraction with fibres is never as good as with a slit spectrograph, since one is limited by the fibre diameter. Achieving the best possible sky subtraction with fibres requires careful planning and observing strategies, depending on the type and magnitude of the science targets, and inevitably involves tradeoffs.

For instance, for bright objects it should be enough to assign a 'sky' fibre next to each target fibre. This technique allows measurements of the sky simultaneously with the science spectra, which is an advantage, but it reduces the fibre multiplexing advantage by a factor of ~2.
When the targets are faint this is not the best strategy, as the relative throughput and the small differences in the spectral resolution of the fibres prevent an accurate sky substraction that can pollute the science target spectra.

In the case of faint targets, a better strategy is to 'beam-switch' (move the telescope back and forth between objects and neighbouring sky), so that the sky spectra are measured on the same fibres as the science objects.
The disadvantage of this method is that it reduces the integration time on source, and also that the sky is not measured simultaneously, hence the sky conditions need to be very stable in the spectral range of interest to get a good sky substraction. E.g. it might not be the best option when observing at the red end, where the sky lines are strong and very variable.


Flux calibration can be done by placing a spectrophotometric standard star in a fibre used in any configuration. It is recommended to take spectra of the standard through 2 or 3 different fibres, to check that the derived flux calibration is not sensitive to possible errors in the measured relative fibre througput.

Note that it is difficult to get absolute flux calibration with fibres, because of the flux fraction lost outside the fibre. An aperture correction, depending on the seeing, need to be applied.