It is important to know the advantages and limitations of WYFFOS/AF2, as well as to compare it with other spectrographs available on the WHT. Here we discuss the following topics:
WYFFOS/AF2 can presently observe up to 110 objects over a field of 1 degree diameter; though it should be noted that there is some vignetting starting at a diameter of 40 arcminutes. Obviously this is a huge multiplex advantage over a long-slit spectrograph like ISIS, where at best one can observe 2 or 3 objects along the slit. With LDSS-2, one can observe up to ~90 objects if a blocking filter is used, but over a much smaller field of ~10 arcmin; a more typical number is 30-40 objects per mask.
The chief limitation of WYFFOS/AF2 for crowded fields is the minimum fibre spacing of 25-30 arcsec (in the *best* case), which is set by the button/prism assembly on the input fibre ends. LDSS-2 is much better in this respect, with object separation essentially limited by slit length; separations of 5-10 arcsec can be achieved.
It should also be noted that, *unlike* 2dF, at present fibres cannot cross with WYFFOS/AF2.
WYFFOS/AF2 can achieve dispersions as high as 0.8 A/pixel with a 2400 line grating with the present 2.7 arcsec fibres and a TEK CCD with 24 micron pixels. This is to be compared to ISIS, which can achieve ~0.2 A/pixel with the same setup. The highest dispersion available with LDSS-2 is 2.4 A/pixel with the 'high' grism, about a factor of 3 lower than for WYFFOS/AF2. When the WYFFOS echelle grating is implemented in 1997, the dispersion will be ~0.3 A/pixel (R=~9000) with the large fibres.
Because of losses in the (26m) fibres and the fibre connectors, the WYFFOS/AF2 throughput is lower than for both ISIS and LDSS-2. Our best present estimates for the WYFFOS/AF2 sensitivity is 17.5 near 4500 A (this is the AB magnitude giving 1 photon/sec/A). In comparison, the LDSS-2 sensitivity is ~18.5, while that of ISIS is ~18.1, at the same wavelength and also with a TEK CCD. The quoted WYFFOS sensitivity is for one of the best fibres (#26), and the throughputs will be correspondingly lower for other fibres (see the Fibre response table).
Another limitation with WYFFOS/AF2 is the large fibre diameter of 2.7 arcsec. Because of the larger sky contribution, for point sources the S/N will be lower than for slit/multislit spectrographs, where the slit width is typically 1-1.5 arcsec. A smaller set of 1.5 arcsec diameter fibres will be commissioned in 1997, and these will be significantly better for point sources.
Sky subtraction with fibres is never as good as with a slit spectrograph, since one is limited by the fibre diameter. At present the best sky subtraction we have achieved with WYFFOS/AF2 is 3% (based on the rms residuals on sky dedicated fibres in a twilight sky frame), but that was based on the system when it displayed the large scattering caused by the deposit on the field-flattening lens. Further information on the sky subtraction will be provided when we have it to hand. Achieving the best possible sky subtraction with fibres requires careful planning and observing strategies, and inevitably involves tradeoffs. For instance, one can assign an `sky' fibre next to each object fibre, which reduces the fibre multiplexing advantage by a factor of ~2. Another strategy is to `beam-switch' (move the telescope back and forth between objects and neighboring sky), but this of course reduces the integration time on source.
True flux calibration can never be done with fibres, since again one is limited by the fixed fibre diameter. However, with the present 2.7 arcsec fibres and good seeing, one can come close. It should also be noted that flux calibration with multi-slit spectrographs such as LDSS-2 is non-trivial, since one has to change between a wide-slit mask for the flux calibrator and a narrower-slit mask for the arc calibration.
All multi-object spectrographs are much less flexible than conventional long-slit spectrographs, since one has to either design multi-slit masks for LDSS-2, or prepare a Configuration file for WYFFOS/AF2. In both cases, the relative astrometry has to be carefully prepared in advance of the observing run, which is a significant overhead. In comparing WYFFOS/AF2 to LDSS-2, with LDSS-2 masks have to be produced well in advance of the observing run, with the hassle of having to produce a new mask if you wish to make changes. With WYFFOS/AF2, the input Configuration files can be modified at short notice. It should also be noted that WYFFOS/AF2 has no imaging capability, unlike LDSS-2.