This form can be used to estimate the limiting magnitudes of
objects that can be observed with SpeX for a given S/N and exposure time. The
user should specify which observing mode is requested, the slit size, desired
S/N, seeing (at 2.2 microns), integration time for an individual image (in sec),
total integration time for the observation (in sec), and the number of spatial
rows to extract from the two-dimensional image. (Note that the plate scale for
SpeX is 0.15 arcsec/pixel.) At each wavelength point over the observed range,
the limiting flux is calculated for the input parameters. Magnitude limits are
given for the centre of each filter passband covered (e.g., at 2.2 microns for
the K band). Saturation will occur if the "on-chip" exposure time (i.e., for a
single image) is too long; this is seen as a drop-out in the flux versus
wavelength plot. (Wider slits will also increase the background.)
For background-limited exposures shortward of about 2.5 microns
on-chip exposure times should be less than 300 seconds for good OH line
cancellation. A 120 second on-chip exposure saturates the detector
at wavelengths longer than about 4 microns for a 0.5 arcsec slit.
The flux limits assume perfect flat-fielding and telluric division, and
a nominal instrument performance. Throughput is particularly sensitive
to seeing. Currently, the median seeing at IRTF is about 0.8 arcsec
at K.
The program uses Steve Lord's ATRAN model of the atmospheric
transmission and emission as a function of wavelength for an altitude of 14000
ft, an airmass of 1.2, and a zenith water vapor content of 1.2 mm. The model is
smoothed beforehand to the requested resolution (which depends on the slit size
and observing mode).
The calculations take about a minute to perform, so please be patient,
and make sure your browser doesn't time out prematurely.
MARCH 2001: Measured sensitivity is
about 0.5 mags brighter than calculated by the flux calculator. Please
remember to subtract this from the calculated sensitivity.
Estimated sensitivity for some configurations is tabulated here.
The most likely reasons for this discrepency
are a combination of guiding errors and noise
introduced by sky variability and division by flats and standard,
since throughput, background and instrument noise are well know.
Observers are reminded to reassess sensitivity for the prevailing
conditions while observing. Particular attention should be paid to
guiding. A good way to gauge how well the guider is performing is by
comparing the FWHM of the spectrum (which includes seeing and
the vertical component of guiding) and the FWHM of a star taken
with the guider/slit-viewer.
SEPT 2000: This new version uses a Lorentzian
seeing profile, normalized to unity in a 3'' x 3'' box, rather
than a Gaussian. From measurements of stars in the guider/slit-viewer,
we have found that a Gaussian profile of given FWHM overestimates
the flux transmitted by the slits. The magnitude of the effect is almost
two for the smallest (0.3 arcsec wide) slit but less for the wider slits.
The older version of the calculator therefore overestimated the
point-source sensitivity.
Questions about SpeX and its expected performance should be
directed to John Rayner (rayner@irtf.ifa.hawaii.edu).