The performance of the CalUnit in various observing modes has been modelled. It was found that, for flat-fielding the majority of instrument configurations, there are sufficient photons from the continuum sources to saturate the detector in the minimum exposure time. The signal/noise achieved in a single frame is limited by the well-depth of the detector to ~300. For higher signal/noise observations, multiple frames are required. For optical observations, the exposure time required to take a flat-field frame is dominated by the time to read out the CCD. The goal read-out time for calibration frames is 10s. For InSb detectors used in the NIR, the read-out time is much less (~0.2s). Table 3 shows some representative exposure times for the Phase I instruments.
The estimates for the wavelength calibrations assume that lines 10^-2 the strength of the brightest lines must be measured with a signal/noise of 10 per pixel. The exact time required will depend on the arc lines present in the spectrum at the wavelength of the observations. The spectral radiance for the brightest hollow cathode lines is taken as 0.04Wm^-2 sr^-1; for the low pressure gas lamps it is 0.01Wm^-2 sr^-1.
The goal for the exposure times for calibration frames was
stated above as 10s to measure a frame with signal/noise 10^3.
The values in Table 3 show that
this is not met. The limiting factor is neither the
transmission of the CalUnit nor the radiance of the lamps, but
the array well depth coupled with the read-out time. Hence, the
calibrations are as efficient as possible with current arrays.
The impact of the time required to calibrate on the overall
observing efficiency can be seen by considering the way in
which the instruments will be used. One of the most demanding
modes of observing is with the 120,000 resolving power with
HROS. A wavelength calibration frame is required ever 15mins to
correct for flexure effects. In the following it is assumed
that lines 10^-3 of the brightest arc lines are required. The
timescales an be broken down as follows:
exposure on source: 15x60s
move the science fold mirror: 15s
calibration exposure: 60s
array read-out: 10s
move science fold: 15s
total time required: 1000s
calibration exposure: 60s
calibration overheads (moving science fold, array read-out):
40s
efficiency of observations (900s/1000s): 90%
Table 3: The time required to complete a
calibration frame with the CalUnit, for a line 10^-2 of the
strength of the brightest lines.