1. Introduction
A SPYBAL is executed to center the spectrum in the cross-dispersion (or GHRS ydirection) on the diode array. This procedure uses an observation of the spectral calibration lamp and performs a null-deflection in the y direction to ensure that the spectrum does not fall off the top or bottom of the diode array. A SPYBAL actually produces science data consisting of a wavelength calibration exposure. The wavelength used is different for each grating and was chosen to give an even distribution of light over the bandpass sampled. The null-deflection determines an offset from the nominal position (telemetry point (ZSPYBLU)) that is subsequently applied to all science observations.
A new SPYBAL is inserted into an observation program each time a different grating is used and every 90 minutes of alignment time. This means that as long as the grating does not change, no new SPYBAL will be inserted, unless one has several exposures and the total alignment time (not real orbital time) exceeds 90 minutes. If one has several short exposures but does not change grating, a SPYBAL will be inserted when one reaches the 90-minute mark, but only after the completion of the executing exposure and before the next. If one has very long exposures over 90 minutes, the exposure will not be interrupted to insert a SPYBAL and no correction will be made to make sure the spectrum is centered on the diode array. When no correction is done there is a concern about spectrum drifting off the diode array, resulting in reduced signal.
It is common for observers to plan observations in a way that minimizes motions of the carrousel to improve the stability of the wavelengths solutions. Certainly anyone who explicitly requests a wavelength calibration (wavecal) at a central wavelength of interest does not want the carrousel moving between the wavecal and subsequent science observations, nor during the science observation. Note that for science exposures (OBSMODE = IMAGE, ACCUM, WSCAN, OSCAN), any use of Number_of_Iterations > 1 (i.e., repeats) will be treated as a single exposure and will not be interrupted by a SPYBAL. Consequently, this observation strategy prevents SPYBALs from being executed and one risks the possibility of the spectrum not being properly centered on the diode array, i.e. very long exposures are not corrected. For example, if you have a single exposure run of 6 hours, a SPYBAL will only be inserted at the beginning of the observation.
While observers have been discouraged from making use of long exposures
without periodic SPYBALs, they have also been given a measure of control
over carrousel movement in the optional parameter called SPYBAL.
This optional parameter can be used to prevent the insertion of a SPYBAL
between a wavecal and the subsequent science observation. The use of this
parameter has, however, led some programs to be constructed in a way to all
but eliminate SPYBALs. Consequently, the question has been raised: "How long
can an observation really be before a SPYBAL must be executed?"
[Previous][Next] [Up]