[Previous] [Next] [Up]
2. Investigation
The investigation of the utility of SPYBALs in keeping the spectrum centered on the diode array focused on two areas: 1) the change in SPYBAL offset with time and 2) the degree to which an individual spectrum may be tilted with respect to the diode array.
SPYBAL Drift with Time
In order to investigate the movement of the spectrum with respect to the diode array over time, programs containing multiple SPYBALs were searched for and the value of the ZSPYBLU parameter was pulled from the archive. ZSPYBLU records the offset amount, in y-deflection units, which the spectrum was moved from its nominal position in order to center it on the array. ZSPYBLU was plotted against time for all first order gratings except for the G140M and G200M gratings (for which we lacked sufficient data). For the Side 2 gratings, Figure 1 on page 3 shows the results for grating G160M, Figure 2 on page 3 for G270M, and Figure 3 on page 4 shows the results for ECH-B. Similarly, for the Side 1 gratings Figure 4 on page 4 displays the results for grating G140L and Figure 5 on page 5 shows the results for ECH-A.
ZSPYBLU values for Side 2 GHRS grating G160M over time.
ZSPYBLU values for Side 2 GHRS grating G270M over time.
ZSPYBLU values for Side 2 GHRS grating ECH-B over time.
ZSPYBLU values for Side 1 GHRS grating G140L over time.
ZSPYBLU values for Side 1 GHRS grating ECH-A over time.
The temporal behavior shown in the figures is quite different between the Side 1 and Side 2 gratings. The Side 2 gratings show little or no change with time. Of all the Side 2 gratings examined, there is some indication that the G160M grating may show some drift with time--although it is modest. The lack of change of ZSPYBLU with time indicates that the spectrum, once centered on the diode array, effectively remains centered for periods of up to 15 hours. The Side 1 gratings show, however, a distinctive trend with time. For both the G140L and ECH-A gratings we see significant drifts with time--drifts of up to 25 deflection units.
The science diodes for the GHRS are 40 microns wide by 400 microns high spaced on 50 micron centers. In terms of deflection space, a single diode maps to 8 deflection units in width and 64 defection units in height.The drifts seen for the Side 2 gratings range up to about 10 deflection steps in 10 hours or about 15% of the height of the diodes. The Side 1gratings, however, show motions of up to 25 deflection steps or almost 40% of the diode height.
We suspect that the difference in behavior between the two sides depends on the recent thermal history of the detectors. In all cases where we see significant drifts with Side 1 gratings, the Detector 1 had recently been in HOLD or STANDBY mode for several to many hours. Consequently, the drifts probably reflect the fact that the Detector 1 had not fully "warmed-up" prior to use. The lack of significant drifts for Side 2 gratings may be due to its more constant thermal environment--recall that the Side 2 Low Voltage Power Supply (LVPS) is always kept on (ever since the problems with Side 1 were diagnosed in 1991).
Spectrum Tilt Across the Diode Array
In order to understand how this drift with time could affect a spectrum if not corrected for by recentering with a SPYBAL, one must take into consideration the tilt of the spectrum across the diode array. Figure 6 on page 7 shows the y-deflection function for the first order gratings, i.e. where on the photocathode the spectrum is located for these gratings. This figure also shows the intrinsic "tilt" of the spectra--while the diode array is aligned in the x-and y-directions, the spectra are clearly not. (The echelle relationships are not shown in Figure 6 and were not considered for the purposes of this part of report.) Figure 7 shows the derivative of each of the curves in Figure 6, i.e. the difference, in y-deflection units, between the ends of a spectrum centered at a given central wavelength. It is obvious from Figure 7 that spectra obtained at some wavelengths are tilted more than others. This effect can be pronounced, ranging from a few y-deflections to more than 50 deflections units. The G140L grating has an average tilt of about 45 deflection units. As a practical guide the short wavelength end of the medium resolution gratings suffer from the least amount of tilt.
Consequences of Tilted Spectra Drifting across the Diode Array
Combining the effects discussed in the previous two sections of this report could result in a situation where the ends of a spectrum drift off the diode array if not periodically corrected by a SPYBAL. A "back-of-the-envelope" calculation can give us some idea of how large this effect can be. Take, for example, the case of a spectrum of a point source obtained with G140L. Ignoring the width of the spectrum (something on the order of the size of the SSA or about 8 deflection units) the ends of the spectrum (which differ themselves by about 45 defection units) will be within about 9 deflection units of the edge of the diode array. If we assume a worst case drift of about 25 deflection units, we find that at the end of this time about 25% of the spectrum will have fallen off the edge of the diode array! In the case of extended object observed in the LSA, the effect is much more pronounced. In this case the width of the spectrum cannot be ignored--the width is equal to the size of the aperture or about 64 deflection units. For the case of a G140L observation of an extended object in the LSA, we start out with a loss of light for about one-half of the spectrum--we start out with spectrum already falling off the array. In the time it takes to drift 25 deflection units, some part of all of the spectrum will fall off the array, resulting in a significant reduction in signal!
GHRS y-deflection function for the first order gratings.
GHRS Spectrum tilt across the diode array. This figure shows the derivatives of the y-deflection curves to give a sense of how much the spectrum is tilted for a given central wavelength. For a given wavelength value, the delta y-deflection value gives the difference between the ends of the spectrum.
- SPYBAL Drift with Time
-
- Spectrum Tilt Across the Diode Array
-
- Consequences of Tilted Spectra Drifting across the Diode Array
-
[Previous] [Next] [Up]
Generated with CERN WebMaker