Appendix A

Segment 2A Gain Sag

16 December 2002 through 3 February 2003

David Sahnow

March 3, 2003

From 16 December 2002 to 3 February 2003, the high voltage on detector segment 2A was reduced by approximately 250 volts, to a lower-than-optimal value, due to high voltage transients on that segment.

Effects

While the voltage was lowered, pulse heights were much lower than normal, resulting in significant walk effects for LiF LWRS data. An example of the effect of the voltage change on the pulse height is shown in the following figure. It shows the pulse height for two different observations of the same object taken in the LWRS aperture; since the exposure times were slightly different, they have been normalized before plotting. In early December, the modal gain for both SiC and LiF LWRS apertures was above 10. After the voltage change, the LiF LWRS modal gain dropped to 5, while SiC was approximately 9. Values below 8 are most likely to show significant walk problems, although this depends on the position on the segment.

The figure also shows that the gain has not dropped so much that counts have been lost below the lower threshold, so the flux calibration was not affected.

The figure below shows the effect of this gain sag on stim lamp images taken during this period. The stim lamps illuminate the detector directly in a roughly uniform manner. Shadows of the grid wires at the front of the detector are visible because of the way the lamps illuminate the detector. Each of the plots in the figure shows a y projection of the photons in the region of the LiF LWRS aperture collected during a stim lamp exposure. The x range shown is from pixel 10950 to 11980, but the effects are similar across the entire segment. All are binned by 2 in the x direction in order to improve the signal-to-noise ratio.

(a) shows an exposure taken in February 2001, well before the high voltage was adjusted. The positions of the grid wire shadows are noted with dashed lines. This was a ~15000 second exposure taken in HIST mode. The shadows of the grid wires are clearly visible.

(b) shows a ~7500 second TTAG exposure with far fewer counts taken on 6 December 2002, shortly before the voltage was changed. The shadows of the grid wires are again quite prominent, although there has been a slight shift in the location of a few. This shift is a sign of the gain sag, and is the reason the voltage was raised a few days later.

(c) was taken a few weeks after (b), and under essentially the same conditions, except that the voltage is now at the lower value. There has been a significant shift in the positions of the shadows. These shifts are typically 20 to 30 pixels.

(d) shows the same exposure as in (c), but this time with the walk correction applied. The positions of the lines have shifted back towards their correct locations, but the correction is not perfect; In addition, the line widths have increased.

The next figure shows the effects on a calibrated spectrum. The same calibration target was observed both before and during the time when the voltage was low. The spectra from these two HIST exposures show very clearly how the spectrum on segment 2A is distorted because of the gain sag. The differences between the two are so significant that it is difficult to tell that these are spectra of the same object.

Conclusions

During the period when the voltage was low on segment 2A, there is a significant shift in the reported positions of many photons. For TTAG data, much (but not all) of this is corrected by the walk correction module of the CalFUSE pipeline. For HIST data, however, only a crude correction is applied, so data are likely to be seriously affected. Line locations and equivalent width measurements are likely to be in error on this segment, so results should be carefully compared to data on other segments.

Questions?: Please address questions to fuse_support@pha.jhu.edu.

Last modified Monday, March 3, 2003