To: SXT Team From: L. Acton Date: 5 June 1993 Subj: SXT Calibration Note 34 - Preliminary draft. Effect of SXT neutral density x-ray filter on data Refs: SXT Calibraton Note 5, The SXT Neutral Density Filters SXT Calibration Note 29, Preliminary Calibration of SXT Flight X-ray filters. Solar Physics SXT instrument paper (Tsuneta, et al.) CONCLUSIONS: 1. ND filter transmission (0.0805 +/- 0.0008) given in references is OK for use until in-flight calibration is done. 2. Exposures given by GT_EXPDUR are the best available. 3. The ND filter scatters some x-rays. 4. Be very careful in deriving Te for FAINT parts of the image, do not mix ND and non-ND images until calibration is better understood DATA: Behind-the-limb X flare of 920628.0525. RESULTS FOR AL12: The SXT_PREP corrected signal in the entire PFI agrees to better than 0.2% between the sum of 36 frames with ND filter and 36 frames without ND filter. Very faint signal in upper RH corner of the PFI increased 50% in ND frames over non-ND frames. Scattered x-rays in 512 pixels on the disk increased 24% in the ND frames over the non-ND frames. Signal in the brightest pixels (>10,000 DN/sec) decreased by 0.7% in the ND frames as compared to the non-ND frames. RESULTS FOR OTHER FILTERS: Effect of ND filter on Be images is somewhat greater than for Al12 but still minor in the bright pixels. Other filters not yet examined. WORK TO BE DONE: Do in-flight calibration of ND filter. Further work on quantitative evaluation of effect of ND filter for all x-ray filters. --------------------------------------------------------------------- From nitta@sxt3.space.lockheed.com Mon Jun 21 17:53:26 1993 Received: by sxt4.oscs.montana.edu (5.57/Ultrix3.0-C) id AA09897; Mon, 21 Jun 93 17:54:46 -0700 Received: by sxt3.space.lockheed.com (5.57/Ultrix3.0-C) id AA13517; Mon, 21 Jun 93 17:56:18 -0700 Date: Mon, 21 Jun 93 17:56:18 -0700 From: nitta@sxt3.space.lockheed.com (Nariaki Nitta) Message-Id: <9306220056.AA13517@sxt3.space.lockheed.com> To: acton@sxt4.oscs.montana.edu, bruner@sag.space.lockheed.com, freeland@sag.space.lockheed.com, hara@flare2.solar.isas.ac.jp, hudson@isass1.solar.isas.ac.jp, lemen@isass0.solar.isas.ac.jp, linford@isass0.solar.isas.ac.jp, morrison@isass2.solar.isas.ac.jp, ogawara@himiko.astro.isas.ac.jp, shimizu@flare2.solar.isas.ac.jp, slater@sxt.space.lockheed.com, strong@sag.space.lockheed.com, tsuneta@flare2.solar.isas.ac.jp Subject: Neutral density filter and scattered light Status: RO Hello All. I have been analyzing a flare on 22-Apr-93 with N. Gopalswamy of the University of Maryland. While I made a light curve of bright areas in the SXT flare mode full-resolution images with the thin aluminum filter, I noticed a fluctuation that did not seem to be real. The amplitude was 20-40 %. The fluctuation was found to be between the exposures in 1.0 Open/Al.1 and 0.1 NuDen/Al.1. My initial suspect was on the fastest shutter speed, which I imagined might be uncertain. Next thought was an error in background subtraction augmented by the small denominator during exposure normalization. However, after analyzing several flares, I now conclude that the dominant cause is the neutral density filter with enhanced scattered light. The reasons are that the fluctuation is seen only in sequences with alternating Open and NuDen. It is seen IRRESPECTIVE OF what the filter B is and what the exposures are. Note that the fluctuation with such a large amplitude is NOT seen in flares before November 1992, based on the flares on 15-Nov-91, 26-Jan-92, 16-Feb-92 and 2-Nov-92. I did see fluctuation of a comparable amplitude in the last example (X9 flare), but it turned out that it was due to saturated pixels going out of the field of view. The two light curves with and without the neutral density filter became consistent after the saturated pixels were confined in the field of view. There is a tendency that the couting rates become consistent between Open and NuDen (see below). I summed up all the pixels > 0 in each frame that was background subtracted (using leak_sub) and registered in sxt_prep, and divided the figure by the exposure duration as given by gt_expdur. In order to decrease the effect of the flare, I averaged about 10 frames in each of Open and NuDen. Following is what I got. Date Interval Filter Counts Ratio 17-Feb-93 10:49-10:52 (appr.) 1.0 Open/AlMg 1.304e5 0.1 NuDen/AlMg 1.389e5 0.939 same 1.0 Open/Al.1 2.063e5 0.1 NuDen/Al.1 2.138e5 0.965 same 28.0 Open/Be119 6.422e3 9.5 NuDen/Be119 1.179e4 0.545 same 17.0 Open/Al12 1.023e4 6.3 NuDen/Al12 1.621e4 0.631 22-Apr-93 around 14:10 1.0 Open/Al.1 9.277e4 0.1 NuDen/Al.1 1.165e5 0.796 same 1.0 Open/AlMg 6.129e4 0.1 NuDen/AlMg 8.873e4 0.691 same 17.0 Open/Al12 5.841e3 6.3 NuDen/Al2 8.144e3 0.717 18-Apr-93 19:14:16-19:16:36 1.0 Open/Al.1 3.486e5 0.1 NuDen/Al.1 3.693e5 0.944 19:27:18-19:29:28 1.0 Open/Al.1 1.733e5 0.1 NuDen/Al.1 1.848e5 0.937 19:14:30-19:17:06 1.0 Open/AlMg 2.287e5 0.1 NuDen/AlMg 2.457e5 0.931 19:27:16-19:29:26 1.0 Open/AlMg 1.083e5 0.1 NuDen/AlMg 1.227e5 0.883 19:25:44-19:27:54 17.0 Open/Al12 1.054e4 6.3 NuDen/Al12 1.398e4 0.754 19:25:34-19:27:44 28.0 Open/Be119 6.112e3 9.5 NuDen/Be119 8.967e3 0.682 21-Mar-93 03:36:10-03:38:30 1.0 Open/Al.1 1.290e5 0.1 NuDen/Al.1 1.598e5 0.807 03:21:54-03:23:16 28.0 Open/AlMg 9.758e3 9.5 NuDen/AlMg 1.375e4 0.710 03:34:20-03:36:30 17.0 Open/Al12 6.814e3 6.3 NuDen/Al12 1.102e4 0.618 I must admit there may be smaller (e.g. 1-2 %) fluctuation in the data even before November 1992. However, what I present here shows much larger fluctuation, maybe pointing to the need of discussion on how flare data after November 1992 with the neutral density filter should be analyzed . Cheers, Nariaki Nitta