Annotation of /ccfdev/packages/emos-quantumef/ChangeLog

ChangeLog for emos-quantumeff
============================
VERSION 1.12 - 2004-04-20 (RDS)
-----------
        Issue 0016
        Produced from a new calibration of MOS chips used on the
    SWIFT project. The changes affect the QE_TOTAL array of
    both cameras below 1 keV.
        
VERSION 1.11 - 2003-06-03 (RDS)
-----------
        Issue 0015
        Uses realistic channel space pattern fractions for Timing
        mode (SPR-2422)
        Includes key-words to describe the ratio of large pattern
        events (SPR-2421)
        
VERSION 1.10 - 2002-12-16 (RDS)
-----------
        Issue 0014
        Set the pattern fractions in channel space more accurately so
        that they can be used within epatplot
        
VERSION 1.9 - 2002-11-22 (RDS)
-----------
        Issue 0013
        Add in pattern fractions for timing and burst modes
        Add in a CHBINS_FRACTION extension
        
VERSION 1.8 - 2002-06-28 (RDS)
-----------
        Iss 0012
        Return to the values in Version 1.6 which are now further
        substantiated by work by Philippe Marty.
        
VERSION 1.7 - 2002-03-27 (DHL)
-----------
        Iss 0011
        Seems like that SAS v 5.3 won't go out with the required
        vignetting
        corrections so we need a quick fudge to get the MOS2 and MOS1 QE's
        back into an alignment that temporarily gives reasonable results,
        while not maintaining the old Leicester QE's with known
        deficiencies. MOS2 only changed..
        
VERSION 1.6 - 2002-03-07 (DHL)
-----------
        Iss 0011
        Made an error in adding up only patterns 0:12 into the QE_TOT 
        extensions
        So this version should now have patterns 0:31 
        
VERSION 1.5 - 2002-03-01 (DHL)
-----------
        Iss 0011
        Fixed the data for the QE, by making the high energy (>1.5keV) points
        match a detection depth provided on a CCD-CCD case from the Orsay data
        (see below). At lower energies I just take the existing sets from the
        Leicester response matrices (q20 versions) and merge into the higher
        energy data. I note that we are matching a response of 0-12 patterns
        only, so we do not truly expect the curve to fit a simple Si depth
        -due to effect of energy thresholds and event splitting, but the Orsay 
        data are a pretty good match. On echange i made was to get a good fit 
        just below the Si K edge & there is put in a depth of 32um for ALL 
        chips

        Values used were: MOS1, CCD 1-7  25.8,29.6,28.8,25.3,28.4,28.2,27.8
                          MOS2, CCD 1-7  28.0,27.4,26.8,28.8,28.8,26.4,28.2
        Note that in MOS2 CCDS4&5 were never measured on ground so I took an 
        average value based on other CCDS from the same manufactured batch

VERSION 1.4 - 2002-01-30 (RDS)
-----------             
        Iss 0010     

        Add in extra units and remove quotes from region string

VERSION 1.3 - 2002-01-24 (RDS)
-----------             
        Iss 0010

        Add in extra pattern fraction tables (SCR-72)

VERSION 1.2 - 2002-01-08 (DL)
-----------             
        Iss 0009

        To fix SPR2178 I changed the region descriptor from POLYGON type to
        BOX type. This not only clears up ambiguity on the borders of the shape
        but probably speeds up the calculation.
        
VERSION 1.1 - 2001-09-20 (DL)
-----------             
        Iss 0008

        Updated for version 5.2 of SAS. We realise that there was a double
        accounting of the pattern fractions so that at 10keV (for exampel)
        the 50% fraction of 0-12 patterns was doubly applied. This led to
        a gradually decreasing QE with Energy compared with the true expected
        QE
        
VERSION 1.0 - 2001-07-19 (DL)
-----------     
        Updated for version 5.1 of SAS to make the SAS arfgen reproduce 
        consistent data with the LUX-supplied response matrices.
        Addedd additional energy anchor points at 0eV and 30keV to aid
        interpolation.
        High energy points are scaled according to a silicon absorption depth
        which was measured at Orsay - now we have a good cross-match to CCD 
        fabrication batch which allows an estimate even for thoise CCDS which 
        were never measured at Orsay.
VERSION 0.6 - 2000-09-6 (DL)
-----------
        Further analysis of fligh tdata at Licester allowed better 
        reconciliation with the analytical QE models. In addition we 
        introduce far more energy data to allow for better interpolation. There
        are still problems around the O edge and below 300eV, so we need to
        understand modifications in RMF to make progress at this point.
        
VERSION 0.5 - 2000-06-23 (DL)
-----------
        The initial flight calibration revealed discrepancies at the O edge
        and the lowest energy reponses. These were attributed mainly to 
        systematic  errors in the monitoring process at the Orsay synchrotron.
        New empirical fits to the data were provided by Leicester, and these 
        have been compared with new SCISIM runs, utilising the best latest 
        electrode structure used in the Leicester analystical QE model fit. A 
        combination of LUX data, and an extension BELOW the Si L edge for the 
        SCISIM data allows a hopefully better fit. In addition I have modified
        the QE's for the outer CCDs of each camera, where there are systematic
        differences in QE at > 5keV. These can be attributed to ddifferences
        in silicon resistivity translating to different depletion depths. The 
        QE data for this at Orsay is equivocal, so I take average trends in
        increased efficiency between 5 and 12keV and allocate a best estimate 
        of increased depletion depth to this, and in return modify the overall
        QE curves as a result. I am suspicious, though that all QE's are 
        GREATER tahn the central CCD, and this may indicate a systematic bias
        in the measurment method at Orsay.

        The new data sets should give good results to about 10% accuracy at 
        worst case small energy scales (eg EXAFS scales at silicon edges etc.)
        
VERSION 0.4 - 2000-05-23 (DL)
-----------
        FIxed error of missing pattern 1 data for CCD5
VERSION 0.3 - 2000-02-23 (DL)
-----------
        FIxed numerical problem in the data files (some values >>1 )
        
Version 0.2 - 2000-02-17 (UL)
-----------
        + fixed minor problem with REGION expression in data files

Version 0.1 - 2000-02-09 (DL)
-----------
        + created by DL
1	dlumb	1.6	ChangeLog for emos-quantumeff
2			============================
3	rsaxton	1.19	VERSION 1.12 - 2004-04-20 (RDS)
4			-----------
5			Issue 0016
6			Produced from a new calibration of MOS chips used on the
7			SWIFT project. The changes affect the QE_TOTAL array of
8			both cameras below 1 keV.
9
10	rsaxton	1.18	VERSION 1.11 - 2003-06-03 (RDS)
11			-----------
12			Issue 0015
13			Uses realistic channel space pattern fractions for Timing
14			mode (SPR-2422)
15			Includes key-words to describe the ratio of large pattern
16			events (SPR-2421)
17
18	rsaxton	1.17	VERSION 1.10 - 2002-12-16 (RDS)
19			-----------
20			Issue 0014
21			Set the pattern fractions in channel space more accurately so
22			that they can be used within epatplot
23
24	rsaxton	1.16	VERSION 1.9 - 2002-11-22 (RDS)
25			-----------
26			Issue 0013
27			Add in pattern fractions for timing and burst modes
28			Add in a CHBINS_FRACTION extension
29
30	rsaxton	1.15	VERSION 1.8 - 2002-06-28 (RDS)
31			-----------
32			Iss 0012
33			Return to the values in Version 1.6 which are now further
34			substantiated by work by Philippe Marty.
35
36	dlumb	1.14	VERSION 1.7 - 2002-03-27 (DHL)
37			-----------
38			Iss 0011
39			Seems like that SAS v 5.3 won't go out with the required
40			vignetting
41			corrections so we need a quick fudge to get the MOS2 and MOS1 QE's
42			back into an alignment that temporarily gives reasonable results,
43			while not maintaining the old Leicester QE's with known
44			deficiencies. MOS2 only changed..
45
46	dlumb	1.13	VERSION 1.6 - 2002-03-07 (DHL)
47			-----------
48			Iss 0011
49			Made an error in adding up only patterns 0:12 into the QE_TOT
50			extensions
51			So this version should now have patterns 0:31
52
53	dlumb	1.12	VERSION 1.5 - 2002-03-01 (DHL)
54	dlumb	1.13	-----------
55	dlumb	1.12	Iss 0011
56	dlumb	1.13	Fixed the data for the QE, by making the high energy (>1.5keV) points
57			match a detection depth provided on a CCD-CCD case from the Orsay data
58			(see below). At lower energies I just take the existing sets from the
59			Leicester response matrices (q20 versions) and merge into the higher
60			energy data. I note that we are matching a response of 0-12 patterns
61			only, so we do not truly expect the curve to fit a simple Si depth
62			-due to effect of energy thresholds and event splitting, but the Orsay
63			data are a pretty good match. On echange i made was to get a good fit
64			just below the Si K edge & there is put in a depth of 32um for ALL
65			chips
66
67			Values used were: MOS1, CCD 1-7 25.8,29.6,28.8,25.3,28.4,28.2,27.8
68			MOS2, CCD 1-7 28.0,27.4,26.8,28.8,28.8,26.4,28.2
69			Note that in MOS2 CCDS4&5 were never measured on ground so I took an
70			average value based on other CCDS from the same manufactured batch
71	dlumb	1.12
72	rsaxton	1.11	VERSION 1.4 - 2002-01-30 (RDS)
73			-----------
74			Iss 0010
75
76			Add in extra units and remove quotes from region string
77
78	rsaxton	1.10	VERSION 1.3 - 2002-01-24 (RDS)
79			-----------
80			Iss 0010
81
82			Add in extra pattern fraction tables (SCR-72)
83
84	dlumb	1.9	VERSION 1.2 - 2002-01-08 (DL)
85			-----------
86			Iss 0009
87
88			To fix SPR2178 I changed the region descriptor from POLYGON type to
89			BOX type. This not only clears up ambiguity on the borders of the shape
90			but probably speeds up the calculation.
91
92	dlumb	1.8	VERSION 1.1 - 2001-09-20 (DL)
93			-----------
94			Iss 0008
95
96			Updated for version 5.2 of SAS. We realise that there was a double
97			accounting of the pattern fractions so that at 10keV (for exampel)
98			the 50% fraction of 0-12 patterns was doubly applied. This led to
99			a gradually decreasing QE with Energy compared with the true expected
100			QE
101
102	dlumb	1.7	VERSION 1.0 - 2001-07-19 (DL)
103			-----------
104			Updated for version 5.1 of SAS to make the SAS arfgen reproduce
105			consistent data with the LUX-supplied response matrices.
106			Addedd additional energy anchor points at 0eV and 30keV to aid
107			interpolation.
108			High energy points are scaled according to a silicon absorption depth
109			which was measured at Orsay - now we have a good cross-match to CCD
110			fabrication batch which allows an estimate even for thoise CCDS which
111			were never measured at Orsay.
112	dlumb	1.6	VERSION 0.6 - 2000-09-6 (DL)
113			-----------
114			Further analysis of fligh tdata at Licester allowed better
115			reconciliation with the analytical QE models. In addition we
116			introduce far more energy data to allow for better interpolation. There
117			are still problems around the O edge and below 300eV, so we need to
118			understand modifications in RMF to make progress at this point.
119
120	dlumb	1.5	VERSION 0.5 - 2000-06-23 (DL)
121			-----------
122			The initial flight calibration revealed discrepancies at the O edge
123			and the lowest energy reponses. These were attributed mainly to
124			systematic errors in the monitoring process at the Orsay synchrotron.
125			New empirical fits to the data were provided by Leicester, and these
126			have been compared with new SCISIM runs, utilising the best latest
127			electrode structure used in the Leicester analystical QE model fit. A
128			combination of LUX data, and an extension BELOW the Si L edge for the
129			SCISIM data allows a hopefully better fit. In addition I have modified
130			the QE's for the outer CCDs of each camera, where there are systematic
131			differences in QE at > 5keV. These can be attributed to ddifferences
132			in silicon resistivity translating to different depletion depths. The
133			QE data for this at Orsay is equivocal, so I take average trends in
134			increased efficiency between 5 and 12keV and allocate a best estimate
135			of increased depletion depth to this, and in return modify the overall
136			QE curves as a result. I am suspicious, though that all QE's are
137			GREATER tahn the central CCD, and this may indicate a systematic bias
138			in the measurment method at Orsay.
139
140			The new data sets should give good results to about 10% accuracy at
141			worst case small energy scales (eg EXAFS scales at silicon edges etc.)
142
143	dlumb	1.4	VERSION 0.4 - 2000-05-23 (DL)
144			-----------
145			FIxed error of missing pattern 1 data for CCD5
146	dlumb	1.3	VERSION 0.3 - 2000-02-23 (DL)
147			-----------
148			FIxed numerical problem in the data files (some values >>1 )
149
150	ulammers	1.2	Version 0.2 - 2000-02-17 (UL)
151			-----------
152			+ fixed minor problem with REGION expression in data files
153	ulammers	1.1
154	ulammers	1.2	Version 0.1 - 2000-02-09 (DL)
155	ulammers	1.1	-----------
156			+ created by DL