MARS EXPRESS SPICAM GEOMETRY Computation =========================================================================== Revisions -------------------------------------------------------- 2005 09 15 V014 update text and wording, correct wide slit axis dimarellis 2005 02 01 V013 update LOF, dimarellis 2005 01 26 V012 add acronyms, minor corrections, dimarellis 2005 01 06 V011 corrections... dimarellis 2004 12 15 V010 initial, dimarellis Purpose This document describes the geometry computation. This document is organized as: Introduction Definitions Preparation of observations Computation of what is seen IR SUN case List of acronyms: =========================================================================== UV1 UV channel aperture for Star and Nadir UV2 UV channel aperture for SUN IR1 IR channel aperture for Star and Nadir IR2 IR channel aperture for SUN UV1star UV channel aperture, NO slit UV1nstar UV channel aperture, Narrow slit UV1wstar UV channel aperture, Wide slit UV1nad UV channel aperture, full slit UV2sun UV channel SUN aperture, full slit LOS Line of sight SLITPOS Position of center of slit in pixels COB UV, center of each band in pixels LOF UV, Line of sight of center of each band LOS1 UV1, Line of sight (SC axes) of center of slit LOS2 UV2, Line of sight (SC axes) of center of slit, SUN case PL Pointing direction of the SC MREQ Mira request file for observation planning Introduction: =========================================================================== Spicam has 4 apertures: UV1 near +ZSC axis, Star, Nadir, Limb UV2 in XY plane, SUN, around 30 deg from XSC IR1 near +ZSC axis, Nadir, Limb IR2 in XY plane, SUN, around 30 deg from XSC The frame attached to SC (X,Y,ZSC) is defined as: Za center of SC to Nadir face (velocity vector at launch) Xa center of SC to main SC antenna Ya completes right handed frame (solar panel) UV data: -------- In addition, at the entrance of the spectrometer, we may have (or not) a slit. Along the slit, the width is not constant and we have a narrow part and a wide part of the slit. The data obtained in all modes are 5 bands on a CCD in the row axis. The positions and the widths of the bands are somewhat flexible. The slit is oriented in the column axis of the CCD, parrallel to XSC axis. The data of each band is typically the spectrum of a part of the slit. This can be the narrow part or the wide part, depending on the position of each band on the CCD. For UV, there is a full CCD field of view, and each pixel has its own elementary field of view. The full CCD is 408 x 290 pixels IR data: -------- For IR, there are only 2 pixels, one for each polarization. The field of view of each pixel is supposed to be the same. Due to the large value (1 deg) of the IR field of view (see below) compared to one UV pixel, we consider that IR is aligned with UV, and so, no specific configurations. The special cas SUN IR2 will be treated separately (see below) For UV or IR channels, there are 2 steps totally independant: 1. Preparation of observations: In this step we have to define the SC pointing in order to get data in the detector UV (at a specified row) and/or IR 2. Computation of what is seen in each detector: Definitions: =========================================================================== Due to the great number of operation modes and instrument configurations, we need to have a full description for each case of observation. Of course the slit must be used for each extended source like Nadir, Limb or Sun observations. For Star observation, there is no need for the slit, because the Star is a pointlike source surrounded by the dark sky at the entrance of the spectrometer. Nevertheless, in STAR mode, the slit can be used if the "dark sky" is not dark enough (for example, if there is bright sources not far from the star, like a bright limb). For each mode, depending on the slit, we define the configurations: STAR No slit UV1star nominal mode Narrow slit UV1nstar a few times Wide slit UV1wstar a few times In star mode, the configurations with slit are defined only for preparation of observations. These configurations are not useful in the computation step (see next paragraph). In Nadir or Limb, the slit is ALWAYS used. There is no use to do an observation of an extended source without slit, because in this case, each point of the full field of view will produces a spectrum, and then on the CCD, we will get all spectra mixed. NADIR Slit UV1nad nominal mode /LIMB FIX Slit UV1nad nominal mode as virtual star with slit SUN Slit UV2sun nominal mode For the preparation of observations we have to define: ------------------------------------------------------ -the pointing direction of the SC called pointing line PL -the CCD lines to be read where data are supposed to be found -the full orientation of the SC(around the PL), in some cases (not always) For the computation of what is seen: ------------------------------------ What must be computed? There are two cases to take in consideration, depending on the target: A) Extended sources, with SLIT: (NADIR, LIMB, SUN, FIX, ...) -the main line of sight (LOS) which is the direction of the center of the SLIT at CCD line 144 (SLITPOS) in SC axes. -the line of center of field of view (LOF) which is the direction of the point of the slit which produces a spectrum on a specific band, in SC axes. Each band is defined by center (COB in pixels)) and extension defined in elementary pixels. There is a LOF for each band. B) Pointlike sources with/without SLIT: (STAR) -the main line of sight (LOS) which is fully defined by the Spacecraft position and the direction of the Star. Conclusions: 1) For case B, we need to know what is the Star direction (which is not in orbit/attitude file). 2) For case A, we need to know where is the slit in S/C axes. We call LOS1 the direction of the LOS in S/C axes. This value has to be determined either at ground by alignment measurements or in flight (much better). To get this value in flight: Do an observation with Star and find Lyman pixel, Do an observation with Slit and find Lyman pixel, Deduces LOS1 (or 2) in S/C axes. There is a different value for UV1nad and UV2Sun namely LOS2. The ray coming from SUN is reflected by a small miror and then is parallel to LOS1. 3) We may want to compute not only the main direction, but also extension of various fields of view (for each band, full CCD...). For this, the orientation of the S/C is needed (attitude file). These FOV are fixed in the CCD frame, and so, they will be defined in pixels, relative to SLITPOS. Instrument values: ------------------ POINTING: --------- PL (pointing direction) is defined by 2 sets of 2 angles: (used in the MREQ planning file) alpha, delta defined the target direction alpha right ascension in J2000 delta declination in J2000 phi, theta defined in "attitude frame" attached to the SC (X,Y,ZSC) Za center of SC to Nadir face (velocity vector at launch) Xa center of SC to main SC antenna Ya completes right handed frame (solar panel) phi angle of projection of PL in XY plane, start at X theta angle of PL above XY plane /|\ Z | | / | / | . | / . theta |/ . *-------------> Y / \ . / \ . / . /. . X |/ . . . \ phi phi, theta are named "offsets" FIELD OF VIEWS: --------------- IR pixel field of view is 1 deg, much larger than the UV pixel FOV Due to this large FOV, the IR line of sight is LOS. Below we will discuss only UV definitions. UV Pixel field of view: The FOV is defined at focus of parabolic mirror at the slit. The slit is 50 microns wide in its narrow part and 500 microns in its large part. The grating 'optical enlargment' is 1. Then the slit size may also be defined in pixel units, and positions in the slit plane may be expressed with virtual pixels. pixel size is 23 micrometres focal length of parabolic mirror is 120 mmn angle of one pixel is 0.023/120 (in radian) Solid angle of one CCD line through the slit 0.050*0.023/120 = 7.986E-08 steradian POSITIONS: ---------- Positions on the CCD are defined in pixels coordinates: For UV1 aperture: full CCD row along YSC pixel 1 to 408 (lambda decreasing) column along XSC pixel 1 to 290 slit along CCD column narrow slit has a width of about 2 pixels center column number is 204 wide slit start at row 200, width is about 20 pixels position along slit is Line (or row) number (variable) SLITPOS is 204, 144 (column, row in pixel coordinates) ORIENTATION of the slit: (after final analysis) The orientation of the slit is such that the wide part is towards -XSC For UV2 aperture: We have of course the same definitions as UV1 in SC axes, but the incident ray must be computed after reflection on the SUN mirror. For the 5 bands: We need to know the position and the width of each band These values are in Telemetry Data: Y0 1st row of the 1st band bin binning, number of row for each band same for all bands UNLESS "progressive binning" (will be treated separately) We have COBx, x = 1 to 5, center of band x COBx = (Y0 + (x-1)*bin) + bin/2 Fields of view: In addition to COBx, we define several FOV. Each FOV is defined as: rectangular, 4 corner points, Bottom Left, Bottom Right, Top Right, Top Left In pixel coordinates, column, row numbers For bands, we have FOV1 to FOV5 In addition: FOV6 5 bands together FOV7 full narrow slit FOV8 full wide slit FOV9 full slit FOV10 full CCD First we define points of interest on CCD (corners of bands, of CCD,...) ; PTS Numbering (from 1) ; 13------14-------11------12 full CCD ; ; 17 8 ; ; 15 16 9 10 Narrow/Wide slit ; ; 18 7 ; 19 6 ; 20 5 Bands ; 21 4 ; 22 3 ; ; 25---------23 2 --------1 Then a FOV is defined as a list of points, 4 points in general, but in some case we may have 8 points. This allows to have complicated FOV, not only rectangle. (see Geoinit and Geospicam for a full description) Preparation of observations: =========================================================================== For each identified configuration, we have the following: Line is the center of the band 3 Config alpha delta phi theta Line comment UV1star ra dec 90.00 89.83 144 (1) UV1nstar ra dec 90.00 89.83 144 (1, 2) UV1wstar ra dec TBD TBD 240 (1, 3)) UV1nad NA NA 0.00 90.00 144 (4) UV2sun ra dec 30.22 -0.07 102 (5) (1) ra, dec are star coordinates in J2000 Variation in theta induces a shift in the wavelength assignment (spectrum is shifted but is on the same row if phi = 90 deg) Variation in phi induces a shift in the row number (line) With the present values of phi, theta, Lyman Alpha is at pix 367 Several observations were done with theta = 89.70 at the beginning of the mission. Lyman Alpha was NOT on pix 367. A few observations were done with 89.93 (reason unknown) ra, dec, phi, tetha MUST be specified in MREQ. (2) A few observations were done with slit (and star). As the star was seen through the slit, these values give the direction of the line of sight of the slit at line 144 which is the LOS in NAD mode. (3) a few observations without success were done with (9.28 88.95). These values are not correct (bad orientation of X axis) (4) In Nadir, the Z axis is continuously pointed towards the center of Mars, phi, theta have no meaning. (5) phi, theta define the direction of Sun aperture which is in the SC XY plane at around 30 deg of X. Several observations were done with others values, but without data on IR channel. With (30.57, -0.49) the SUN was centered on line 144. This is LOS in SUN The (30.22, -0.07) values are supposed to give data on IR and UV at the same time (observation are OK). See paragraph on IR2SUN. Computation of what is seen: =========================================================================== In summary we have in S/C axes: For UV1 LOS1 = (90.00, 89.83) corresponds to SLITPOS For UV2 LOS2 = (30.57, -0.49) corresponds to SLITPOS AFTER reflection on the SUN mirror We supposed that we have: Mexpos MEX position file Mexatt MEX attitude file We use Naif/Spicelib for all computations (with Kernels files). The general scheme of computation is: INPUT parameters TIME on CCD, position of 1st band binning TARGET (Star, Sun, Others...) FIRST compute LOS in IAU_MARS (body-fixed frame) compute derived parameters Then COBx in CCD frame LOFx in S/C axes in IAU_MARS Then for each corner of FOVx same as LOFx Depending on the observation target we have: STAR with Mexpos with Star direction ---> compute LOS with Mexatt with LOS1 ---> compute COBx, LOFx, FOVx NADIR, LIMB,... with Mexpos with Mexatt with LOS1 ---> compute LOS1 ---> compute COBx, LOFx, FOVx SUN special case of NADIR use LOS2 instead of LOS1 ---> compute LOS2 ---> compute COBx ---> reflection on mirror ---> compute LOFx, FOVx In Summary we have: Target Pixels SC axes Mars axes CCD axes ------ -------- ------- --------- NAD SLITPOS LOS1 MLOS COBx LOFx MLOFx FOVx ... SUN SLITPOS LOS2 MLOS COBx LOFx (*) MLOFx (after reflection) Notes: 1) For STAR, the interesting parameter is LOF3 (the band where is the spectrum of the STAR). As this band was centered at CCD 144 we have LOF3 = LOS1 if the pointing is perfect. This is true for all observations except the first one (orbit 17). 2) Other than STAR For Nadir observations, in general (bin = 4, Y0 = 135), and we have also LOF3 = LOS2. But others observations were done with others values and then, LOF3 is not LOS. This is also true for all Limb observations. In NADIR, due to NOT inertial direction of pointing, the FOVx are not the same at the beginning and at the end of the exposure time. More, we may have FOVx which overlap one another if X axis is not perpendicular to orbital plane. In Limb mode, although we are in inertial pointing, due to the motion along the orbit, the FOV are different relative to MARS. 3) General Geometrical parameters: All geometrical parameters will be related to LOF3, center of the third band. 4) Visualisation of FOV: In order to have an idea of what is seen, the FOVx will be projected on the plane perpendicular to the direction between MEX and Mars Nearest Point (MNP) of LOS, at MNP. MNP is either intersection with Mars surface or impact parameter (the point of LOS which is nearest of Mars if no intersection). What to do if impact parameter is behind Mars relative to Line of sight? IR SUN case: =========================================================================== In contrary of the IR1, the IR2 is very small. The entrance optics produces an image of the SUN which is 400 microns on an optic fiber which is 200 microns. Furthermore, there is a circular diaphram of 100 microns which limits the FOV. As the SUN is around 20 arcmin (seen from Mars) we need a pointing with an accuracy of 10 arcmin or less otherwise we have nothing in the fiber. On ground, the IR2 and UV2 were co-aligned, but in such a way that the image of the SUN was at Line 102 on the CCD (and not 144). I forgot this adjustment and the first tries on SUN were done with SUN on Line 144, and of course this explain why there was no signal on IR! The precise misalignment between UV and IR is unknown, and so, we do not know which part of the SUN is pointed by IR channel. POSITIONS: ---------- With the nominal SUN pointing, we have signals on UV and IR. For IR, we consider that the Line of view is the SAME as the 3 rd band of the UV channel. The IR computations are done as UV with an offset of 55 pixels in +XSC (3 rd band of UV) relative to center of CCD.