MARS EXPRESS SPICAM DATA Calibration =========================================================================== Revisions -------------------------------------------------------- 2005 01 21 V016 separate UV and IR reberac 2005 01 21 V015 add annexes dimarellis 2005 01 13 V014 update wavelength assignment dimarellis 2004 10 13 V013 first issue dimarellis Purpose This document describes the calibration of the Spicam UV data which are delivered in the Spicam UV data product. Introduction =========================================================================== All informations needed to calibrate the UV data file are given. This document is organized as: Introduction UV Data Time Temperatures Dark current Occultation mode Others modes Wavelength assignment Star Nadir/Sun/Limb Photometry Annexes: others documents on calibration =========================================================================== UV DATA =========================================================================== Time: =========================================================================== The exact timing of UV data is the following: start of Exposure time = UTC header time - 1 s + 126 ms (-1) because the picture was taken in the previous second (+126) processing in DPU, after data acquisition, before time tag The time of data can be defined as (middle of exposure time): Time of data = Start of Exposure time + Exposure Time / 2 Temperatures: =========================================================================== In the data file, 2 temperatures are recorded (in header) - temperature of cold face of Peltier (= CCD) - temperature of hot face of Peltier The temperature of CCD is supposed to be the same as the cold face. The thermistors used to detect temperature are: YSI 44003A, 0.2%, 1000 Ohms at 25 deg C inserted in parallel with 50 Kohms plus 1 Kohm in serial under 5 volts Table below gives the temperature (deg C) versus the level Temp Value -30 242 -25 239 -22 237 -20 236 -18 235 -15 232 -10 228 -5 224 0 219 5 213 10 208 15 202 20 196 25 190 30 185 35 179 40 174 45 170 50 165 55 161 60 158 65 155 70 152 Note: These temperatures are not used to evaluate dark current. Dark current: =========================================================================== The measurement in a CCD pixel is given by: M = S + D M = signal D = dark current, which has 2 contributions D = d1 + d2 d1 = dark current from photocathode of intensifier d2 = dark current from CCD Depending on the operating mode of the instrument, there are several ways to evaluate the dark current. Occultation Mode: ----------------- This mode is used with Stars or SUN. In this mode, as the target is at one time fully occulted, there is data without signal from target. These data are the dark current D. Procedure: Determine data where there is no signal (beginning or end of data). Average a few (10, 20...) lines in order to decrease the Poisson noise on dark current. Substract this value to the data where there is signal from target. This procedure is done for each pixel in order to take into account the pixel non uniformity of the dark current. It is the best way to correct data for dark current. Others Modes: ------------- These modes are used in NADIR or LIMB. In these modes, the target is never occulted and so there is no data without signal. In fact, in some data, for example when looking at the dark side (without Sunlight), we may suppose that there is no significant signal from the target and so the previous procedure must be used. If there is no data without signal from target, we have to evaluate dark current with another method. Each component of the dark current D is computed separately. D = d1 (intensifier) + d2 (CCD) CCD (d2): This value is computed by averaging the masked pixels (397 to 406, first pixel is 1), for each line of the CCD. Intensifier (d1): It is more difficult to evaluate the dark current coming from the intensifier, because there is no masked photocathode. The intensifier dark current is computed as follow: with data from occutation mode compute data without signal (see procedure) compute average of CCD dark with masked pixels substract CCD dark from data to get d1 We have the following relation: D = 1.07 * d2 Comments on Dark from intensifier: is very low High voltage level has no (little) influence Then the computed value is subtracted to each pixel. Note that in this case, the pixel non uniformity is ignored. Wavelength assignment: =========================================================================== In the spectrometer (grating), the spectrum recorded on the CCD can be shifted depending on the position of the source at the entrance of the spectrometer. Depending on the operating mode, the source is either the slit, or the star without slit. The wavelength assignment will not be the same. Star: ----- In this mode there is no slit. The position of the star in the entrance field of view of the spectrometer is not well known, because it depends: of the pointing direction (coordinates given to SC) of the relative alignment beteween Spicam and SC The wavelength assignment is done with the following procedure: - In the spectrum of the Star, determine the position of the Lyman Alpha absorption line which is strong and above pixel 350. - Compute the wavelength with a dispersion of 0.527 as for others modes. - Or, with the known calibrated spectrum of the Star, determine the pixel position of another strong absorption line. The spectrum of the Star can be retrieved from Astronomical database. Nadir/Limb/Sun: --------------- In these modes there is a slit at fixed position. From data taken in dark side of the planet, with the slit, the Lyman Alpha position is pixel 367 (first pixel is one). The grating dispersion is about 0.527 nm /pixel (from ground cal) The wavelength is given by (from JLB 2005 01 13): wavelength (nm) = 322.17 -0.54732*pixel number (from 0 to 407) Photometry: =========================================================================== The photometric calibration is fully described in a dedicated document written by J.L. Bertaux. ===> See Annexes Below is only an overview of calibration pinciples. Then we need to compute the relation between the number of DN (digital number) recorded in each pixel and the numbers of photons received. This relation is function of: wavelength high voltage level of the intensifier The relation is, for each pixel: Photons = S(lambda) * DN S is efficient (or equivalent) surface taking into account all instrument characteristics (mirror surface, transmissions, quantum efficiency, CCD gain...) DN is the pixel content (corrected from dark current) There is a very simple way to compute S with Star observations. The ratio between a calibrated spectrum (from Astronomical database) and the spectrum (in DN) recorded outside atmosphere is S(lambda). This is in flight calibration. This calibration is done each time there is a Star observation and allows us to follow the sensitivity of the instrument. Once the S(lambda) is computed, it can be used in the others modes to compute the radiance of the target. To do that we need to know the gain of the intensifier The gain of the MCP (Micro Channel Plate) of the image intensifier may be adjusted by TC (telecomand) with a HV level from 500 to 900 volts, commanded by a digital level HT from 0 to 255. When a photo-electron is created in the photocathode (a photo-event), it results in a pulse of light from the phosphor, distributed over a few pixels of the CCD. It is detected by the CCD reading electronics by a number of ADUs (Analog to digital Units). At HT = 20 ( a usual low gain level necessary to avoid saturation for dayside nadir observations), there are about 2 ADU created per photo-event, while for HT = 200, there are 40 ADU per photo-event. There is a potential danger to activate the intensifier at a large gain on an intense source of light. Therefore, this high gain has been used cautiously, on the day side but at high altitude, and then on the night side, when the pericenter has shifted to the night side. One constraint of SPICAM operation is that during one observing period (ON - OFF), the parameters of the instrument may not be changed. In particular, when a fixed inertial attitude of the spacecraft is chosen, then one must be sure that the signal will not be too strong to damage the intensified detector, if one uses a high gain on a bright target. The relation between the level of HT (between 0 - 255) and the gain is given below: Gain = exp(7.46113 * ln(500 + 1.57*HT) - 46.3864) HT Gain 0 0 1 1.0 10 1.2 20 1.5 40 2.4 60 3.6 80 5.2 200 37.3 Annexes: =========================================================================== File : SPICAM_UVT31DOC.PDF in DOCUMENT directory The calibration factors of the Spicam UV spectrometer channel. Note describing the calibration process (JL Bertaux) File : SPICAM_UVSEFF.DAT in CALIB directory ASCII file given the Seff as a fonction of wavelength.