MARS EXPRESS SPICAM DATA Calibration
===========================================================================
Revisions
--------------------------------------------------------
2005 11 24 new calibration process - from A. fedorova, 05/05/2005
update of the "dark current" section.
A new method to determine the dark current is presented.
The calibration process is more complicate than the first one
but correction for DC channel 2 is more accurate.
update of the "PSF" section (one file for each channel)
update of the "Absolute Units" section files
2005 01 28 corrections (content file description + possible calibration of
channel 1)- Anna Fedorova
2005 01 21 separate UV and IR description - reberac
2005 01 06 V014 add infrared calibration information
from Fedorova-Kiselev issue 2004 11 02
2004 10 13 V013 first issue dimarellis
Purpose
This document describes the calibration of the Spicam IR data which
are delivered in the Spicam IR data set.
Introduction
===========================================================================
All informations needed to calibrate the IR data file are given.
This document is organized as:
Introduction
IR Data
Time
System Monitor
Command
Dark current
Wavelength assignment
Solar Spectrum
Absolute Units
PSF
All files mentioned in this document are in the CALIB directory.
===========================================================================
IR DATA
===========================================================================
Time:
===========================================================================
The exact timing of IR data is the following:
TBD
System Monitor:
===========================================================================
In the data file "system monitor" information, that needs for calibration
is saved for each measurement cycle.
- detector 0 temperature in Volts,
- detector 1 temperature in Volts,
- AOTF temperature in K,
- RF power at 110 MHz (the middle of AOTF frequency range) in Volts,
- base plate temperature in K,
- supply voltage control in Volts
Command:
===========================================================================
17 command parameters
EXIT, SOURCE, DETS, TIME, GAIN, DAC,
WINDOW 0, WINDOW 1, WINDOW 2 parameters:
for each windows (base-frequency, number of points, step)
COMMAND DESCRIPTOR, DOTS DESCRIPTOR
DETS - Detectors used for spectrum measurement: 0 -detector 0 only,
1 -detector 1 only, 2 -both detector 0 and detector 1,
3 -detector 0 and AOTF RF power.
TIME - AOTF chopping period: 0 -1.4ms, 1 -2.8ms, 2 -5.6ms, 3 -11.2ms.
GAIN - Amplifiers gain factor: 0 -1, 1 -3, 2 -8.25, 3 -26 (design values).
DAC - AOTF RF power control: 0-255 (8 MSBs of AOTF 12-bit DAC value, 4 LSBs
of DAC value are set to zero: DAC value = 16 x RF power control).
Dark current:
===========================================================================
The measurement in a IR channel is given by:
M = S + D
M = measurement
S = signal
D = dark current
S = M - D
The current has a temperature (or time) and frequency dependance
The first method takes into account the temperature and frequency dependances,
while the second method is time,temperature and frequency dependent.
Depending on the command set, there are several ways to evaluate the dark
current.
Method 1 :
---------
The dark current was determined for the main commands :
1) DAC=1744, GAIN = 8.25, chopping period = 5.6 ms - main mode of nadir
observation.
Because of many measurements of dark current in flight the dependance
is represented by quadratic polynom approximation:
D(F)=a(F)*X2+b(F)*X+c(F) where D signal in ADU/gain factor,
X temperature of detector in Volts
F frequency in MHz
We found a, b and c coefficients for sequence of frequency from 84 to 147
MHz with step 0.5 MHz.
File TOK_COEF1744_825.TXT there are seven columns in the file for 3
coefficients for each detector (F(MHz),a0,b0,c0,a1,b1,c1 - index 0
for detector 0, 1 - for detector 1).
2) DAC=1504, GAIN = 3.0, chopping period = 5.6 ms - nadir and limb
observation Due to small amount of points it was possible to make just a
linear approximation.
D(F)=a(F)*X+b(F) where D signal in ADU/gain factor,
X temperature of detector in Volts.
We found a and b coefficients for sequence of frequency from 84 to 147 MHz
with step 0.5 MHz.
File TOK_COEF1504_ORB.TXT there are five columns in the file for 2
coefficient
for each detector (F(MHz),a0, b0, a1, b1 - index 0
for detector 0, 1 - for detector 1).
3) DAC=1744, GAIN = 3.0 and 1.0, chopping period = 2.8 ms
There is no temperature dependance of dark current (no observations)
The dark current is presented by D(F) dependance.
File DARK_1774_3_28.TXT there are three columns in the file for dark
current of detector (first column is F - frequency in
MHz, second and third columns - D signal in ADU for CH0 and CH1
respectively)
S(F) = M(F) - D(F)
Method 2 :
---------
The dark current was determined for the main commands :
1) DAC=1744, chopping period = 5.6 ms - main mode of nadir, limb and star
occultation.
a) nadir, phobos and limb (just when gain=8.25)
The dependance is represented by a natural logarithm approximation:
D(F)=a(F)*ln(time+b(F))+c(F) where D signal in ADU (gain factor 8.25),
time - in second - beginning of spectrum
F frequency in MHz
For gain correction it needs to multiply by coefficient gain(for current
orbit)/8.25
File TOK_COEF1744_56_825.TXT there are seven columns in the file for three
coefficients for each detectors :
F(MHz),a0,b0,c0,a1,b1,c1
index 0 for detector 0, 1 - for detector 1).
For temperature correction use shift coefficient:
File TEMP_DEP_SHIFT1744_825.TXT there are five columns in the file for 2
coefficients for each detectors (F(MHz),shia0,shia1,shib0,shib1
- index 0 for detector 0, 1 - for detector 1).
shift(F)=shia0(F)*mean(det0temp(1))+shib0(F)
D(F)=D(F)+shift(F)
b) star occultation and limb (just when gain=3)
The dependance is represented by a natural logarithm approximation:
D(F)=a(F)*ln(time+b(F))+c(F) where D signal in ADU (gain factor 3),
time - in second - beginning of spectrum
F frequency in MHz
for gain correction it needs to multiply by gain(for current orbit)/3.0
File TOK_COEF1744_56_3.TXT there are seven columns in the file for 3
coefficients for each detectors (F(MHz),a0,b0,c0,a1,b1,c1
- index 0 for detector 0, 1 - for detector 1).
For temperature correction use shift coefficient:
File TEMP_DEP_SHIFT1744_3.TXT there are five columns in the file for 2
coefficients for each detectors (F(MHz),shia0,shia1,shib0,shib1
- index 0 for detector 0, 1 - for detector 1).
shift(F)=shia0(F)*mean(det0temp(1))+shib0(F)
D(F)=D(F)+shift(F)
2) DAC=1504, GAIN = 3.0, chopping period = 5.6 ms - nadir and
limb observation .The dependance is represented by a natural logarithm
approximation:
D(F)=a(F)*ln(time+b(F))+c(F) where D signal in ADU (gain factor 3),
time - in second - beginning of spectrum
F frequency in MHz
for gain correction it needs to multiply by gain(for current orbit)/3.0
File TOK_COEF1504_56_3.TXT there are seven columns in the file for 3
coefficients for each detectors (F(MHz),a0,b0,c0,a1,b1,c1
- index 0 for detector 0, 1 - for detector 1).
A temperature dependance wasn't found
3) DAC=1744, GAIN = 3.0 , chopping period = 2.8 ms
The dependance is represented by a natural logarithm approximation:
D(F)=a(F)*ln(time+b(F))+c(F) where D signal in ADU (gain factor 3),
time - in second - beginning of spectrum
F frequency in MHz
for gain correction it needs to multiply by gain(for current orbit)/3.0
File TOK_COEF1744_28_3.TXT there are seven columns in the file for 3
coefficients for each detectors (F(MHz),a0,b0,c0,a1,b1,c1
- index 0 for detector 0, 1 - for detector 1).
A temperature dependance wasn't found
S(F) = M(F) - D(F)
Wavelength assignment:
===========================================================================
The Wavelength-RF dependance
F(frequency, kHz), lam(wavelength, nm),t(temperature, C)
Channel 0: (A.Kiselev)
lam= 136700000/F-0.0000000000653*F**2+74.43+0.0285*t+0.0001*t**2
The accurancy of such calibration is about � 0.2 nm (within 1100 - 1500 nm).
Channel 1:
lam= a/F+b
a=-3.6228649*t**2+2464.6217*t+13690971
b=-5.4920304e-6*t**2+4.4824233e-3*t+71.220396
The accurancy of such calibration is about � 0.3 nm (within 1100 - 1500 nm).
Solar spectrum:
===========================================================================
file "KURUCZ_1101.TXT" - is a solar spectrum converted with SPICAM PSF (Point
Spread Function)
wavelength, mn - Solar radiance, W/m2/ster/microns
Absolute units:
===========================================================================
For channel 0:
1) DAC=1744 S(lam)=ADU(lam)/GAIN/ck_ch0(lam) 'CKF_1744_28_CH0.TXT'
2) DAC=1504 S(lam)=ADU(lam)/GAIN/ck_ch0(lam) 'CKF_1504_56_CH0.TXT'
ck_ch0(lam) - calibration coefficients in ADU/(W/m2/mkm/sr)
In files: first column - lam(nm), second column - ck_ch0(lam)
Possible calibration of channel 1:
ck_ch1(lam)=ck_ch0(lam)/coef(lam) - calibration coefficients for channel 1
in ADU/(W/m2/mkm/sr)
file 'CKF_CH1.TXT' (lam(nm), coef(lam))
PSF
===========================================================================
A. Kiselev: Y = -10 + A*(SIN(X))^2/X^2*(1+0.08*X) , X = (F0-F)/21.0
where F - frequency [kHz]
A - amplitude.
Dependence F(lam) could be found from the Wavelength-RF dependance
file PFS.TXT (more accurate PSF) (first column - F, second column - PSF(F))
Update:
one file for each channel: files PFS_CH0.TXT and PFS_CH1.TXT