POA: Revised Documentation for POA_CALFOS
Revisions to the handbook to accompany the final
release are currently being prepared.
Here is the text from the HST Data Hand Book (pages 31-16 ff
from Chapter 31: Calibrating and Recalibrating FOS Data)
relating to the correction of the GIM problem revised so as to be
consistent with 'poa_calfos' . Future versions of the hand book will
contain this material.
31.4.5 Correction for image motions (OFF_CORR)
As with all spectrographs the ability to obtain correct
wavelengths and clean line profiles is critically dependent on the
meachanical, optical and electronic stability of the entire
system. Several precautions where taken in the design and operation,
but nevertheless the effects of image motions of various kinds
affected almost all data obtained. On top of that, contributions by
the sources discussed below are approximately of the same size, so
that a coherent solution could only be found post-operational in the
comprehensive model built into 'poa_calfos' and based on the analysis
of the entire FOS archive.
The detector related image motions stem from the electron-optical
design. The FOS Digicons accelerated single photo-electrons released
in the photocathode and imaged them onto a 1-dimensional array of
diodes where they were counted. The electron optics consisted of the
accelerating E-field of 20 kV and a parallel (longitudinal = Z)
B-field of 105 Gauss. The strip on the photocathode to be imaged onto
the diode array could be chosen by application of currents to crossed
pairs of large coils that added B-fields in X,Y direction (order a few
Gauss). The whole structure was embedded in a tube made of "mu-metal"
to provide shielding against external magnetic fields.
- GIMP: Early on during SV in 1990 it was detected that the
geomagnetic field (order 0.5 Gauss) penetrated through the magnetic
shield of the red side detector at considerable strength (attenuation
only about factor 10). As a result, the location of the electron beam
moved about on the diode array along small ellipses (diameters of up
to 1 diode = 4 pixel), orientation and dimensions depending on orbital
path and attitude (pointing) of HST. Whether or not the blue side
detector suffered as well from this geomagnetic image motion problem
(GIMP) remained unclear at the time, but a factor 7 smaller effect was
found to match the sparse data. The full effect of this is seen in all data obtained prior to 3-Apr-1993.
- OBGIMP: On 5-April-1993 an algorithm
was implemented in the FOS on-board computer using the X,Y deflection
system to compensate for GIMP in 15 second increments (see ISR CAL/FOS 66). POA analysis has
verified that this compensation was suboptimal for the red side
(amplitude 4 pix) and about a factor 10 to large for the blue detector
(amplitude .5 pixel). All data obtained after 5-April-1993 need a
compensation for OBGIMP.
- YBOFF: Because the X,Y deflection system used magnetic fields, the
ExB electron drift (Lorentz-force) moved the electron beam in X-direction even if only a Y-direction field was applied. This effect
went unnoticed and uncompensated for, so that the regular updating of
YBASE values (i.e. the definition of locations of spectra on the
photocathode) resulted in a stepwise function of X-direction offsets in
time. Data obtained late in the FOS life suffer the most, with offsets
of up to 6 pixels in late 1996. All data need this YBOFF
The optical bench of FOS was manufactured from carbon rods and invar
joints, and therefore pretty insensitive to thermal effects. However,
the wheel carrying the focusing gratings used by both, the red and
blue side trains respectively, and its stepper motor drive have always
been a matter of concern even before launch. The usual explanation
(also carried into the scientific literature) of seemingly random
offsets in dispersion direction in the FOS data base has been to
assign it to non-repeatibility in the position lock mechanism. Since
any under/over rotation would produce a dislocation of the spectra in
a direction perpendicular to the dispersion direction long before an
X-offset would be noticed, this was not a valid explanation. However
TMPOFF: The drive shaft of the filter-grating wheel motor apparently
was able to exert lateral force onto the wheel, thus altering optical
alignment between collimator, grating and detector to produce a
deviation in the exit angle for a given wavelength as seen from a
particular location on the photocathode (or diode array for
that). This effect is very strongly correlated with the ambient
temperature, in particular the one read from the FGWA motor
sensor. The amplitude of the effect is 0.06 pixels per degree change
in temperature, and since the general temperature in the HST aft
shroud has increased by more than 10 degrees between 1990 and 1997 all
data (red and blue side) need this correction.
After conversion from counts to count rates and the correction for
paired-pulse loss, the raw data is shifted in memory location
according to the sum of offsets determined from all effects above. The
correction is applied to the arrays of raw data, errors, data quality
For spectrophotometric ACCUM, IMAGE and RAPID-READOUT modes a unique
correction is determined for each data group or ystep based on the
orbital position of the spacecraft at the mid-point of the observation
time for each subintegration. While the correction is calculated to
sub-pixel accuracy, it is applied as an integer value and is therefore
accurate only to the nearest integral pixel. This is done to avoid
resampling the data in the calibration process. Special handling is
required for data obtained in ACCUM mode since each data frame
contains the sum of all frames up to that point. In order to apply a
unique correction to each frame, data taken in ACCUM mode are first
unraveled into separate frames. Each frame is then corrected
individually, and the corrected frames are recombined.
The correction is applied by simply shifting pixel values from one
array location to another. As a typical example, if the amount of the
correction for a particular data group is calculated to be +2.38
pixels, the data point originally at pixel location 1 is shifted to
pixel 3, pixel 2 shifted to pixel 4, pixel 3 to pixel 5, and so
on. Pixel locations at the ends of the array that are left vacant by
this process (e.g., pixels 1 and 2 in the example above) retain their
unshifted values and are assigned a data quality value of 700.
The correction for the post-onboard-GIMP (post 5-April-1993) datasets
involves first removing the original onboard correction and then
applying the same correction as used for pre-onboard-GIMP data.
The orbital position is not calculated from the orbital
parameters in the science header (.shh), because these elements are
those valid at SMS generation only (usually approx 7 days before the
observation). Instead, 'poa_calfos' uses a highly accurate
ephemeris program from NORAD that is based on daily updates of the
elements of HST determined from the tracking of HST. Both, the
onboard-GIMP effects and the original GIMP effect are then determined
from appropriate models of the geomagnetic field (the 10th order
version of the International Geomagnetic Reference Field, IGRF), and
an elctron-optical model of the digicons, and scaled
appropriately. The effect of YBASE value changes are determined with
reference to the YBASE values in use for the dispersion relation.
Aperture/mode specific biases are included. Finally, a weighted
average is obtained from 7 different temperature monitors in the FOS
container and converted into a temperture related offset. For details
on the physical backgrounds see ST-ECF Technical Report POA/FOS
004. For a discussion regarding how the on-board GIMP correction
was driven and how the effect can be monitored in the engineering data
stream see ST-ECF Technical
Report POA/FOS 003.
The basic parameters for the OFF_correction are saved into the POA
(group parameter) keywords, a comprehensive listing can be found here.
The released versions of 'poa_calfos' OFF_CORR correction do not resample
the data (same behaviour as the previous 'calfos'). The data are shifted
in the FOS X-direction (the dispersion direction); this shift is calculated
with sub-pixel resolution, but applied as the rounded value whole number
value. The 'pfos_pix2wav'
tool in the v1.2.1 release now encorporates the sub-pixel aspect of
the correction. This technique will not touch the FOS data itself, but
apply the correction on a line list of pixel positions.
The offset correction (OFF_CORR) is not applied to target acquisition
data and polarimetry data. The correction can be applied SPECTROSCOPY,
IMAGE and TAPID-READOUT mode data while using 'poa_calfos', or more
accurately, all the data that fit the POA processing criteria (link).
| ESA | NASA | ST-ECF Archive | ESO
| ST-ECF | STScI