Caveats when Analyzing COS Data

With CALCOS 2.11c there are some issues with the Calibration
products.

FUV Pulse-Height Thresholds

FUV data taken in TIME-TAG mode include a pulse-height value for
each photon event. Eventually, pulse-height thresholds will be
applied to improve the quality of the data. At present, however,
all pulse heights (0-31) are included in the data extraction.  This
may result in a higher (but still very low) detector background.
It may also result in the appearance of several low-gain detector
features on segment B.

See Section 3.4.4 of the COS Data Handbook for more information on
how to use pulse-height filtering in CALCOS, but be aware that
adjusting pulse-height thresholds may cause other complications.
Since all of the current calcos reference files have been generated
using the full range of pulse heights, they may not be appropriate
if you choose to apply a different set of thresholds.  In particular,
the sensitivity, background, and flat field are all dependent on
the pulse-height thresholds chosen.


Change in FUV Detector Gain

Early FUV spectra showed a distinctive pair of emission and absorption
features, dubbed the "divot and clod."  The divot is a deficit of
counts located ~ 4000-5000 pixels from the blue edge of the spectrum
and the clod is a corresponding excess of counts located ~ 1000
pixels blueward of the divot.  It turns out that the X coordinates
of high-gain photon events falling on a particular region of the
detector were being systematically mis-calculated, resulting in the
counts' being shifted from one region of the detector to another.
Lowering the detector gain eliminated these features.

The gain was lowered on 12 August 2009.  FUV spectra obtained
before that date may show spurious features.


FUV Flat Field

Beware of fixed-pattern noise on the FUV detector.  COS FUV spectra
will not be properly flat fielded for some time (no on-orbit flat
exists yet), so observers should be cautious not to over-interpret
features in the data.  Data obtained at multiple FP-POS offsets are
less susceptible, but even for those spectra the fixed-pattern noise
will be present, albeit at a reduced level.  A similar warning
applies to areas with QE grid shadows, but in those areas the
data-quality flag will indicate the presence of a shadow.


Change in Instrument Focus

As part of the initial COS alignment and focus procedure, the NUV
channel focus position was changed a number of times during SMOV.
These changes were followed by an OTA secondary mirror adjustment
on 20 July 2009.  All NUV-channel data taken since that date can be
considered in focus.

Two FUV focus adjustments have been made since launch. The first,
made on 26 July 2009, moved the M gratings slightly further from
best focus than their initial launch positions. The second adjustment,
on 28 August 2009, brought all the FUV gratings to their current
best-focus positions.