The overall ISOCAM consortium has studied detector models to correct for
transient and drift effects. However, up-to-now no correction method
is working perfectly. For this reason, and certainly being a little
conservative, we recommend a substantial increase of the number of
exposures ( ) to be taken at each polarizer position.
Transients in the camera involve several time constants which are
empirically known to be inversely proportional to the flux. The flux
F [ADU/s/g] includes the flux from your source and the zodiacal
emission. You can find values for the zodiacal emission in the various
ISOCAM filters in the ISOCAM Observer's Manual (s. Table 22). The
main transient is often a kind of exponential
with time constant as estimated by the following empirical formula:
This formula can be taken to estimate
the minimum number of
exposures to be taken at each polarizer position before
switching and for those cases where we hope to be able to correct
drifts with future transient correction methods.
However, as sated above we have to take a conservative
approach and recommend for the value of :
where F is the source flux in units [ADU/s/g]
and 
is the integration time in seconds. For a conversion of astronomical
units such as [Jy] or [mJy/arcsec 
], into engineering units [ADU/s] please
see ISOCAM Observers Manual (Table 20).
Equation (3.2) is still our best recommendation even
for particularly critical observing project such as measuring the
polarization of faint emission around a bright source. Due to the
displacement of the sky image on the array from one polarizer to the
next, the measure of the faint emission is hampered by the memory
effect on the source.
For example if you are observing at gain = 2 with = 2s and
expect a source flux of 100 ADU/s/g the recommended number of exposure
reads: = 100.