The signal processing of PHT-S staring mode data starts at the level where one signal in V/s has been derived for each integration ramp. In the following processing steps, the signal records of ramps that are completely saturated are skipped.

7.4.1 Dark signal subtraction

The dark signal data are stored in Cal-G file PSDARK (128 pixels of PHT-S), see Sections 14.7 in particular 14.7.1.

7.4.2 Transient correction and calibration of signals

For each pixel

the signals

are directly converted to flux densities $F^t_{\nu}(i)$ using a transfer function:

$\displaystyle F^t_{\nu}(i)$	$\textstyle =$	$\displaystyle \varphi(i,s(i,t),t)~~~~[Jy],$	(7.14)
$\displaystyle {\Delta}F^t_{\nu}(i)$	$\textstyle =$	$\displaystyle \frac{{\Delta}s(i,t)}{s(i,t)}F^t_{\nu}(i)~~~~[Jy],$	(7.15)

The function $\varphi(i,s(i,t),t)$ has been implemented as a table with logarithmic signal and time spacings. The value is obtained from linear interpolaton. We give the flux a superscript

to indicate that the flux itself is supposed to be constant and has no time dependence. Time

along a measurement is computed from

where $t_{mid}$ is the time at the midpoint of a given ramp and

the time of the first readout of a given measurement. Signals with $t\,<\,1\,$ s are not processed.

This transformation yields transient corrected flux densities $F^t_{\nu}(i)$ along the measurement. Due to the transient behaviour of the PHT-S detectors, the transient correction is larger at the beginning of a measurement where the signal is less stabilised than after a long period of integration. The stabilisation time depends on the absolute brightness of the source.

To include this information, a weight is given to each $F^t_{\nu}(i)$ based on the degree of transient correction:

Ancillary data required:

The transfer function $\varphi(i,s,t)$ and corresponding weight table

are stored in Cal-G files PSDYNAMIC and PSDYNWT, see Section 14.9 for detailed descriptions.

7.4.3 Computation of the weighted mean flux

The weighted mean of the array of flux densities $F^t_{\nu}(i)$ is computed using weights obtained from the combination of the weighting factor

and the statistical uncertainty ${\Delta}F^t_{\nu}(i)$ :

$\begin{displaymath} w(i,t) = \frac {w'(i,t)}{{\Delta}^2F^t_{\nu}(i)}, \end{displaymath}$

(7.18)

$\displaystyle \langle F_{\nu}(i) \rangle$	$\textstyle =$	$\displaystyle \frac {\sum_{t>1s} F^t_{\nu}(i)w(i,t)}{\sum_t w(i,t)}~~~[{\rm Jy}],$	(7.19)
$\displaystyle \Delta \langle F_{\nu}(i) \rangle$	$\textstyle =$	$\displaystyle \sqrt{\frac{\sum_{1}^{N}w(i,t) \times ( \langle F_{\nu}(i) \rangle - F^t_{\nu}(i))^2} {(N-1)\sum_{1}^{N}w(i,t) }}~~~~[{\rm Jy}].$	(7.20)

The calibration for PHT-S is most valid for detector temperatures of $2.7<T_{det}<3.1~$ K. Since there were occasions during operations where PHT-S operated outside this temperature interval, a warning will be written to the SPD header in case the temperatures as given by the ERD keywords PSERTEM1 and PSERTEM2 are outside the interval.

If $F_{\nu}(i)$ for pixel

is outside the flux density range then a warning message will be written to the SPD header that the flux for pixel

is out of the calibration range.

Ancillary data required:

The flux limits for which the calibration is reliable are stored in the header of the Cal-G file PSDYNAMIC (Section 14.9) under keywords FLXiL and FLXiU (lower and upper bound, resp.). where

is the pixel number.

7.4 Signal Processing: Staring PHT-S

7.4.1 Dark signal subtraction

Ancillary data required:

7.4.2 Transient correction and calibration of signals

Ancillary data required:

7.4.3 Computation of the weighted mean flux

Ancillary data required: