AboutROSAT |
ROSAT Home Page |
ROSATImages |
---|

The bulging detector window causes a large-scale spatial
non-linearity (counter gas pressure 1.5bar, bulging 1.2mm).
The primary electron cloud drifts along the electric field lines
which run perpendicular to the equipotential plane
(PSPC entrance window).
Thus event positions within the central ring of the window support
structure are slightly shifted towards the detector center while
for event positions outside the central ring the shift depends
on the event position.
(*cf.* Fig. F.12 and Fig. F.13, top)
This shifting introduces spatial non-linearities.
The shape of the correction has been determined at a photon energy
of *E*=0.3keV, the positional and energy-dependent effects
were assumed to be separable.
The energy-dependent term is a smooth function of the energy
(*cf.* Fig. F.13, bottom).
Is is normalized to unity at *E*=0.3keV.

In practice SASS determines the shift using a binned map with
512 512 pixels, *i.e., * uses a grid of bin size 16 detector pixels
and has to interpolate between four neighboring grid point.
The interpolation of the correction vector is performed according to
Eq. (F.13) and (F.14).

Within SASS the energy dependent scale factor is taken from a table with 256 elements (channels). For the value is taken instead.

**Figure F.12:**
Window correction tables for PSPC-B.
The `Golden Disk' (arrays and ):
The position correction in *x* and *y* is plotted for every
8th's array element; the energy is assumed to be 0.93keV,
for which the energy dependent factor of the position
correction equals 1.0.
The origin of the displacement vector of element (*n*,*m*) is
drawn at .
The basic information was taken from files `p$manpar:tabx_093_b.dat` and `p$manpar:taby_093_b.dat`.

**Figure F.13:** Window correction data files according to Eqs.(F.13)
and (F.14): top panel: shift of coordinates for PSPC-C:
white corresponds to negative values and black to positive
(top left: , top right: ). Bottom panel: energy
dependent factor , normalized to unity at 0.93keV.
The basic information was taken from files
`p$manpar:tabx_093_c.dat`, `p$manpar:taby_093_c.dat`, and `p$manpar:scal3_new.dat`.

**Parameters:**

- ] pulse height of event, corrected for ADC
non-linearity, gain saturation, temporal and
spatial variations (PI channel)
- ,
- detector coordinates (electronic position correction applied)
- position dependent correction arrays (Fig. F.12 and F.13, top panel)
- energy dependent correction factor (Fig. F.13, bottom panel)
*r*- random number (real number in [0.0,1.0])
- ,
- detector coordinates (electronic position correction and window correction applied)

**SASS source code:**
package `PSPC`, bubble `CT`, subroutine `DCORW`

**SASS files:**

- for PSPC-B (Fig. F.12)
`p$manpar:taby_093_b.dat`- for PSPC-B (Fig. F.12)
`p$manpar:tabx_093_c.dat`- for PSPC-B (Fig. F.13, top panel)
`p$manpar:taby_093_c.dat`- for PSPC-B (Fig. F.13, top panel)
`p$manpar:scal3_new.dat`- (Fig. F.13, bottom panel)