For spectrophotometry the same sensitivity calculations as for photometry
apply but here to each pixel of the detector array (cf. Chap. 6.4).
The central wavelength
refers to each pixel (see Tab. 10 where average data for 8
pixels are given). Accordingly, the source flux is
the flux expected at the particular wavelength. The wavelength
calibration of the PHT-S arrays is pretty well linear. The wavelength of
the physical centre of the nth pixel in 
is:
For integration time calculations the flux incident on one detector
pixel has to be determined. If only a line flux 
(in 
) is known, the corresponding flux 
(in 
) incident on one pixel can be determined from the
following equation:
is the spectral width per pixel for PHT-SS and PHT-SL,
respectively:
Note that there are small gaps between the individual pixels leading to a slightly smaller pixel bandwidth than the wavelength difference between the pixel centres (see above).
The equation for the total power of point-like sources 
is:
where is given in 
, and i=1,2,...,8.
For extended sources the total power 
becomes:
where 
is given in 
, and i=1,2,...,8.
The NEP for point-like sources is given by:
and for extended sources it becomes:
measured in :
The constants C4 and C5 are tabulated in Tab. 13 for eight wavelength ranges of the short wavelength (PHT-SS) and the long wavelength channel (PHT-SL), respectively.
The detector noise 
for the PHT-S detectors is:
Depending on the wavelength range selected the appropriate value has to be used.
The total noise is:
Again a differential measurement for background subtraction is assumed.
Now the integration time or the S/N ratio can be calculated from Eq. 4 and 5 in Sect. 4.
If chopping observations are requested, the exposure time per wavelength range is twice (rectangular and triangular chopping) or 3 times (saw-tooth chopping) the on-source integration time, depending on the requested chopper mode.
