The spectral responsivity of a photomultiplier depends on the photocathode
composition, the window composition, and the mode of illumination (i.e., from
the vacuum side or the substrate side).
For many common situations, these factors have been combined into
standard spectral responses known as ``S-numbers''
(S-1, S-4, S-20, etc.)
that are used by most, but not all, manufacturers.
The spectral response should therefore be indicated by a character column
named
SNUMBER,
containing the S-number, if it is known (e.g., 'S-13').
Other valid string values are
'BIALKALI'
for ``bialkali'' photocathodes with glass windows;
'QBIALKALI'
for ``bialkali'' photocathodes with fused-quartz windows;
'GAAS'
for gallium arsenide ``negative-electron-affinity'' photocathodes
with glass windows;
and
'QGAAS'
for gallium arsenide ``NEA'' photocathodes with fused-quartz windows.
Any spectral response markedly different from these should be flagged as
SNUMBER= 'OTHER',
and the spectral response supplied in a separate table file, as described
under
DET= 'OTHER' (see section 
).
Photomultipliers are used in different modes, which have different properties. This is specified by a character column MODE, whose values may be 'PC' for pulse counting; 'DC' for DC photometry; or 'CI' for charge integration.
If MODE= 'PC', additional information is needed to describe the pulse-overlap (``dead-time'') correction. This is given in another character column, DEADTYPE, which can have the values 'EXTENDING', for a paralysable counter, or 'NONEXTENDING', for a non-paralysable counter. If the counter's behavior is unknown, and cannot be determined, set DEADTYPE= 'UNKNOWN'. The estimated value of the dead-time parameter itself, in seconds, goes in a Real*4 column called DEADTIME, and the uncertainty of the dead-time parameter, again in seconds, goes in a Real*4 column called DEADTIMEERROR.
Users should be aware that, because pulse pile-up partly offsets coincidence losses, the effective dead-time parameter depends on a combination of the resolution of the discriminator, the discriminator setting, and the characteristics of the individual photomultiplier under actual conditions of use, such as temperature and voltage, that affect the pulse shape and pulse-height distribution. Therefore, it is essential to keep these parameters fixed during a run. Also, the effective dead-time parameter should be determined from actual photometric data gathered for the purpose of determining its value accurately; nominal values of pulse-resolution times from manufacturers, or pulse-resolution times determined with pulse generators, are not suitable for correcting photometric measurements.