The detectors developed for UVCS are two dimensional photon
counting, centroiding, microchannel plate sensors with electronic readout.
Two virtually identical detectors are used for the Ly-
and O VI channels of the spectrometer assembly. A KBr coated, low resistance Z stack of microchannel
plates (MCP's) provides detection and amplification, and a multilayer
cross delay line anode (XDL) accomplishes position readout. The format
is 
, electronically digitized to 
pixels. This
detector and its detailed performance characteristics are discussed by
Siegmund et al, (1994).
KBr photocathode material is deposited directly on the MCP Z stack as
an opaque layer. Incoming photons interact with the photocathode,
resulting in photoelectron emission (Siegmund et al, 1987; Siegmund et al, 1990) and a subsequent charge avalanche
in the MCP's, giving an overall charge multiplication of ~ 2 X 10^7. This
charge cloud is drifted from the MCP output to the delay line anode. Here
the charge is divided equally between two orthogonal sets of charge
collection fingers on the XDL in the X and Y axes. These are connected to
X and Y external serpentine delay lines (Siegmund et al, 1994) so that the X and Y photon
event centroid positions may be deduced from the signal arrival time
differences at the two ends of each delay line.
The KBr photocathode quantum detection efficiencies (QDE), which were measured prior to installation in the instrument,
for both UVCS detectors is given in
Table IV.
The photocathodes are solar blind, cutting off at about 1600 Å.
MCP photon event pulse amplitude distributions are narrow (<30%
FWHM for O VI) allowing amplitude discrimination of background events
if desired. The overall background rate is low, with a rate of <1 event cm
s
uniformly distributed across the FOV. This was expected
for intrinsic beta decay MCP noise (Siegmund et al., 1988; Fraser et al., 1987) for 40K
in the Philips MCP glass.
Single pixel spot image event counting rates in excess of 100
events/pixel/sec have been achieved with no degradation of the MCP gain.
The detector is stable to better than 1 pixel for global counting rates up to
events per sec, with a calibrated photometric rate/deadtime
characteristic (<40% dead time @ events s). The thermal drift
is <1 pixel over the expected UVCS temperature range. Flat field images
are dominated by MCP fixed pattern noise and are stable, however the
expected MCP multifiber modulation is at an unusually low level (a few
%).
The Ly-
detector has an optically flat MgF
entrance window, thus
blocking out all thermal ions or electrons from the detector, and all
wavelengths <1100 Å. The O VI detector is ``open face'' with a 90%
transmission mesh, biased at +15V, for attenuation of low energy ions,
placed 9mm in front of the MCP's at the entrance to the detector vacuum
housing. A feature of the XDL system is a commandable artificial pulse
(stimulation pulse) injected into each delay line at a rate of ~40 sec,
resulting in a spot image effectively at the center of the detector aperture.
The data obtained from this feature gives an in orbit evaluation of
electronics health. A number of other command and diagnostic functions
are provided for the XDL detectors. Specific types of command
functions include high voltage control, signal threshold level adjustments,
and stimulation pulser control. Housekeeping data also includes power
supply voltage monitors, X and Y pulse height data, fast event counters,
and temperatures.