R. Farrow et al., XSPRESS - X-RAY SIGNAL-PROCESSING ELECTRONICS FOR SOLID-STATE DETECTORS, Nuclear instruments & methods in physics research. Section B, Beam interactions with materials and atoms, 97(1-4), 1995, pp. 567-571
With recent improvements in synchrotron sources and X-ray optics great
pressures have been placed on detector systems to produce higher coun
t rates and better resolutions. Present high performance 13 element ge
rmanium detector systems can give reasonable count rates with good res
olution (similar to 10(4)-10(5) Hz per channel and similar to 250 eV F
WHM @ Fe-55 with 0.5 mu s shaping time). However, these systems are re
stricted by limitations in both the detector and in the analogue pulse
processing after the detector. With respect to the detector, increasi
ng the number of channels without degrading the energy resolution is a
great challenge due to increased crosstalk and capacitance. The analo
gue pulse processing electronics are significantly limited by the dead
time introduced by the shaping amplifier. This dead time causes pulse
pile-up at higher rates which leads to non-linearity and poor resolut
ion. This paper describes the XSPRESS system which has been developed
at Daresbury Laboratory for the new Wiggler II beamline 16. This syste
m overcomes previous limits in both signal processing and detector fab
rication to give great improvements in system performance. The signal
processing electronics departs from standard analogue processing techn
iques and employs sophisticated adaptive digital signal processing har
dware to reduce the dead time associated with each event to a minimum.
This VME based technology allows us to vastly increase the count rate
for each channel yet still retain the ability to gain very good resol
ution. The detector has been developed through a collaborative agreeme
nt with EG and G Ortec and packs an unprecedented 30 germanium crystal
s into an extremely small area whilst still retaining the energy resol
ution of smaller arrays. This system has increased throughput rate by
an order of magnitude per channel and when all channels are implemente
d, an increase of at least two orders of magnitude for the whole array
should be seen. Data has been taken using this system on the SRS at D
aresbury Laboratory and these results will be given along with a detai
led explanation of the operation of this system.