Parallel processing method for high-speed real time digital pulse processing for gamma-ray spectroscopy

A new data acquisition (DAQ) system was developed to fulfil the requirements of the gamma-ray
spectrometer (GRS) JET-EP2 (joint European Torus enhancement project 2), providing high-resolution
spectroscopy at very high-count rate (up to few MHz). The system is based on the Advanced Telecommunications
Computing ArchitectureTM (ATCATM) and includes a transient record (TR) module with 8
channels of 14 bits resolution at 400 MSamples/s (MSPS) sampling rate, 4GB of local memory, and 2 field
programmable gate array (FPGA) able to perform real time algorithms for data reduction and digital pulse
processing.
Although at 400 MSPS only fast programmable devices such as FPGAs can be used either for data processing
and data transfer, FPGA resources also present speed limitation at some specific tasks, leading
to an unavoidable data lost when demanding algorithms are applied. To overcome this problem and
foreseeing an increase of the algorithm complexity, a new digital parallel filter was developed, aiming to
perform real time pulse processing in the FPGAs of the TR module at the presented sampling rate. The
filter is based on the conventional digital time-invariant trapezoidal shaper operating with parallelized
data while performing pulse height analysis (PHA) and pile up rejection (PUR). The incoming sampled
data is successively parallelized and fed into the processing algorithm block at one fourth of the sampling
rate. The following data processing and data transfer is also performed at one fourth of the sampling rate.
The algorithm based on data parallelization technique was implemented and tested at JET facilities,
where a spectrum was obtained. Attending to the observed results, the PHA algorithm will be improved
by implementing the pulse pile up discrimination.
© 2010 Elsevier