Nuclear spectrometry measurements involve laboratory instruments that primarily employ digital signal processing (DSP).
Compared to traditional analog designs, DSP enables high stability, improved resolution, and higher throughput in all applications.
All the advantages listed above provide improved spectrum quality and analysis results. In connection with the development of low-power digital signal processing modules, progress has been made in implementing DSP in portable devices and building analyzers on their basis in standard PC boards.
Digital signal processing.
The ionizing radiation spectrometer converts the electric charge pulses from the detector into voltage pulses, the amplitude of which is measured and stored in the spectrometer's memory as a histogram. This histogram of pulse amplitudes is analyzed by software for the qualitative and quantitative determination of the isotopes present in the sample.
In an analog system, the pulse is formed by analog electronic circuits whose characteristics tend to change depending on temperature and other measurement parameters. In a digital system, the pulse is formed by a digital signal processor implemented as a very large integrated circuit, which is a highly stable spectrometer. One of the most important advantages of a digital system is the virtually infinite number of combinations of available pulse formation parameters. This makes it possible to select the best operating mode for each specific detector, which would allow achieving the best resolution and throughput on this specific detector.
The only similarity revealed when comparing the block diagrams of an analog and digital spectrometer is the presence of an emitter follower at the input to receive a signal from the preamplifier, and an interface for communication with a computer at the output. Immediately after the input follower in the digital spectrometer, a fast ADC digitizes the shape of each input signal and converts it into a string of numbers. A digital filter processes this information using a special algorithm. After the digital filter, the functions of baseline restoration, fine-tuning of gain, and stabilization of the spectrum with digital accuracy and stability are performed.
Digital filters . The shape of a digital filter is shown in the figure.

The filter has the form of a trapezoid with sides that can be concave and a flat top that can be inclined or have zero width (then the trapezoid is reborn as a triangle). The traditional trapezoid filter is used for working with coaxial semiconductor detectors. With a large volume of detectors with a ballistic deficit, filters with an inclined top are used, which allows for partial compensation of the deficit. If we are talking about working with planar semiconductor detectors and scintillation detectors, then in this case a triangular filter shape is mainly used or the width of the flat top is made minimal. At this stage of development of modern digital spectrometers, it is possible to set several dozen possible values for each filter parameter: rise time, width of the flat top, concavity of the trapezoid side, angle of inclination of the flat top. A large selection of digital filter parameters makes it possible to fine-tune the spectrometer for a specific detector used.
Comparative characteristics of digital and analog devices
The first comparisons of analog and digital spectrometers were conducted at the Los Alamos National Laboratory in the USA. The first serially produced digital spectrometer DSPec (manufactured by ORTEC, USA) and an analog system based on NIM blocks were selected for comparison; a 4002D power supply (ORTEC), a 3106D high-voltage power supply (manufactured by Canberra, USA), an 8077 amplitude-to-digital converter (Canberra), an 8232 digital stabilizer (Canberra), a 672 spectrometric amplifier (ORTEC), and a 4610 multichannel analyzer (Canberra). Two coaxial detectors made of pure p-type germanium manufactured by Canberra with efficiencies of 23% and 25%, respectively, were used in the work.
Two sources of ionizing radiation were used in the experiment: 57 Co and 60 Co to evaluate the devices in the low and high energy ranges, respectively. Spectra were collected at the input at 1, 3, 10, and 30 kHz loads with shaping times of 2, 4, 6 μs for the analog and rise times of 4, 8, and 12 μs for the DSPec. For the 50 kHz load, a shaping time of 2 μs and a rise time of 4 μs were used.
As a result of the comparison, it was possible to establish that in most experiments the digital spectrometer showed better characteristics compared to the analog device. In particular, it was revealed that the DSP has a higher throughput with the same resolution as the analog system.