Abstract: A spectrometry method and apparatus for filtering step-like pulses output by a preamplifier to measure the energy or other desired characteristics of events occurring in a detector attached to the preamplifier. The spectrometer provides a set of filters that are capable of measuring the desired characteristic with different accuracy or resolution, detects the pulses in the preamplifier signal, and measures the time between consecutive pairs of detected pulses. For each detected pulse, the spectrometer selects a filter from the set of available filters, based on the measured time intervals between the pulse and its immediately preceding and succeeding detected pulses, applies it to the pulse, and indexes the output of the filtering operation with one or more indices identifying the selected filter.

Abstract: Techniques for measuring the baseline of the energy filter in nuclear and other spectrometers that filter pulses output by a preamplifier to measure the energy of events occurring in a detector connected to the preamplifier. These spectrometers capture the peak amplitudes of the filtered pulses as estimates of the underlying event energies and subtract a baseline value from these captured peak values in order to compensate for the energy filter's non-zero amplitude in the absence of any preamplifier output pulses. A second, baseline filter is connected to the preamplifier's output, where the basewidth of this baseline filter is significantly shorter than that of the energy filter. Times are determined when the baseline filter is not filtering preamplifier output pulses, output values from the baseline filter are captured during such determined times, and these baseline values captured from the baseline filter are used to create an accurate estimate of the energy filter's baseline value.

Abstract: For pulse processing spectrometers that use a fast channel to detect input pulses, a slow channel to measure the pulses' energies, pileup inspection circuitry pulses, and binning to produce an output spectrum of captured energy values, the correction technique extends the Harms method to produce an estimate <C> of the average ratio of the number of input pulses detected in the fast channel per non-piled-up pulse whose captured energy value is to be binned into the output spectrum and, for each such non-piled-up pulse, adding the value <C> to the corrected output spectrum in the bin dictated by the captured energy value. The uncorrected spectrum is formed in the traditional manner by simply adding 1 to the equivalent bin. Three techniques are disclosed for producing the estimate <C>, a running average, a bucket averaging, and a circular buffer method.