4. METHOD OF SOLUTION
The important subroutines and their functions are as follows:
BGSLOP: a subroutine for calculating the slope of the off-peak plateau in the low-energy side (BOB-75 only).
CALAF: a subroutine visualizing the obtained calibration curve in relation to the given standard data in energy calibration and efficiency correction.
DECAY: a function calculating the decay correction factor. It is effective only for the built-in Eu standard sources.
ENGEF: a subroutine for energy calibration and efficiency correction. The calibration can be carried out either by constructing the calibration curve (or the efficiency curve) with the internal or external standard peaks, or by giving the values of the coefficients of the curve. The calibration curve for the energy is fitted to a polynomial of the channel number, while the logarithm of the counting efficiency is expressed in a polynomial of the logarithm of the energy.
FINDSK: a subroutine for determining the line shape function of a typical peak in the spectrum; the obtained information is used in the succeeding subprogram, PKFIT, for analysing the rest of the peak groups in the spectrum. The procedure via FINDSK is, however, performed only when the peak shape changes remarkably with each data as in the case of the gamma-scanning of the reactor fuel rod.
FUNX: a function determining the smoothing condition for each peak group, to be used for the NaI(Tl) spectrum or for the spectra given by the low-energy gamma-ray spectrometer.
GRAFIC: a subroutine visualising the result of the least squares fit.
LINEQ: a subroutine for calculating the inverse matrix.
LIVET: a subroutine for calculating the live counting time when the data have been acquired in the clock-time mode.
NLIEQ and NLNIJ: subroutines giving the values of the fitting function with/without tail and its first derivatives with respect to the involved free parameters at given values of the independent variable (the channel number).
NONLIN: a subroutine for the non-linear least squares fit.
PKFIT: a subroutine for the peak-fitting procedure. The peak shape is approximated by an asymmetric Gaussian curve with/without tail in the low-energy side. The asymmetry factor and/or the shape of the tail are either fixed or unfixed in the least squares fit. The shape of the background level can be chosen among the linear, exponential, and various sigmoid curves.
PKREC: a subroutine for the peak search.
POLIN: a subroutine for calculating the values of the coefficients of the calibration or efficiency curve and the involved errors.
POLIX: a subroutine for computing the energy and the counting efficiency of a given peak.
SMART: a subroutine for the data-smoothing following Savitzky- Golay's method.
SUBTRAX: a subroutine for detecting a weak peak buried in the natural background spectrum. The background spectrum is smoothed and then subtracted from the smoothed data of the relevant spectrum. The residues are cleared unless positive values continue over one FWHM. The resulting spectrum is smoothed once more and subjected to the succeeding peak-searching procedure.
Note: the program does not correct for the pile-up effect appear- ing in the measurement at a high counting rate, nor does it account for channel shift during measurement.