Detectors calibrations with PDS experiment integrated

The PDS was assembled and intensively tested at LABEN SpA, Vimodrone, Milano, Italy. After successful thermal-vacuum tests and vibration tests, the following main response calibrations were performed on GRBM detectors in late 1994/early 1995:

• window transparency: each GRBM detector was illuminated in 30 different points on its surface with 2 collimated radioactive sources of 122 keV (⁵⁷Co) and 662 keV (¹³⁷Cs) photons at a distance of 20 cm and giving a spot size on the slab of 20 mm. These measurements, compared with those similar undertaken in Bicron, show the effect of the surrounding materials like PMT and their electronics on GRBM detectors response and are important to verify the correct modeling of the PDS structures for the Monte Carlo simulations;
• PHA channel vs. energy relation: each slab has been irradiated with collimated monochromatic sources of photons with energies ranging from 60 to 662 keV located at a distance of 4 m in order to illuminate the entire shield. The PHA channel centroids of the photo-peaks of the measured spectra at the various energies were determined by fitting the peaks with a Gaussian. In this way a preliminary energy to channel conversion was obtained for each GRBM detector. In addition, from these measurements it was possible to evaluate the detectors response at the various energies;
• calibration of GRBM and AC band thresholds and related effects: the energy values and spectral shapes corresponding to the different steps of the thresholds delimiting the GRBM ( LLT-ULT) and AC (>ACT) bands have been determined using some of the radioactive sources used in the previous set of measurements, located at the same distance from the detector. For the GRBM LLT and ULT calibration, 60 and 662 keV source was used respectively. What is observed in the 240 channels spectra collected with the thresholds at different steps is not a straight cut at a given channel value but an $\sim$exponential decrease of the counts below LLT or above ULT. This effect is due to the fact that the electronic thresholds assume a fixed pulse shape, which is not exactly true, especially at low energies, where the signal is weaker and more affected by statistical and non-linearity effects, and must be taken in serious account in modeling detectors response. The ACT threshold steps values have been measured using a 122 keV source. The count rate in the AC ratemeter with the threshold at its minimum value was measured. The spectrum of the 122 keV source consists mainly of a photo-peak, with a little Compton scattering contribution. Thus, the ratios between the AC counts measured with each step value and those measured with the minimum step depend on the area of the Gaussian photo-peak which is lost because of the cut, i.e. to the threshold energy. Another relevant threshold effect is that of the GRBM ULT on the gain calibrator light pulser line. Photons produced by the light pulser are generated in a NaI scintillator which has a different decay constant with respect to CsI. As a consequence, the electronics discriminator circuits have a different effect on the calibrator signals with respect to the CsI signal. The result is a distortion of the calibrator photo-peak, with its higher energy photons decreased by the ULT. This effect as a function of ULT must be considered in evaluating calibrator line value to check detector gain.