Comparison of three resistor network division circuits for the readout of 4×4 pixel SiPM arrays

The purpose of this study is to investigate the behavior of a flexible SensL's silicon photomultiplier array (SPMArray4) photodetector for possible applications in PET imaging. We have designed and evaluated three different resistor network division circuits to read out the signal outputs of a 4×4 pixel SiPM array. We have applied firstly (i) a symmetric resistive voltage division circuit, secondly (ii) a symmetric resistive charge division circuit and thirdly (iii) a charge division multiplexing resistor network reducing the 16 pixel outputs to 4 position signals. In the first circuit the SensL SPMArray4-A0 preamplification electronics and a SPMArray4-A1 evaluation board providing the 16 pixels voltage outputs were used, before the symmetric resistive voltage network. We reduced the 16 voltage signals firstly to 4X and 4Y coordinate signals. Then those signals were further reduced to 2X and 2Y position signals connected via a resistor network. In the second readout circuit we have used the same technique but without the preamplification stage. The third circuit is based on a discretized positioning circuit, which multiplexes the 16 signals from the SiPM array to 4 position signals. The 4 position signals (Xa, Xb, Yc and Yd) were digitized using a free running sampling technique. An FPGA (Spartan 6 LX16) was used for triggering and signal processing of the pulses. We acquired raw images and energy histograms of a BGO and a CsI:Na pixilated scintillator under 22Na excitation. A clear visualization of the discrete 2×2×5 mm3 pixilated BGO scintillator elements as well as the 1×1×5 mm3 pixilated CsI:Na crystal array was achieved with all applied readout circuits. The symmetric resistive charge division circuit provides higher peak to valley ratio than the other readout circuits. Τhe sensitivity and the energy resolution remained almost constant for the three circuits.