Speaker
Description
The ePIC Longitudinal Forward Hadronic Calorimeter (LFHCal) is a steel plastic scintillator sampling calorimeter with SiPM-on-tile readout, covering a pseudorapidity range of (1.2 < $\eta$ < 3.5). It consists of 60 layers of steel absorbers and scintillator tiles, corresponding to approximately 6 interaction lengths. The large number of SiPMs required for the detector makes fast, reliable, and consistent quality control essential. To address this challenge, we have developed a robotic testing setup for the quality control of Hamamatsu S14160-1315PS SiPMs. The system measures the I–V curve, single-photon spectrum, breakdown voltage, gain, and dark current of each SiPM. With a throughput of approximately 500 SiPMs per day, the automated system significantly reduces the time and manual effort required for large-scale testing. In addition, a separate automated optical inspection setup incorporating convolutional neural network (CNN)-based image classification has been developed to identify surface imperfections in SiPMs.
This presentation describes the design, operation, and results of both automated setups. Furthermore, using the large dataset collected from approximately 20,000 SiPMs tested to date, we compare the performance of several methods for extracting the breakdown voltage $(V_{\mathrm{br}})$ from I–V curves.