A Portable, Low Power Radiation Detection and Identification System for High Count Rate, Long Term Monitoring

Authors

Samuel J. Murray, University of Nebraska-LincolFollow

First Advisor

Sina Balkir

Second Advisor

Michael Hoffman

Date of this Version

Summer 8-2023

Citation

S. J. Murray, "A Portable, Low Power Radiation Detection and Identification System for High Count Rate, Long Term Monitoring," Ph.D. dissertation, Department of Electrical and Computer Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska, August 2023.

Comments

A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfilment of Requirements For the Degree of Doctor of Philosophy, Major: Electrical Engineering, Under the Supervision of Professors Sina Balkır and Michael Hoffman. Lincoln, Nebraska: August, 2023

Copyright © 2023 Samuel J. Murray

Abstract

This dissertation presents the design of a novel radiation detection and identification system that can operate continuously over a period of 8 days while detecting at 30,000 counts per second, consuming a total of 11 mW. The entire system is highly integrated, containing a gamma ray detector, a high voltage detector power supply, and a multichannel analyzer (MCA) system-on-a-chip (SoC), which are all combined into a compact form using a multi-level, configurable printed circuit board design. The MCA SoC, fabricated using a 65 nm CMOS technology, features two enabling resources to allow low power detections at high count rates for detectors that produce slow or fast pulses alike, supporting a variety of detector compounds such as CsI, NaI, and LaBr. The resource designed for slower detectors is based on an array of current mirrors that sample and remove an error current that would otherwise cause significant spectral corruption at higher count rates, while the other is meant for faster detectors and mitigates the error current corruption by stabilizing the signal baseline. Both on-chip circuits are configurable and allow the use of low power, event-driven downstream circuitry and signal processing techniques. Furthermore, the SoC contains a hardware neural network accelerator capable of using the spectral data acquired by the MCA to estimate the constituent radioisotopes within the source, executing in 694 μs and using 1.94 μJ of energy. The SoC also features a 32-bit RISC-V microcontroller that can execute custom programs and transmit the acquired spectral data in less than 10 ms using a serial readout link.

Advisors: Sina Balkır and Michael Hoffman