主催: The Japan Society of Mechanical Engineers
会議名: 福島廃炉研究国際会議2024(FDR2024)
開催日: 2024/10/10 - 2024/10/13
We developed a signal processing system for radiation- hardened cameras that can receive and convert digital differential signals transmitted over long distances into video signal even when the drive frequency of a CMOS image sensor (CIS) fluctuates due to gamma radiation. The camera consists of a camera head component and a signal processing component, which are connected by a long-distance cable. The output signal of the CIS in the camera head component is a digital differential signal (LVDS signal). In the LVDS signal, the clock and pixel data are output in parallel. The pixel data includes a code indicating the start of a frame. The signal processing component receives the code and converts the pixel data into the video signal. In the gamma radiation test of the camera, the image was not output after 150 kGy. To investigate the cause, the clock and pixel data were evaluated. The clock frequency increased by two times after 150 kGy, while the clock slew rate (SR) did not degrade due to gamma radiation. In the code of pixel data after 150 kGy, some bits did not exceed the center voltage of the LVDS signal and were not judged as 1 (error bit). Therefore, we devised a signal processing system that can recognize the code even when gamma radiation causes the driving frequency of CIS to fluctuate. For the developed signal processing system, a part of the code ([0xFFFFFF]) was extracted that can predict the location of error bits and the increase in error bits as frequency increases. The extracted signal increases from the negative shift of the lower limit voltage at the same SR regardless of TID, so the number of error bits can be predicted to increase starting from the higher bits as the frequency increases. Therefore, in addition to the [0xFFFFFF], [0x7FFFFF] were programmed as recognition signals instead of the code for the signal processing system, and the system was designed to switch between these recognition signals depending on the increase in frequency. We determined that the proposed signal recognition system can receive the LVDS signal containing the error bit and output image. However, there was distortion in the output video received using [0x7FFFFF], which will need to be addressed in future work.