FPGAs kullanılarak MRI gradient, RF-iletici ve RF-alıcı dizi zincirleri için çok kanallı optik iletişim sisteminin geliştirilmesi

Modern Magnetic Resonance Imaging (MRI) systems are increasingly using array configurations for gradient, transmit, and receive chains due to their ability to improve imaging performance and spatial control. However, this introduces cabling complexity; array systems require multiple coaxial cables and dense on-coil electronics, leading to bulky cable assemblies and handling challenges. Long coaxial cables between the console and magnet rooms can heat up under high currents, change resistance, and be difficult to handle when placed near amplifiers or FPGAs. Moreover, RF chokes on coaxial help to reduce shield currents but increase the cable bulk, making handling difficult. Optical fibers, Small Factor Pluggable Plus (SFP+) or Plastic Optical Fiber (POF), are a good substitute as there are no conductive parts or currents traveling. Field-Programmable Gate Arrays (FPGAs) with built-in high-speed transceivers will be used to create an optical fiber communication network to facilitate the creation of the physical layer design. FPGAs built-in AXI 1G/2.5 Ethernet Subsystem library can also help create a data link layer design, which is proposed for the network layer. To reduce the cost of the system, a two-stage architecture is proposed with a spectrometer and distributor board communicating on SFP+ cables and the distributor and controller boards communicating on POF cables. This architecture can control multiple channels with reduced cost and cabling parameters. The SFP+ will work with the Ethernet Subsystem IP at 2.5 Gbps, and the network layer design will have decreased overhead. A simple Universal-Asynchronous Receiver Transmitter (UART) code will manage the POF part of the architecture, working at 20 Mbps or 100 Mbps. Simulation and Integrated Logic Analyzer (ILA) tests have been done to verify the design of each code block. Moreover, a complete working system showing Pulse-Width Modulation (PWM) waveforms for the gradient and RF-Transmit chain and data storage and reconstruction on the console PC are demonstrated for the RF-Receive chain.