Microchip Technology has released a revision to its PolarFire field-programmable gate array (FPGA) Ethernet sensor bridge, a smaller production-ready board for integration into Nvidia edge AI systems.
Microchip said the need for this revision comes from the shift to compact, high-performance AI systems that is pushing sensor connectivity to smaller form factors to provide more space for racks and architectures.
The revision to the Ethernet sensor bridge enables a 60% smaller form factor with support for twice the number of cameras and is USB-C powered for rack deployment. Microchip said these changes allow engineers to replace multi-interface sensor connections with a 10 Gb Ethernet-based architecture that reduces complexity.
This could be used for AI applications powered by Nvidia Jetson and IGX systems like:
- Medical
- Industrial
- Humanoid robotics
The updated board includes support for up to four cameras that are compatible with Nvidia Jetson and features on-board optical latency measurement circuitry for validation using Nvidia’s display analysis tool.
Other features on the board include:
- Cameras
- Cables
- Reference design
- Board schematics
- Design collateral
- Integrated timing and power management ICs
- Interface support including MIPI CSI‑2, I²C, UART and GPIO
- Multiple sensor and video interface support like SLVS-EC 2.0, 12G-SDI, HDMI and DisplayPort
