FET1126B-BJ-S System on Module (SoM) by Forlinx is a Rockchip RV1126B/RV1126BJ processor-based AI computing module that is industrially qualified. This small module targets modern embedded systems all too well, giving it a strong reputation in edge AI processing, machine vision, and real-time video analytics, so it is excellent at smart cameras, industrial monitoring, AIoT gateways, and intelligent automation systems.
The FET1126B-BJ-S is based on a quad-core ARM Cortex-A53 CPU with a speed of either 1.6 GHz (commercial version) or 1.3 GHz (industrial version), offering balanced performance, yet consumes not much power. Alongside the CPU, it has a 3 TOPS neural processing unit (NPU) that has the potential to execute AI models effectively on the device without any connections to the cloud being made.
Another highlight of such a module is that it has rich I/O and connectivity. Functional pins are completely led out through stamp holes and LGA pads, and offer display interfaces such as MIPI-DSI and RGB, Ethernet interface (RGMII) support, UART, CAN FD, SPI, I2C, PWM, and ADC. Also, the GPIO design can be completely interchanged with the 40-inch current Raspberry Pi system and thus can be more easily integrated with other architectures and prototypes.
The FET1126B-BJ-S may be configured with 1 GB to 4 GB LPDDR4 RAM and 8 GB to 64 GB eMMC storage in a flexible configuration, and this can be handy in a wide range of tasks, such as light data processing in addition to more serious compute-intensive jobs.
The FET1126B-BJ-S can operate at industrial temperature ranges of -40 °C to +85 °C and boasts a long life span in a variety of severe environments due to its design to be used in a practical setting. With a broad temperature span, high testing, and stability of its supply over long periods, it is a reliable product that can be used in long-term industry and commercial uses.
The company also developed the OK1126Bx-S carrier board, a compact baseboard designed for Forlinx RV1126 series SoMs, providing essential interfaces for edge AI and vision applications. It features MIPI-CSI camera connections, display, Gigaflit Ethernet, USB, CAN FD, UART, SPI, I2C and GPIOs, which are popular with smart cameras and industrial, AI-based systems. The board makes development simple through enabling a reliable power supply, debugging and simple peripheral expansion.
The OK1126Bx-S platform comes with a complete Linux software support package provided by Forlinx, including a Board Support Package (BSP), kernel source, device drivers, and sample applications. It delivers mainstream AI platforms via the RKNN toolkit by Rockchip that provides on-device inference of vision and edge AI workloads. It is compatible with popular frameworks like TensorFlow, PyTorch, Caffe and MXNet with support for mixed-precision AI workloads (INT8/INT16).
In my point of view, Forlinx does not list public pricing for the RV1126 SoM or compatible OK1126Bx-S carrier board. Pricing is available on request and depends on configuration and order volume. The board is currently available for evaluation and commercial orders through Forlinx and its authorised distributors.