Yuan AIR6N0-C-MB NX 4xGMSL2 NVIDIA Jetson Orin NX and Orin Nano Compact Carrier Board with 4x GMSL2 Camera Inputs

Cameras on a robot arm, a vehicle or a large machine often sit metres from the compute, which is further than MIPI cables reach. The AIR6N0-C-MB NX 4xGMSL2 pairs YUAN’s AIR6N0-C-MB NX space-saving carrier board with a four-channel GMSL2 camera board that matches the carrier footprint.

It takes the Jetson Orin NX 16GB or 8GB Super module at 157 or 117 TOPS, or the Orin Nano 8GB or 4GB Super at 67 or 34 TOPS. The Orin NX pairs a 1024-core NVIDIA Ampere GPU with 32 Tensor Cores and a 6- or 8-core Arm Cortex-A78AE CPU, and can encode one 4K60 or six 1080p60 streams in AV1 or H.265. On the Orin Nano the M.2 slots run at PCIe Gen3 and video encoding falls to the CPU cores, so the Orin NX is usually the better fit for multi-stream recording.

The GMSL2 board on the AIR6N0-C-MB NX 4xGMSL2 takes four cameras at up to 1920 x 1080 at 60 fps, with 1280 x 720 at 90 fps and 1080 x 1280 at 120 fps among the listed modes. A DB9 connector carries CAN bus for the vehicle or machine network, jumpers bring out I2C and a UART, and a microSD slot adds removable storage. The carrier keeps Mini DisplayPort 1.4, Gigabit Ethernet, a USB 3.2 Gen2 Type-C port and two M.2 2230 slots for NVMe and wireless. YUAN offers this board in PCIe and MIPI versions, so the build is confirmed when the board is ordered.

The AIR6N0-C-MB NX 4xGMSL2 takes 12-36 V DC and is rated from 0 to +60 °C with airflow (storage -20 to +80 °C), with CE and FCC marks and JetPack 6.2 or later. YUAN’s QCAP SDK handles capture, recording and RTSP, SRT or WebRTC streaming, and the QDEEP SDK adds detection, segmentation, OCR and licence plate recognition models.