reComputer J3011 Edge AI Device with Jetson Orin Nano 8GB, Fanless 40 TOPS, Dual GbE PoE, RS232/422/485 & CAN

Overview

reComputer Industrial J3011 is a fanless Edge AI Device built on the NVIDIA Jetson Orin Nano™ 8GB module delivering 40 TOPS AI performance. It provides hybrid industrial connectivity including 2 RJ-45 GbE ports (LAN1 with PoE PSE 802.3af 15 W), 1x DB9 supporting RS-232/RS-422/RS-485, 4x DI/4x DO with isolated grounds, and 1x CAN. The compact enclosure (159mm x 155mm x 57mm, 1.57kg) supports desk, DIN rail, wall-mounting, and VESA. Operating temperature is -20 ~ 60°C with 0.7m/s airflow, and typical power efficiency shown is 10W – 20W (per product visuals).

Key Features

• NVIDIA Jetson Orin Nano 8GB module for up to 40 TOPS AI performance.
• 1024-core NVIDIA Ampere architecture GPU with 32 Tensor Cores; 6-core Arm® Cortex®-A78AE v8.2 64-bit CPU (1.5MB L2 + 4MB L3).
• 8GB 128-bit LPDDR5, 68 GB/s bandwidth.
• Industrial I/O: 2x RJ-45 GbE (1x PoE PSE 802.3af 15 W), RS232/RS422/RS485 via DB9, 4x DI/4x DO, 1x CAN.
• USB: 3x USB3.2 Gen1, 1x USB2.0 Type-C (Device mode), 1x USB2.0 Type-C for Debug UART & RP2040.
• Displays: 1x HDMI Type A; 1x DP. Cameras: 2x CSI (2-lane 15pin).
• Storage: 1x M.2 Key M PCIe (M.2 NVMe 2280 SSD 128G included).
• Expansion: Mini PCIe for 4G/LoRaWAN® (module optional); M.2 Key B for 4G/5G (module optional); optional SMD Wi‑Fi/Bluetooth (custom production).
• Power options: DC 12V–24V terminal block (2-pin) or 19V power adapter (without power cord).
• Fanless passive heatsink; 1x 4‑pin fan connector (5V PWM) available on board.
• Rugged design: -20 ~ 60°C with 0.7m/s airflow; 95% @ 40°C non‑condensing; 3 Grms vibration; 50G shock.
• Certifications indicated: CE, FCC, RoHS, REACH, UKCA, KC.

Note on SSD Compatibility

For SSDs, Seeed recommends the 128GB, 256GB, and 512GB NVMe M.2 2280 models from Seeed due to JetPack version differences affecting SSD compatibility.

Specifications

AI Performance	40 TOPS (Jetson Orin Nano 8GB)
GPU	1024-core NVIDIA Ampere architecture GPU with 32 Tensor Cores
CPU	6-core Arm® Cortex®-A78AE v8.2 64-bit CPU; 1.5MB L2 + 4MB L3
Memory	8GB 128-bit LPDDR5, 68 GB/s
Video Encoder	1080p30 supported by 1–2 CPU cores
Video Decoder	1x 4K60 (H.265) | 2x 4K30 (H.265) | 5x 1080p60 (H.265) | 11x 1080p30 (H.265)
Carrier Board
Storage	1x M.2 Key M PCIe (M.2 NVMe 2280 SSD 128G included)
Networking	LAN1: RJ45 GbE PoE (PSE 802.3af 15 W); LAN2: RJ45 GbE (10/100/1000Mbps)
USB	3x USB3.2 Gen1; 1x USB2.0 Type-C (Device mode); 1x USB2.0 Type-C for Debug UART & RP2040
Display	1x HDMI Type A; 1x DP
Fan Header	1x 4-pin Fan Connector (5V PWM)
DI/DO & CAN	4x DI; 4x DO; 3x GND_DI; 2x GND_DO; 1x GND_ISO; 1x CAN
COM	1x DB9 (RS232/RS422/RS485)
SIM	1x Nano SIM card slot
Expansion	Mini PCIe for 4G/LoRaWAN® (module optional); M.2 Key B supports 4G/5G (module optional)
Wi‑Fi/Bluetooth	Supports SMD Wi‑Fi/Bluetooth (customized production)
TPM	1x TPM 2.0 connector (module optional)
RTC	1x RTC socket (CR1220 included); 1x RTC 2‑pin
Camera	2x CSI (2-lane 15pin)
Power (Terminal)	DC 12V–24V terminal block (2‑pin)
Power Adapter	19V power adapter (without power cord)
Dimensions (W x D x H)	159mm x 155mm x 57mm
Weight	1.57kg
Installation	Desk, DIN rail, wall‑mounting, VESA
Environment	Operating Temperature: -20 ~ 60°C with 0.7m/s airflow; Operating Humidity: 95% @ 40°C (non‑condensing)
Reliability	Vibration: 3 Grms @ 5 ~ 500 Hz, random, 1 hr/axis; Shock: 50G peak acceleration (11 msec)
Warranty	2 Years

What’s Included

reComputer Industrial J3011 (system installed)	x1
Mounting bracket	x2
DIN rail bracket	x1
Bracket screw	x4
16‑Pin Terminal Block for DIO	x1
19V Power adapter (power cord sold separately)	x1
2‑Pin Terminal block power connector	x1

Power Cord Note

The product includes a power adapter but does not include an AC cloverleaf power cord. Purchase the appropriate AC cloverleaf power cord for your region.

• AC Cloverleaf Power Cord - US
• AC Cloverleaf Power Cord - EU

 

Applications

Application Fields

• AI Video Analytics
• Machine Vision
• Autonomous Mobile Robot (AMR)
• Generative AI

Capable of Bringing Generative AI to the Edge

Build AI agents that summarize, search, and extract insights from video and images using Vision‑Language Models (e.g., LLaVA).

Build Multi‑streams AI Video Analytics

The NVIDIA Multi‑Standard Video Decoder accelerates video decode for SD/HD/UltraHD content to support multi‑stream analytics at the edge.

Fastest Way to Deploy Generative AI and Computer Vision Models

Seeed provides jetson-example projects for one‑line deployment of edge AI workloads including Ollama, Llama3, YOLOv8, and more.

Manuals

• Datasheet
• Schematic
• 3D File
• Assembly Guide
• MTBF Test Report
• Github
• Seeed Nvidia Jetson Product Catalog
• Nvidia Jetson Comparison
• Nvidia Jetson Product Comparison
• Seeed Nvidia Jetson Success Cases
• Seeed Jetson One Pager

ECCN/HTS

HSCODE	8471504090
USHSCODE	8517180050
UPC
EUHSCODE	8471707000
COO	CHINA

Details

Jetson Orin Nano 8GB, 40 TOPS AI performance, fanless design, open-source hardware. Supports robotics, generative AI, video analytics, and offers multiple connectivity options.

reJ3011 Edge AI Device has dual LAN, HDMI, USB 2.0, Nano SIM, DC input, three USB 3.2 ports, COM port, DI/DO/CAN, reset button, and status LEDs. (32 words)

Features an M.2 Key B Mini PCIe slot, control and UART header, fan header, 2-pin RTC connector, 260-pin SODIMM connector, and TPM header on the front; DIP switches, M.2 Key M slot, 2-lane 15-pin CSI interface, RTC socket, optional Wi-Fi/Bluetooth module, and external antenna interface on the rear. Supports diverse connectivity and expansion for edge computing. Designed for high-performance AI tasks with integrated hardware interfaces. Compact and versatile, it enables efficient processing in demanding environments. Ideal for industrial automation, smart cities, and AIoT applications requiring reliable, real-time performance in a rugged form factor. Built to streamline deployment and enhance scalability in edge AI solutions.

No unusual incident occurred in aisle 324 when a shelf collapsed and boxes fell off, blocking the aisle at 3:30 PM.

Nvidia Deepstream car recognition system running on Nvidia-desktop with a performance of 0.73 FPS, processing frames at an average rate of 0.79.

reJ3011 Edge AI supports Llama3, ollama, LLaVA, Whisper, Stable Diffusion, NanoDB.