reComputer Industrial RK3576 Overview

Last updated:

Seeed Studio reComputer Industrial RK3576 is a fanless industrial edge AI computer built on the reComputer RK3576 Module. Rockchip RK3576 combines four Cortex-A72 cores at 2.2GHz, four Cortex-A53 cores at 2.0GHz, an Arm Mali-G52 MC3 GPU, and a 6 TOPS INT8 NPU. Standard configurations are D5-0432 with 4GB LPDDR5 / 32GB eMMC and D5-0864 with 8GB LPDDR5 / 64GB eMMC, with Armbian pre-installed.

Isolated RS-485, CAN FD, and digital I/O, together with dual Gigabit Ethernet, USB, and M.2 expansion, connect controllers, sensors, meters, and storage. The aluminum-alloy enclosure uses passive cooling and supports 9–36V DC terminal input, PoE PD, and DIN-rail mounting for industrial vision inspection, data acquisition gateways, and equipment monitoring.

Features

  • Local AI processing: Rockchip RK3576 combines four Cortex-A72 cores, four Cortex-A53 cores, an Arm Mali-G52 MC3 GPU, and a 6 TOPS INT8 NPU. reComputer AI Lab provides model and deployment resources for computer vision, language, vision-language, and speech applications.
  • Isolated industrial I/O: Two RS-485 channels, one CAN FD channel, one digital input, and one digital output connect field devices such as controllers, sensors, and meters.
  • Industrial power and installation: A 2-pin terminal accepts 9–36V DC. One Gigabit Ethernet port supports PoE PD. The fanless enclosure and DIN-rail mounting suit control-cabinet installations.
  • Storage and wireless expansion: An M.2 Key-M 2280 slot with PCIe 2.1 ×1 supports an NVMe SSD or a compatible AI accelerator. miniPCIe and nano SIM provide expansion for optional cellular and other wireless modules.
  • System supervision: A hardware watchdog with a 1–255-second range and a high-accuracy RTC support device supervision and timekeeping.

Specifications

Product configurations

ParameterD5-0432D5-0864
SKU100050947100094701
RAM4GB LPDDR58GB LPDDR5
eMMC32GB64GB

Processing, connectivity, and expansion

ItemSpecification
SoCRockchip RK3576
CPU4 × Cortex-A72 at 2.2GHz + 4 × Cortex-A53 at 2.0GHz
GPUArm Mali-G52 MC3
NPU6 TOPS (INT8)
Operating systemArmbian pre-installed
Ethernet2 × Gigabit Ethernet; one port supports PoE PD
RS-4852 × isolated channels
CAN1 × isolated CAN FD channel
Digital I/O1 × isolated DI; 1 × isolated DO
USB2 × Type-A USB 3.0; 1 × Type-C for OTG and DisplayPort
Display1 × Display output; DisplayPort 1.4 via USB Type-C
M.21 × M.2 Key-M 2280, PCIe 2.1 ×1, for NVMe storage or a compatible AI accelerator
miniPCIe1 × slot for optional 4G LTE, LoRaWAN, or Wi-Fi HaLow modules
SIM1 × nano SIM slot
Wi-Fi / BluetoothOnboard Wi-Fi 6 and Bluetooth 5.4
Audio1 × 3.5mm headphone jack with microphone input
Watchdog / RTCHardware watchdog, 1–255 seconds; high-accuracy RTC
Buttons / indicatorsReset and MaskROM; PWR, USER, and ACT LEDs
Buzzer1
Video decodingUp to 8K@30fps, H.265 / H.264 / AV1 / AVS2
Video encodingUp to 4K@60fps, H.265 / H.264

Power and deployment conditions

ItemSpecification
DC input9–36V DC through a 2-pin terminal block
PoE1 × PoE PD input
Cooling / enclosureFanless passive cooling; aluminum alloy
MountingDIN rail
Operating temperature0–60°C with 0.7 m/s airflow
Storage temperature−20–90°C
Operating humidity10–95% RH
Warranty2 years

4G LTE, LoRaWAN, Wi-Fi HaLow, SSDs, and AI accelerators are optional expansion modules selected for the deployment.

Application Scenarios

  • Industrial vision inspection: Process camera data locally for object detection, classification, and inspection.
  • Data acquisition and gateways: Collect device data over RS-485, CAN FD, and digital I/O, then forward it through Ethernet or a configured wireless connection.
  • Equipment and energy monitoring: Record machine status, meter data, and operating conditions for local analysis and supervisory systems.