T-Chip Group
Intelligent hardware website
Innovate intelligent hardware and solutions.
Server Business Website
Server products and efficient solutions.
Consumer Business Website
StationPC NAS products and services.

EC-A1688JD4 Industrial AI PC

Up to 16 TOPS Computing Power

· High-Performance AI Processor

· Deployment of Large AI Models

· Supports Multiple Deep Learning Frameworks

· Multi-Channel Video AI Processing Performance  

· Rich Algorithms with High Practicality

· Full Aluminum Alloy Shell for Heat Dissipation

· Rich Expansion Interfaces

· Provides Development Materials

Manual Wiki Download

Gemma

Llama

Qwen

ChatGLM

EC-A186JD4
EC-A1688JD4 Industrial AI Computer

Powered by the SOPHON BM1688 / CV186 AI processor. BM1688 delivers 16 TOPS INT8, CV186 delivers 7.2 TOPS INT8. Supports mainstream AI models and deep learning frameworks. Features SOPHON SDK for easy algorithm deployment with low migration cost. Supports 16-channel 1080P@30fps decode and 10-channel 1080P@30fps encode. Offers HDMI 2.0, RS485, RS232, and USB 3.0.

High-Performance AI Processor

Powered by SOPHON AI processor BM1688 / CV186, based on ARM Cortex-A53 with a maximum frequency of 1.6GHz, widely applicable in AI inference, computer vision, and other fields.

Private Deployment

Supports private deployment of large-scale parameter models under the Transformer architecture, such as Qwen series, Llama series, Gemma series, DeepSeek-R1 series, ChatGLM series, and other large language models.

Deep Learning Frameworks

Supports traditional network architectures such as CNN, RNN, and LSTM, as well as various deep learning frameworks including TensorFlow, PyTorch, PaddlePaddle, Caffe, and ONNX. Enables pedestrian detection, face detection, face recognition, liveness detection, and other video structuring applications.

Video AI Performance

Supports 16-channel H.264 1080P@30fps video decoding, 10-channel H.264 1080P@30fps video encoding, and full-process processing (decoding + AI analysis) for 16-channel 1080P HD video, meeting the requirements of various AI application scenarios such as face detection in video streams, license plate recognition, and smart cities.

Rich and Practical Algorithms

Supports person/vehicle/non-motorized detection, video structuring, behavior analysis, and more. Features high security and reliability for flexible product development. Comprehensive software/hardware support enables easy AI inference for cloud and edge applications.

Full Aluminum Alloy Shell

Featuring an industrial-grade all-aluminum alloy casing with fanless cooling design, it ensures stable 24/7 operation to meet diverse industrial application requirements.

Rich Expansion Interfaces

Equipped with expansion interfaces including Gigabit Ethernet, HDMI 2.0, RS485, RS232, CAN, USB 3.0, Type-C, TF Card slot, and SIM Card slot to meet peripheral expansion needs in various scenarios.

Development Materials

Provides source code, tutorials, technical documentation, and development tools to lower the development barrier, simplify the development process, and make development more efficient.
Up to 16 TOPS NPU Computing Power

Both EC-A1688JD4 and EC-A186JD4 feature built-in neural network acceleration engine TPU. BM1688 delivers computing power of 32 TOPS (INT4), 16 TOPS (INT8), 4 TFLOPS (FP16/BF16), and 0.5 TFLOPS (FP32). CV186 delivers computing power of 12 TOPS (INT4), 7.2 TOPS (INT8), and 1.5 TFLOPS (FP16/BF16). Capable of intelligent data processing, speech recognition, and image analysis, meeting edge computing AI requirements of most terminal devices.

Application Scenarios

Widely applicable in industry fields such as robotics, drones, smart transportation, edge computing, retail AI, and cameras.

Robots
Robots
Drone
Drone
Smart Transportation
Smart Transportation
Edge Computing
Edge Computing
Retail AI
Retail AI
Smart Network Cameras
Smart Network Cameras
Interfaces
Specifications

EC-A1688JD4

EC-A186JD4

Basic Specifications

SOC

SOPHON BM1688

SOPHON CV186AH

CPU

Octa-core ARM Cortex-A53 @ 1.6GHz

Hexa-core ARM Cortex-A53 @ 1.6GHz

TPU

16T@INT8, 32T@INT4, 4T@FP16/BF16, and 0.5T@FP32 computing power

7.2T@INT8, 12T@INT4, and 1.5T@FP16/BF16 computing power

ISP

Time-sharing multiplexing for up to 6 sensor input videos, with maximum widths of 4608 (non-tile mode) and 8192 (tile mode). Supports Sensor self-band dynamic and 2-frame wide dynamic range, and the maximum performance supports: 12M@30 HDR or 8K@15 SDR or 16M@30 SDR. Support RGB-IR, AI ISP interface, 3A (AE/AWB/AF, 3A control user adjustable). Support fixed mode noise removal, dead pixel correction, shadow correction, lens distortion correction, purple edge correction, Bayer noise reduction, 3D denoising, image edge enhancement, dehazing, dynamic contrast enhancement, image video Mirror, Flip and other functions.

Decoding/ Encoding

Video decoding: H.264 decoding (Max performance: 1920×1080@480fps or 8192×4320@30fps) Video encoding: H.264 encoding (Max performance: 1920×1080@300fps or 8192×4320@15fps) Image codec: JPEG/MJPEG Baseline codec (JPEG codec capability: 1080P@480fps, maximum resolution of 32768×32768)

RAM

8GB LPDDR4 (4GB/8GB/16GB optional)

4GB LPDDR4 (4GB/8GB/16GB optional)

Storage

32GB eMMC (32GB/64GB/128GB/256GB optional)

Storage expansion

1 × TF Card 1 × M.2 M‑KEY (inside the device) - Supports expansion of PCIe NVMe SSD (2242/2260/2280) - Supports expansion of SATA 3.0 SSD (2242/2260/2280) - PCIe NVMe SSD supported by default

OS

Ubuntu

Software Support

Supports private deployment of ultra-large-scale Transformer-based models, including large language models such as Deepseek-R1, Gemma, Llama, ChatGLM, Qwen, and Phi series. Supports traditional network architectures including CNN, RNN, and LSTM, as well as mainstream deep learning frameworks: TensorFlow, PyTorch, PaddlePaddle, Caffe, and ONNX. Supports custom operator development and Docker-based containerized management.

Power

DC 12V (5.5mm × 2.1mm, support 9V~24V wide voltage input)

Power consumption

Normal: 7.2W(12V/600mA), Max: 12W(12V/1000mA)

Normal: 8.5W(12V/710mA), Max: 14W(12V/1170mA)

Size

188.0mm × 88.44mm × 50.65mm

Weight

Net weight: 0.79kg, Total weight with packaging: 1.14kg

Environment

Operating Temperature: -20℃～60℃, Storage Temperature: -20℃～70℃, Storage Humidity: 10%～90%RH (non-condensing)

Interface Specifications

Internet

Ethernet: 2 × RJ45 (1000Mbps) WiFi: Extend WiFi/Bluetooth module via M.2 E-KEY (2230), supports 2.4GHz/5GHz dual band WiFi6 (802.11a/b/g/n/ac/ax) and Bluetooth 5.2 4G: Extend 4G LTE via Mini PCIe (Reused with 5G) 5G: Extend 5G via M.2 B-KEY (Reused with 4G and USB3.0(1), not pasted by default)

Video output

1 × HDMI2.0 (4K@60Hz)

Audio output

1 × 3.5mm Audio jack (Support MIC recording, American standard CTIA)

USB

2 × USB3.0 (Max: 1A; Top: USB3.0(1), reused with 5G; Bottom: USB3.0(2)), 1 × Type-C (Debug)

Others

1 × SIM Card, 1 × Phoenix connector (8Pin, 3.5mm pitch): 1 × RS485, 1 × RS232, 1 × CAN 2.0
Accessories

RM500Q-GL 5G

$ 399.00

EC20 4G-LTE

$ 39.00

WXT14R2201 WiFi Module

$ 15.00

Go to more product
Customization

Firefly team, with over 20 years of experience in product design, research and development, and production, provides you with services such as hardware, software, complete machine customization, and OEM server.

Hardware Solution

Software Solution

Turnkey Solution