Open search

Embedded box PC

AI for industry, medical technology, and “smart” cities

Distec expanded its portfolio with the industrial embedded box PC EC-3200 from iBase for applications with artificial intelligence (AI). The Nvidia Jetson TX2-based product with CAN-connection is for industry, medical, technology, and “smart” cities.

Distec is provider for TFT flat screens and system solutions for industrial and multimedia applications (Source: iBase)

The processor combines exceptional speed and energy efficiency with a dual core Denver 2 and a quad core ARM Cortex-A57 processor, said the company. “Compared to the Jetson TX1, the energy efficiency and performance has more than doubled,” explained Thomas Schrefel, Product Manager Embedded at Distec. “This is made possible by the ultra-modern Nvidia Pascal architecture with 256 processing units and up to 1,33 Teraflops. The box PC thus offers AI computing power for edge applications with 8 GiB memory and 59,7 GiB per seconds memory bandwidth.” The product is suitable for use in industrial robots, medical devices, and “smart” cities and for devices that support collaboration in companies.

Harsh industrial environments

The product is designed for a temperature range from -20 °C to +60 °C. It has 8 GiB LPDDR4 memory and 32 GiB eMMC. HDMI, Gigabit Ethernet, two USB 3.0, and two USB 2.0 ports are available as I/O interfaces. On the housing there is a DB9 connector for CAN or EIA-232, a micro USB slot, and an external 10-pin GPIO connection. Two M.2 expansion sockets are available for expansion with optional SSD storage (2280 Key-M) and WAN modules (3042 Key-E). The operating system is based on a specially made Ubuntu 16.04 with Jetpack 3.2.1 and L4T 28.2.

Today's edge- and cloud-based AI products require better computing and video analysis skills in order to perform sophisticated real-time data processing and to overcome latency problems. The box PC is a solution with a fanless design for uninterrupted operation. It takes advantage of the Jetson TX2's GPU-accelerated parallel processing to handle data-intensive and mission-critical workloads.


Publish date