Electronics Weekly Electronics Design & Components Tech News
We’re talking support for up to four cameras, for machine vision or robotics or smart HMI…
Priced at around $70, it is built around a Texas Instruments AM67A vision processor.
Specifically, there is also a quad-core 64-bit Arm Cortex-A53 CPU subsystem running at 1.4GHz. An Arm Cortex-R5 subsystem for low-latency I/O and control. And a GPU with video and vision accelerators. Finally, there is a dual general-purpose C7x DSP with Matrix Multiply Accelerator (MMA) capable of the afore-mentioned 4 Top/s (TOPS stands for trillions or tera operations per second, in the context of AI operations).
For memory, as well as a microSD card socket, there is 4GB of LPDDR4
Networking, for WiFi 6 and Bluetooth 5.4 (with BLE), is via the BeagleBoard.org BM3301 module based on TI CC3301.
In terms of interfacing, there’s a PCI-Express Gen3 single lane controller, Gigabit Ethernet and USB3.1-Gen1 ports. General I/O includes:
- 2x MIPI CSI
- PCI-Express Gen3 x 1 interface (requires external adapter)
- 4x USB3 (5Gbps) type-A host ports
- 1x USB2 (480Mbps) type-C device port and power input
- 40-pin expansion header
You can read read more about the BeagleY-AI on the beagleboardorg website. Note that the beagleboard hardware design is open-source (although the name is copyright and its use requires payment).
For more context on the board, you can read our Technoogy Editor’s news article on its release. He highlights, for example, that it is in Raspberry Pi form factor, and uses a Debian XFCE Linux distribution, available through GitHub.
Where can you buy it from? Suggested sellers include (alphabetically): DigiKey, Element14, Mouser Electronics and Seeed.
Eee. When I were a lad, we were grateful for just one GOP… Four trillion operations per second? You don’t know you’re born…
