ROHM has recently announced the release of new high noise immunity monolithic LED driver, the BD81A76EFV-M, to support vehicle class lighting system requirements. These state-of-the-art MCUs are on the verge of revolutionizing the electronics industry by making AI integration easier across a breadth of applications, including delivering predictive maintenance to tools of all kinds and breathing new life into previous generations of equipment.
The new series of AI MCUs use ROHM s original “Solist-AI™” driver that uses a simple 3-layer neural network algorithm to perform AI learning and inference on-chip so that constant, cloud, or network connection is not necessary. This “edge AI” strategy has major benefits in latency, data privacy and power consumption efficiencies.
Featuring a 32-bit Arm® Cortex®-M0+ core and ROHM’s proprietary “AxlCORE-ODL” AI accelerator, these MCUs achieve AI processing performance that is roughly 1,000 times faster than existing software-based MCUs (at 12MHz operation).
This ultra-fast processing allows for real-time anomaly detection and number output, making them perfect for predictive maintenance in industrial devices such as motors and sensors, while also being ideal for smart home devices too as for utilization in residential facilities.
Offering outstanding low power consumption, the ML63Q253x-NNNxx and ML63Q255x-NNNxx series consume approximately 40mW during AI processing. They also support critical peripherals including CAN FD controllers, 3-phase motor control PWM and dual A/D converters providing broad control and data processing support.
And ROHM plans to support developers aiming to extract the maximum potential of these new AI MCUs. Solist-AI™ Sim -AI Simulation Tools on the Web- It is an AI simulation tool that can be used from the company website, that enables both learning and inference effectiveness to be verified prior to physical use.
Also a full development support environment including an integrated development environment, an AI operation verification simulator, and evaluation boards, helps to accelerate the design and development process.
Mass production of 8 models of low-power TQFP package has started from February 2025, and pre-mass models are available for online purchase, ROHM will rewrite the edge calculation of intelligent, power saving new rules.
This breakthrough is a key step towards the future that may one day see devices that possesses the ability to learn and adapt for themselves, orchestrating their own movements for research and advanced manufacturing (as well as for a new class of intelligent electronics).
