ROHM Semiconductor ML63Q2500 AI-Equipped Microcontrollers
ROHM Semiconductor ML63Q2500 AI-Equipped Microcontrollers (MCUs) provide a network-independent solution for early anomaly detection before equipment failure. This contributes to more stable and efficient system operations by decreasing maintenance costs and the risk of line stoppages. These devices adopt a simple three-layer neural network algorithm to implement ROHM's proprietary on-device AI solution, Solist-AI™. This feature allows the MCUs to independently perform AI learning and inference without cloud or network connectivity. The modules integrate a 32-bit Arm® Cortex®-M0+ processor, ROHM's proprietary AI accelerator AxlCORE-ODL, and a variety of peripheral circuits. The ML63Q2500 MCUs permit real-time operational status monitoring while avoiding network latency issues and security risks. With a low 40mW power consumption during AI processing, the series is ideal for anomaly detection and predictive maintenance in industrial applications.
Features
- Network-independent solution that supports early anomaly detection before equipment failure
- Independently executes both learning and inference without relying on network connection
- Contributes to more stable, efficient system operation by reducing maintenance costs and risk of line stoppages
- Adopts 3-layer neural network algorithm to implement ROHM's proprietary on-device AI solution Solist-AI
- Flexibly adapts to different installation environments and unit-to-unit variations, even within the same equipment model
- Uses ROHM's proprietary AI accelerator AxlCORE-ODL
- Enables fault prediction and degradation forecasting using sensing data in devices
- Delivers approximately 1000 times faster AI processing compared to ROHM's conventional software-based MCUs (theoretical value at 12MHz operation), enabling real-time detection and numerical output of anomalies
- High-speed learning (on-site) at point of installation is possible, ideal for retrofitting into existing equipment
Applications
- Factory Automation (FA) sensors
- Industrial and office equipment
- Robots
- Residential facilities
- Home appliances
- Batteries
- Power tools
- Motors
- Household appliances
- Measurement instrumentation
Specifications
- CPU:
- 32-bit RISC CPU (Arm Cortex-M0+ CPU)
- Arm Thumb®/Thumb®-2 instruction supported
- Serial wire debug port
- Minimum instruction execution time
- 30.5µs (at 32.768kHz system clock)
- 20.83ns (at 48MHz system clock)
- 24 bits x 1 channel, counting by system clock (SYSCLK) (initial clock: LSCLK)
- Internal memory:
- Rewriting program memory area by software
- Background operation (CPU can work while erasing and rewriting to data flash memory area)
- Flash ROM:
- 128/256KB program area
- 8KB data area
- 16KB work memory (data RAM)
- 2.3V to 5.5V operating voltage range
- -40°C to 105°C operating temperature range
- 6x 16-bit timer
- 2x SSIO
- AxlCORE-ODL AI accelerator
- 48MHz (PLL oscillation)/40MHz (crystal oscillation) maximum high-speed operating frequency
- 32.768kHz (internal RC oscillation/crystal oscillation) maximum low-speed operating frequency
- 256KByte ROM capacity in 48-pin package (ML63Q2537)
- 256KByte ROM capacity in 64-pin package (ML63Q2557)
Comparison Chart
Block Diagram
Reference Boards
ROHM Semiconductor RB-D63Q2537 & RB-D63Q2557 Reference Boards
Allow to test the operation of ML63Q2537 and ML63Q2557 32-bit Microcontrollers (MCUs).