DALLAS - Texas Instruments (TI) MSPM0G5187 mixed-signal microcontrollers (MCUs), now available from Mouser Electronics, integrate a TinyEngine neural processing unit (NPU) with an Arm Cortex-M0+ MCU for edge-AI inference, sensing, and control applications.

The MSPM0G5187 operates at up to 80 MHz and provides up to 128 KB of embedded flash memory with error correction code (ECC) and up to 32 KB of SRAM with ECC protection or hardware parity. The flash memory is organized into two main banks with address-swap support for field firmware updates.

The integrated TinyEngine NPU is a hardware accelerator designed for machine-learning inference, particularly deep convolutional neural networks. TI specifies up to 2.56 GOPS at 80 MHz, with support for 8-, 4-, and 2-bit weights and 8- and 4-bit data. The NPU operates autonomously from the main CPU and supports convolutional, pointwise, depthwise, pooling, and residual neural-network layers.

TI identifies applications for the NPU including time-series classification, arc-fault detection, motor-fault detection, and electrocardiogram processing. The company also provides an application note specifically addressing edge-AI applications with the MSPM0G5187.

The MCU combines the NPU with mixed-signal peripherals intended for sensing and control. The device includes a 12-bit, 1.6-MSPS analog-to-digital converter (ADC) supporting up to 26 external channels, plus a high-speed comparator with an integrated 8-bit reference digital-to-analog converter (DAC). The MCU also incorporates a 12-channel direct memory access (DMA) controller and flexible timer peripherals.

For connectivity, the MSPM0G5187 includes four configurable UNICOMM serial interface modules that can support combinations of up to three UARTs, two I²C interfaces or two SPI interfaces. A USB 2.0 full-speed interface includes an integrated physical layer, clock source, and termination resistors and supports host operation and crystal-less device operation. A digital audio interface supports I²S, TDM, left-justified and right-justified formats.

For more information, visit https://www.mouser.com/.