A new 32-bit RISC-V microcontroller combines 200MHz processing, vector extensions and optional external memory, targeting machine learning and digital signal processing workloads.
WCH Electronics has introduced the CH32V407 microcontroller, a new member of its 32-bit RISC-V range designed to provide higher performance for embedded applications. The device runs at up to 200MHz and incorporates vector extensions intended to accelerate machine learning and digital signal processing tasks.
Based on the company’s QingKe RISC-V core, the industrial-grade interconnected microcontroller is designed for zero-wait operation at its 200MHz clock speed. Its vector instruction capabilities are intended to support accelerated computing and parallel processing, potentially making it suitable for increasingly demanding edge applications.
The CH32V407 includes 512KB of flash storage and 200KB of on-chip static RAM. An optional version can add up to 8MB of external PSRAM, providing additional memory capacity for workloads that require larger working datasets.
Connectivity and peripheral options are extensive. The device supports up to 77 general-purpose input/output pins, with external interrupts, alongside two USB 2.0 high-speed interfaces capable of 480Mb/s throughput. Other interfaces include Fast Ethernet, 10 USART serial interfaces, CAN, SDIO, I²C, I²S and three SPI interfaces.
The microcontroller also integrates analogue and display-oriented capabilities, including two ADCs, two DACs, a 150MHz DVP digital video interface and an LCD-TFT display connection. A programmable voltage monitor, hardware random-number generator and multiple low-power operating modes are also provided.
For developers, the combination of vector processing and expanded memory is particularly relevant to edge machine learning and DSP applications, where processing performance and memory availability can constrain compact embedded systems.
The CH32V407 is priced at $2.40 per unit in five-unit quantities, according to the source. Evaluation boards are also available, with pricing starting at $12.90, providing a route for developers to assess the platform before integrating it into their own designs.
