It’s becoming increasingly common for single-board computers to ship with Kendryte’s popular K210, and Sipeed’s latest MAIX Nano M1n with PCIe M.2 and USB Type-C form factor is not left behind as it ships with the K210 embedded along with a 128MB of SPI flash and a 24-pin edge connector.
The K210 AI chip includes a powerful CPU that is based on a dual-core 64-bit RISC-V processor, with up to 400 MHz frequency and a custom convolutional NPU whose TOPs parameter is significant to the chip’s AI performance. It is designed to accelerate edge neural networking workloads with an accelerator that has its own memory alongside 6 MB RAM for the dual-core CPU and an additional 2 MB for the NPU coprocessors. It also has a good ability to realize Face detection, FFT spectrogram, and Object recognition applications.
Based on K210, the M1n which is compatible with deep learning frameworks can be programmed using the Arduino IDE, Platform.io, or MicroPython. For those looking to do more with its part, the MAIX Nano M1n includes an M.2 interface by which users can connect the Type-C downloading board that helps to download the code to the module. The downloading board has pins of 2.54 mm that are directly connected to the pins from the K210 chip for users to add more DIY functions.
Some features of the Sipeed MAIX Nano M1n include:
- Onboard camera DVP carrier: 224pin 0.5mm pitch FPC carrier, DVD – 1.3, AVDD – 3.0V
- Debugging support: High-speed JTAG and UART interface for debugging
- High-performance microphone array processor
- Supports a valid voice stream output
- Through-hole headers for 16 GPIO pins plus LCD display connectivity
- Supports 12-bit, 16-bit, 24-bit, 32-bit precision input audio signal and an audio input up to 192K sample rate.
- Supports multi-channel raw signal direct output
- Has 16-bit internal audio signal processing accuracy
- More than 300mA external current supply-demand @ 5V
- Temperature range: -30C to 85C
