The CI1312 is a high‑performance neural‑network smart voice chip with integrated BNPUV3 processor, CPU, audio codec and multiple interfaces.
The CI1312 is a next‑generation intelligent voice chip developed by 启英泰伦. It integrates the proprietary BNPUV3 neural‑network processor, a 32‑bit CPU, flash and SRAM memory, a low‑power audio codec, and a variety of peripheral interfaces, enabling voice recognition, speaker verification, command‑word learning, voice detection and deep‑learning noise reduction.
Key features include a 220 MHz CPU, 2 MB Flash, 640 KB SRAM, 5 high‑speed GPIOs, 2 UARTs, 1 I²C, 3 PWM channels, and a built‑in power‑management unit. The chip operates from a 3.6 V–5.5 V supply, works in temperatures from –20 °C to +85 °C, and is offered in a compact SOP‑16 package (9.9 mm × 6.0 mm × 1.7 mm). This manual is written in Chinese.
| CPU frequency | 220 MHz |
|---|---|
| Flash memory | 2 MBytes |
| SRAM | 640 KB |
| Audio ADC SNR | ≥95 dB |
| Power‑supply voltage range | 3.6 V – 5.5 V |
| Package | SOP‑16, 9.9 mm × 6.0 mm × 1.7 mm |
| Operating temperature range | -20 °C to +85 °C |
| GPIO pins | 5 |
| UART interfaces | 2 |
| I²C interfaces | 1 |
| PWM channels | 3 |
The chip works from –20 °C up to +85 °C.
It accepts a supply voltage between 3.6 V and 5.5 V.
The device provides five high‑speed GPIO pins.
The integrated audio ADC has an SNR of at least 95 dB.