Articles taggedTech

A walkthrough of ordering two PCBs designed in KiCad from JLCPCB, covering generating and uploading the order files, choosing manufacturing options and a stencil, replying in English to a post-order confirmation email, payment methods and actual cost, and the delivery timeline.

The left click on my Logitech MX Ergo S trackball started double-clicking, so I repaired it by replacing the tact switches. This post records the tools needed for disassembly, how to choose the right switch, removing the main board, and the desoldering and soldering steps, plus the differences from the older MX Ergo and pitfalls when buying parts.

Connected an SCD41 CO2 sensor to M5StackChan as a built-in MCP tool. The AI reads live CO2 levels in conversation and suggests ventilation when levels exceed 1000 ppm. Covers wiring, firmware, and testing.

Learn how to self-host Xiaozhi AI server with Ollama to run M5StackChan AI.AGENT locally for private voice conversations without cloud AI.

This article explains how to add custom MCP servers to M5StackChan AI.AGENT for external tools, local services, and AI workflows.

A walkthrough of M5StackChan setup, covering unboxing, smartphone pairing, Wi‑Fi configuration, AI conversations, and Avatar mode using the stock firmware.

Driving a WS2812B (NeoPixel) RGB LED from a CH32V003 using SPI and DMA, reusing the ch32fun driver header in a MounRiver Studio Ⅱ project. Covers wiring, the two register-name fixes for the official SDK, and a sample that cycles through six colors.

Generate a 50 Hz PWM signal with TIM1 on the CH32V003 microcontroller using the official SDK in MounRiver Studio Ⅱ and drive an RC servo motor (MG90D), with reusable servo.h/servo.c code and oscilloscope verification of the 1.5 ms pulse.

Setting up the CH32V003, a low-cost RISC-V microcontroller, in the VSCode-based MounRiver Studio Ⅱ: installing the IDE, creating a project, wiring a test circuit with the WCH-LinkE, enabling SDI printf debugging, and blinking an LED.

An introduction to the technology stack and architecture behind the rpine lab homepage (rpine.net). The frontend uses Next.js, React, and Tailwind CSS, while the infrastructure runs on Cloudflare Workers, with CI/CD pipelines powered by GitHub Actions and Cloudflare Workers' built-in build features.