A complete guide to enabling hardware acceleration and setting up an automated conversion service on your Rockchip-powered SBC. This post will walk you through installing the necessary drivers and deploying a systemd service to watch for and convert video files automatically.
Part 1: Installing Hardware Acceleration Libraries
For ffmpeg to use your Rockchip VPU (Video Processing Unit), it needs the correct driver libraries. There are two primary methods to get these essential components installed on your system.
Method A: Using a PPA (Recommended)
This is the easiest and most reliable method as it uses a pre-built package archive.
Important: After either installation method, reboot your SBC to ensure all drivers are loaded correctly.
Part 2: The Automation Scripts and Service
This service automatically finds and converts your files. It consists of a conversion script and two systemd files to manage and schedule it.
1. The Conversion Script
This script finds HEVC files, converts them using hardware acceleration, and deletes the original upon success. Save it as /home/manupa/bin/convert_hevc.sh.
#!/bin/bash
WATCH_DIR="/home/manupa/Downloads/SD/TVSeries"
LOG_FILE="/home/manupa/Downloads/converter.log"
log() { echo "$(date +'%Y-%m-%d %H:%M:%S') - $1" >> "$LOG_FILE"; }
log "--- Starting conversion scan ---"
find "$WATCH_DIR" -type f \( -iname "*hevc*.mkv" -o -iname "*h265*.mkv" \) | while IFS= read -r i; do
o="$(dirname "$i")/$(basename "$i" .mkv).mp4"
log "Found: $i"
if [ -f "$o" ]; then
log "Skipping, output exists. Deleting original: $i"
rm "$i"
continue
fi
log "Converting to: $o"
ffmpeg -c:v hevc_rkmpp -i "$i" -c:v h264_rkmpp -c:a copy "$o"
if [ $? -eq 0 ]; then
log "SUCCESS. Deleting original: $i"
rm "$i"
else
log "ERROR: ffmpeg failed on $i"
fi
done
log "--- Scan finished ---"
2. Systemd Service
Tells systemd what to run. Save as hevc_converter.service.
[Unit]
Description=HEVC to H264 Conversion Service
[Service]
Type=oneshot
ExecStart=/home/manupa/bin/convert_hevc.sh
3. Systemd Timer
Tells systemd when to run it. Save as hevc_converter.timer.
Documenting the journey of building a robust data logging and visualization solution.
You might remember my previous project where I built a miniature weather station. In that setup, I displayed live sensor readings using an MQTT broker app on Android. (If you missed it, you can check out the details here: Solar Weather Station with MQTT and the MQTT client I used: MQTT Client on Google Play).
While this initial approach worked for live data, I quickly ran into a couple of significant limitations:
Constant Internet Dependency: The system required a continuous internet connection to view the readings.
No Historical Data: Crucially, the MQTT setup didn't store any past data, making trend analysis impossible.
To overcome these challenges, I needed a more robust solution: a web application with a dedicated data logger.
What's a Data Logger and Why Did I Need One?
In my previous setup, sensor data was published to an MQTT broker and displayed instantaneously. However, this data wasn't being saved anywhere. To address this, I needed to:
Capture the Data: Subscribe to the MQTT topics.
Store the Data: Save these readings into a persistent database.
For this, I developed a Python script that subscribes to the relevant MQTT topics and logs the incoming sensor data into an SQLite database. Each entry is timestamped with the current UTC time, ensuring that the logged data is traceable and clear for analysis.
Building the Web Application: Tech Stack and Features
With the data logging mechanism in place, I built a web interface using:
Backend: Flask (a Python web framework)
Frontend: Basic HTML, CSS, and JavaScript
This web application provides several key features:
Average Sensor Readings: Displays summarized information for various sensors.
Historical Data Visualization: Allows users to view past data, classified by minutes, hours, and days.
Data Export: The entire database can be downloaded as a CSV file for offline analysis or use by customers.
Navigating Deployment Challenges: From VMs to App Services
I tried deploying the system in a VM in Azure, but I faced several issues:
Lack of HTTPS out-of-the-box: Securing the application required manual SSL certificate configuration.
No Friendly URL: Access was via a public IP address, which isn't user-friendly.
Security Concerns: Managing security on a VM can be complex.
Manual Management: Updates and maintenance were time-consuming.
Performance Issues: Lower-tier (SKU) VMs often froze under load.
After discussing these issues with friends, (Thank you Tharindu 😉) the clear recommendation was to containerize the application using Docker and deploy it as an Azure App Service. This approach offered a much smoother path.
Azure App Services provides various ways to run web applications, including options for static web apps, web apps with databases, and WordPress sites. Given my need for a dynamic application with features like persistent storage and a separate, continuously running data logger, I opted for the Web App service.
To streamline deployment, I packaged the entire project, including the Flask application and the Python data logger, into a Docker image. This image was then pushed to GitLab's Container Registry and made public, allowing Azure App Services to easily pull and deploy it.
You can select your container from different sources when setting up the App Service.
I'm thrilled to say the system is now running smoothly! The final website is accessible, secure, and much easier to manage.
This project was a fantastic learning experience. I now have a secure (HTTPS-enabled) application with minimal maintenance overhead, thanks to Azure App Services. It also benefits from inbuilt scaling capabilities and protection mechanisms like Azure Front Door (if configured).
Future Enhancements 🛠️
While the current version is a significant step up, there are a couple of features I plan to add next:
User Authentication: Implement a sign-in page to manage access for different users.
Integrated Database Solution: Migrate data storage from the SQLite database within the Docker container (which can be difficult to access for direct downloads) to Azure's inbuilt database services. This will make data management and backups more robust and accessible.
Solar Weather Station with Supercapacitors | A Modern Approach
Powering the Future: Integrating a Miniature Weather Station with Solar and Supercapacitors
Published on: May 10, 2025
In our ongoing exploration of innovative solutions for sustainability, I've embarked on a project to integrate a miniature weather station with solar power. To ensure reliability and longevity, we've opted for the use of a supercapacitor as our backup energy storage solution. This post details the journey, the technology, and the performance insights.
Starting the Journey with Supercapacitors
I began by testing the application of supercapacitors in various projects. This experience has been invaluable as I now want to apply what we’ve learned to make our weather station more sustainable and efficient.
Converting Our Miniature Weather Station to Solar-Powered Operation
The next phase of this project involves converting the miniature weather station to run on solar power. By integrating a supercapacitor, which acts as a buffer between the variable output of solar panels and the constant demand from our weather station sensors, we can achieve reliable and consistent readings even during periods of low sunlight.
Why Supercapacitors?
2.7V 30F Supercapacitor
Supercapacitors offer several advantages over traditional batteries such as lithium batteries. Here are the key reasons why they are well-suited for our solar-powered weather station project:
High Power Density
Quick Current Delivery: Deliver high currents quickly, suitable for intermittent and peak loads.
Immediate Response to Demand: Provide immediate power when needed, ensuring consistent readings during low sunlight periods.
Long Cycle Life
Improved Efficiency: Maintain efficiency better under frequent discharges and recharges.
Reliability in Variable Conditions: Robust performance in varying environments (e.g., hot direct sunlight), ensuring consistent operation over time.
Environmental Considerations of Traditional Batteries
While supercapacitors offer benefits, it's also worth noting the environmental impact of traditional battery solutions they can help mitigate:
E-waste Concerns: Traditional batteries can pose significant e-waste issues due to toxic components (like heavy metals) and high recycling costs or complexities.
Resource Intensive: The manufacturing of many battery types requires substantial resources (e.g., lithium, cobalt), the extraction and processing of which can have considerable environmental implications.
Opting for solutions like supercapacitors, especially in applications where their characteristics are a good fit, can contribute to reducing these environmental burdens.
The Solution: A Block-by-Block Breakdown
My project breaks down into several key stages, utilizing the following main components:
ESP32C3 Supermini: The microcontroller brain.
BME680 Module: The environmental sensor.
TPS63802 Buck Boost Converter Module: For stable power delivery.
Solar Panel (5V 200mA): For energy harvesting.
Supercapacitors (2x 2.7V 30F in series): For energy storage.
The process involves:
Energy Harvesting: Capturing sunlight using a solar panel.
Energy Storage: Storing the harvested energy using supercapacitors.
Power Management: Providing a stable voltage to the microcontroller and sensor despite fluctuating storage voltage.
The Brain: An ESP32C3 microcontroller handling sensor reading, WiFi, and data transmission.
The Sensor: A BME680 measuring temperature, humidity, pressure, and gas resistance.
The Software: Firmware running on the ESP32C3 to control everything.
Let's look at each part in more detail.
1. Harvesting Sunlight: The Solar Panel
The energy source for this project is a 5V 200mA solar panel. This is a standard, relatively small panel suitable for low-power applications. It captures sunlight and converts it into electrical energy.
2. Storing the Energy: Supercapacitors
Instead of a traditional battery, I'm using two 2.7V 30F supercapacitors connected in series. Supercapacitors are fantastic for this kind of application because they can charge and discharge very quickly and tolerate a huge number of cycles compared to batteries.
Series Connection: Connecting the two 2.7V caps in series allows them to handle a higher total voltage (up to 5.4V). However, connecting capacitors in series reduces the total capacitance – in this case, two 30F caps in series behave like a single 15F capacitor.
Charging Protection: A diode is connected in series between the solar panel and the supercapacitors. This is crucial! It acts as a one-way valve, allowing current to flow from the solar panel to the supercapacitors, but preventing the stored energy in the supercapacitors from flowing back into the solar panel when the sun isn't providing enough voltage (like at night).
Balancing Act: When connecting capacitors (especially supercapacitors) in series, it's essential to ensure the voltage is shared equally across them. I've implemented voltage balancing using 4.8 MegaOhm resistors connected in parallel with each supercapacitor. These resistors help bleed off excess voltage, allowing the voltage to equalize. For more on this, see resources like the Analog Devices document on supercapacitor balancing.
3. Stable Power for the Electronics: The Buck-Boost Module
The voltage stored in the supercapacitors will fluctuate (from near 0V up to ~5.4V). Microcontrollers and sensors need a stable operating voltage. This is where the TPS63802 Buck-Boost module comes in. It takes the fluctuating voltage from the supercapacitors and converts it into a stable 3.3V output, ideal for the ESP32C3 and BME680. (The module can also be set to 4.2V or 5V).
TPS63802 Buck-Boost Converter Module.
4. The Brains: ESP32C3 Microcontroller
The core of the system is the ESP32C3, a modern, low-power microcontroller with built-in WiFi. It's responsible for:
ESP32C3 Supermini.
Initializing and reading data from the BME680 sensor.
Managing WiFi connectivity.
Formatting the sensor data.
Publishing the data to an MQTT broker.
Potentially managing power modes (area for future optimization).
5. The Sensor: Bosch BME680
For environmental data, I chose the BME680 sensor. This single module from Bosch measures:
BME680 Environmental Sensor.
Temperature
Relative Humidity
Barometric Pressure
Gas Resistance (related to air quality)
The Software: Firmware and Libraries
Bringing the hardware to life requires robust software. The firmware running on the ESP32C3 is based on an open-source project specifically designed for interfacing with the BME680 and connecting via WiFi/MQTT. The code I'm using is from this repository: manupawickramasinghe/bme680-wifi-sensor-firmware.
This firmware is built using the ESP-IDF framework (Espressif IoT Development Framework) and leverages an open-source driver for the BME680: gschorcht/bme680-esp-idf.
Here's a summary of the key features provided by this firmware:
Multi-Interface Support: Communicates with BME680 via I2C or SPI.
ESP32/ESP8266 Support: Compatible with various ESP platforms.
WiFi Connectivity: Includes SmartConfig for easy provisioning.
MQTT Data Publishing: Sends sensor readings to an MQTT broker.
Automated Builds and Testing: CI/CD using GitHub Actions for code quality.
Structured Data Output: Formats sensor readings into JSON for MQTT.
This firmware provides a solid foundation, handling the complexities of the BME680 interface, WiFi connectivity, and data transmission.
How It All Comes Together
In summary, sunlight hits the solar panel, its energy is directed via the diode into the supercapacitors for storage (kept balanced by resistors). The TPS63802 buck-boost takes the variable voltage from the supercaps and outputs a stable 3.3V. This 3.3V powers the ESP32C3 and the BME680 sensor. The ESP32C3, running the custom firmware, reads data from the BME680, connects to the local WiFi network, and publishes the environmental readings as a JSON message to an MQTT broker.
The assembled electronics package, ready for deployment.
This creates a self-sufficient system capable of monitoring temperature, humidity, pressure, and air quality, powered entirely by the sun and stored energy in the supercapacitors.
Performance Insights: What the Data Tells Us
Analyzing the collected data revealed a lot about the system's behavior:
Temperature Trends: We see the expected daily warming/cooling and the sensor's fine resolution (small ±0.02 °C noise).
Location Impact: A significant peak spike up to ~46°C highlighted a major issue: the sensor wasn't measuring ambient air, but the heat from hot concrete it was initially placed on. Relocating helped mitigate this.
Power Fluctuation Noise: Subtle temperature jitter was more apparent in the morning and evening, likely linked to power instability during low light/charging transitions.
Crucially, the data reporting frequency varied significantly based on light conditions:
Morning (7 AM+): Starts slow (~15 min/message), increasing to ~5 min/message as sunlight intensifies.
Mid-day (10 AM+): Becomes very frequent (almost every second) under full sun. Clouds cause it to drop back to 5-15 mins.
Evening (After 5 PM): Frequency drops drastically (5-15 mins -> 30 mins), stopping completely after 6:30/7 PM as supercapacitors deplete without solar input.
This changing frequency is a direct indicator of the system reacting to available solar power and supercapacitor voltage. More light means more power, allowing the ESP32 to operate and transmit more often. Less light means less power, forcing it to slow down or eventually enter deep sleep until sufficient charge is regained.
Current Limitations & What's Next
While the project is operational and demonstrates the concept effectively, I've identified a few areas for improvement and future development:
Sensor Accuracy: The BME680 is sensitive to heat generated by the ESP32C3 itself, which can affect temperature and humidity readings if not properly isolated.
Future: Design an enclosure that physically separates the sensor from the main microcontroller board, ensuring better airflow around the sensor.
Placement Heat: Initial tests showed that heat absorbed and radiated from the surface the device sits on (e.g., concrete) can significantly skew readings.
Solution: The device is now mounted on a wooden pole to minimize ground heat interference and improve air circulation.
Power Efficiency: The current firmware isn't fully optimized for ultra-low power consumption, potentially draining the supercapacitors faster than ideal during prolonged low-light conditions.
Future: Implement aggressive ESP32C3 Deep Sleep modes, optimize WiFi connection/disconnection times, minimize active mode duration, and refine supercapacitor balancing for minimal quiescent current.
Solar Charging Circuit: The current simple diode-based charging is functional but incurs a voltage drop.
Future: Explore more efficient solar charging ICs (e.g., MPPT controllers suitable for low power) to maximize energy harvesting and reduce losses.
Breathing New Life: Building a Home Server from Old Laptops
Breathing New Life: My DIY Home Server Project! 💻➡️🏠
Turning e-waste into a useful home lab with a bit of tinkering.
The Spark ✨: Old Laptops Get a Second Chance
Recently, a friend gifted me some old laptops. Instead of letting them gather dust or become e-waste, I saw an opportunity! 💡 Why not build a home server? It's a great way to learn, host services, and make use of hardware that might otherwise be discarded.
The challenge? These laptops came barebones – no SSD, no RAM, and crucially, no power adapter! But where there's a will (and some spare parts), there's a way.
The Hardware Hustle 🛠️: Powering Up!
The biggest hurdle was power. Luckily, I had a nifty USB-C PD Fast Charger Decoy Board (100W capable!) lying around. These little boards can negotiate specific voltages from USB-C Power Delivery chargers.
The laptop needed 19V via its barrel jack. I configured the decoy board's DIP switches to request 20V from my USB-C charger (close enough!). Then came the slightly nerve-wracking part: soldering! 👨🏭
Identified the positive (+) and ground (-) pins on the laptop's power input.
Soldered the decoy board's output wires directly to these points.
Carefully cut a small slot in the laptop's casing to neatly house the board.
Used strong double-sided tape to secure the board inside.
Plugged in the USB-C charger, held my breath... and success! 🎉 The laptop powered on, drawing juice through the new setup. Phew!
Next, I installed some spare DDR3 RAM and an old SATA SSD I had available. Hardware complete! ✅
The Software Stack 🐧🐳: Ubuntu & Docker
With the hardware sorted, it was time for the operating system. I opted for Ubuntu Server – it's stable, widely supported, and great for server tasks. Installation was straightforward via a USB drive.
To manage the various services I wanted to run, I decided to use Docker. Containers make it super easy to install, run, and manage applications in isolated environments. No more dependency nightmares! 🐳
Installing Docker on Ubuntu:
# Update package list
sudo apt update
# Install prerequisites
sudo apt install -y apt-transport-https ca-certificates curl software-properties-common
# Add Docker's official GPG key
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
# Add Docker repository
echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
# Install Docker Engine
sudo apt update
sudo apt install -y docker-ce docker-ce-cli containerd.io
# Add your user to the docker group (to run docker without sudo - log out/in after)
sudo usermod -aG docker ${USER}
# Verify installation
docker --version
Note: You'll need to log out and log back in for the group change (`usermod`) to take effect.
My Container Crew 🚀: Services Running
Here are the initial containers I set up:
🗄️ ArchiveBox (Good Karma Kit)
This awesome tool creates local, browsable archives of websites. Perfect for saving articles, documentation, or anything important online before it disappears. The Good Karma Kit aspect seems focused on archiving public interest content.
Access it at `http://[your-server-ip]:8000`. Replace `~/archivebox_data` with your desired host path for data storage.
🚢 Portainer
Managing Docker containers via the command line is fine, but Portainer provides a fantastic web UI. It makes it easy to view logs, manage containers, networks, volumes, and more. Highly recommended for managing your Docker environment!
Docker Run Command (Portainer CE):
# First, create a volume for Portainer data
docker volume create portainer_data
# Then, run the Portainer container
docker run -d \
-p 8000:8000 \
-p 9443:9443 \
--name portainer \
--restart=always \
-v /var/run/docker.sock:/var/run/docker.sock \
-v portainer_data:/data \
portainer/portainer-ce:latest
Access it at `https://[your-server-ip]:9443` (HTTPS) or `http://[your-server-ip]:8000`. You'll set up an admin user on first access.
🎬 Plex Media Server
No home server is complete without a media solution! Plex organizes your movies, TV shows, music, and photos, allowing you to stream them to virtually any device, anywhere. It scans your media folders and automatically fetches metadata and artwork.
Access the setup wizard at `http://[your-server-ip]:32400/web`. Make sure the paths you map (`/path/to/...`) exist on your host system and have correct permissions.
🏠 Home Assistant
Home Assistant! It's an incredibly powerful open-source home automation platform. You can integrate smart devices, create automations, track sensors, and build amazing dashboards.
Access it at `http://[your-server-ip]:8123`. Replace `/path/to/homeassistant/config` with your desired config path. `--network=host` is often needed for device discovery.
📊 btop++ (Resource Monitor)
While not typically run as a Docker container, `btop++` is an excellent TUI (Text User Interface) resource monitor. It gives a detailed, real-time view of CPU, memory, disk, and network usage right in your terminal. Very handy for seeing how the server is performing!
Installation Command (Ubuntu):
sudo apt update
sudo apt install -y btop
Usage:
btop
Smart Scheduling ⏰: Working Around Data Caps
My ISP offers unlimited data during off-peak hours (midnight to 7 AM). To take advantage of this for potentially data-intensive tasks (like ArchiveBox fetching sites), I set up a cron job. Cron is a time-based job scheduler in Unix-like operating systems.
I scheduled tasks to automatically stop certain containers at 7 AM and start them again at midnight. This helps manage bandwidth usage and costs effectively. 💰
Example Cron Job (Edit with `crontab -e`):
# Stop ArchiveBox container at 7:00 AM daily
0 7 * * * docker stop archivebox
# Start ArchiveBox container at 00:00 AM (midnight) daily
0 0 * * * docker start archivebox
You'd add similar lines for any other containers you want to schedule.
What's Next? 🤔 More Containers!
This is just the beginning! I'm excited to explore more self-hosted applications. What other containers do you recommend for a home server setup?
Some popular ideas include:
Nextcloud ☁️: Your own private cloud for files, calendars, contacts, and more (like Google Drive/Dropbox).
Pi-hole 🚫: Network-wide ad blocker. Say goodbye to most ads on all your devices!
