Personal Projects

As I work on more and more electronics-based projects, I began thinking about how to deal with solder fumes in my small living space. To address this, I decided to build a fume extractor using some spare three-wire variable-speed computer fans I had. A custom control board based on the Arduino Micro (ATmega32U4) triggers a TC4420 MOSFET driver IC to drive an low-side IRFZ48N N-channel power MOSFET, switching the fans using 33 Hz PWM for better low-speed control at the cost of some audible noise. Additionally, the microcontroller determines the fan speeds, input voltage, and ambient atmospheric conditions (AHT10 via I²C), sending all data to an I²C OLED display. The fan speeds are individually calculated using hardware interrupts triggered by falling edges on the fans' tachometer lines. To remove fluctuations, all data is smoothed using moving averages before being displayed.

The design's emphasis on modularity allows for future modifications and/or expansions. All fasteners thread into brass heat-set inserts to facilitate disassembly. The folding handles double as an angle- and height-adjustable stand, and easily removable front and rear wire grilles keep debris out of the fans and ensure user safety. To improve the product's electrical versatility, I incorporated a USB Power Delivery module, allowing it to work with any available USB power supply or power bank.

Despite my best efforts, the final product certainly is not perfect. The housing containing all the circuitry is extremely horizontally cramped due to poor long-term planning when designing the other mechanical components – allocating even a few more millimeters would have saved a lot of hassle. Furthermore, the legs were originally quite prone to self-loosening despite retrofitting a rubber spacer to increase friction. I later managed to fully address this by adding a two-position clip between the carbon tubes to prevent relative movement in the deployed and stowed positions. On a more positive note though, the triple 140 mm fans are capable of generating a surprising amount of airflow, and the final product certainly fulfills its core functionality, deeming the project an overall success and excellent learning opportunity.

This project is currently in the conceptual design phase. In my experience, the overall design of entry-level oscilloscopes has been unchanged for quite some time, and as such I have identified that the interface can be overwhelming to beginners (especially with small screen sizes), and there can be some difficulties with offloading data for future use. My plan is replace some of the traditional controls with a touch screen and five-way joystick to simplify the control panel and create a more intuitive user experience, especially for the modern generation of electronics beginners. Additionally, wireless streaming abilities will allow for screen mirroring and additional data processing on the user's laptop or other device.

My plan is to make a battery-powered device centered around a Teensy 4.1 microcontroller (ARM Cortex-M7 at 600 MHz) that handles data acquisition and digital processing. The onboard SD card slot and RAM expansion chips help make this possible. The main board will then communicate with an ESP32-C6 SBC for Bluetooth and/or WiFi. I will also implement the usual analog front end (for selecting coupling modes, initial signal conditioning, etc.) along with the aforementioned user interface.

My motivation for this project (beyond wanting an oscilloscope for my own use) is to explore the world of signal processing and high-frequency electronics, branching out from my previous experience with mechatronics and other more "physical" electronics. Although the final capabilities of the product are unlikely to be impressive due to budget limitations (good ADCs get really expensive really quickly!) and initial design concessions (ex. using discrete ICs instead of FPGAs like many commercial models), it will certainly be a great learning experience and something I can be proud of.

One of my first ever major CAD projects, this fully functional model is an early example of my ability to turn my understanding of mechanical principles into functional real-world designs. The design was successfully tested up to 1000 rpm using a power drill on the input shaft. One major issue I encountered was initially having too much clearance between the shift barrel and the selector forks, resulting in rough or entirely missed shifts. This taught me the importance of being intentional and careful when specifying fits between components—fractions of a millimeter can make a huge difference in the operation of even a simple mechanical device like this one. Learning experiences like this are critical takeaways to apply to future projects.