Original pitch: a desktop control panel with switches, knobs, and indicator LEDs for managing parts of my room, such as lighting and window curtains
Given the time left in the semester, narrowed the pitch to one feature: a knob that controls the brightness of an LED strip mounted behind my monitor
Key Parts
Arduino Nano: a small microcontroller board that runs the code and controls the other components
Potentiometer: a knob that changes its voltage output as you turn it
MOSFET: an electronic switch that lets a small signal control a much larger flow of power
PWM (pulse-width modulation): switching power on and off very quickly; fast enough switching looks like smooth dimming instead of flicker
Breadboard: a board for wiring circuits together without soldering
USB pigtail cables: short USB cables with bare wire ends for connecting into the MOSFET's screw terminals
The Progress
The circuit has two sides: a low-power side that reads a potentiometer, and a high-power side that runs the LED strip off a wall charger through a MOSFET
A microcontroller pin (max. 40mA, avg. 20mA) cannot supply enough current to power an LED strip (approx. 100mA - 3A) on its own, so the MOSFET sits between the Nano's control signal and the strip's power source
The Nano placed across the breadboard's center trench, keeping its two rows of pins electrically separate.The potentiometer's outer pins tied into the breadboard's power rails, with its middle pin wired straight to the Nano's A0 input.
The MOSFET's input side wired to the wall charger, before the output side was connected to the LED strip.
Wired the potentiometer to 5V, ground, and an analog input pin, and the MOSFET's control pins to ground and a PWM output pin
Firmware reads the potentiometer's position, turns it into a PWM signal, and sends that to the MOSFET, which sets the strip's brightness
The full breadboard circuit: Nano, potentiometer, and MOSFET all wired together.
Successes and Failures
MOSFET signal pins: bare through-holes instead of screw terminals. I bent each jumper wire through two of the three holes for PWM and ground so it would hook onto the module. Shaping the wires securely was harder than expected.
"Not dimmable" LED strip: the listing said the strip was not dimmable, which refers to standard voltage dimmers. The PWM switching method I used works fine.
Clone-board bootloader: the Nano is an ELEGOO clone. A troubleshooting source told me to select the older "ATmega328P (Old Bootloader)" option, but that made every upload fail; switching to the standard, newer bootloader setting is what actually fixed it.
Unstable brightness: the strip pulsed brighter about twice a second at lower settings, pointing to a noisy signal rather than a bad strip. Fixed by reseating a loose ground connection and adding a digital low-pass filter, the software equivalent of an RC low-pass filter. Each update keeps 90% of the previous smoothed value and adds 10% of the latest reading.
Reversed knob direction: turning the knob dimmed the strip backward from what felt intuitive. Fixed in software by reversing the output mapping instead of rewiring the potentiometer.
Success: the core wiring, including the potentiometer input and the high-power charger-to-strip path, worked as planned on the first attempt.
The Outcome
Turning the knob smoothly raises and lowers the LED strip's brightness in real time, with no visible flicker or lag
Still a breadboard prototype with exposed wiring, not a finished enclosure
Before: no ambient lighting behind the monitor.After: the strip lit at knob-controlled brightness.
Turning the knob changes the LED brightness in real time.
ECE Skills Gained
Reading analog sensor input and converting it to a control signal
Using a MOSFET as a switch so a low-power signal can control a higher-power load
Wiring and powering a circuit on a breadboard, including grounding
Filtering noisy sensor data in software
Uploading and debugging firmware on non-standard hardware
Final Thoughts
Narrowing the scope was the right call: I ended up with one finished, working feature instead of an unfinished multi-feature panel
The hands-on troubleshooting made Computer Engineering coursework feel more concrete: reading a sensor, picking the right component, and watching code change something physical
Would like to keep building: add the switches and indicator LEDs from the original pitch once I am more comfortable soldering, and possibly the curtain-control feature
