AlbertoBarragan

My business card has no battery, but when you tap it on a phone, 21 LEDs light up in an animation and the phone pulls up my website. I built it from an open source hardware design, customized the layout in KiCad, had the boards made, and programmed them myself.

Microcontroller

ATtiny816

NFC chip

NXP NTAG I2C Plus

LEDs

21, Charlieplexed

Antenna

PCB coil, ~2.75 µH

Frequency

13.56 MHz

Battery

None

Power from the phone

When NFC is on, a phone puts out a magnetic field oscillating at 13.56 MHz. The copper spiral around the edge of the card is an antenna, and by Faraday's law that changing field induces a voltage in it. The coil is about 2.75 µH, and together with the roughly 50 pF inside the NFC chip it forms an LC circuit that resonates right at 13.56 MHz, since f = 1/(2π√(LC)). A small 1.5 pF capacitor lets the tuning be adjusted.

The NFC chip

The card uses an NXP NTAG I2C Plus. Most NFC tags only use the harvested energy to run themselves, but this chip rectifies it to DC and outputs the extra power to run other parts. That's what powers everything on the card.

Sending the link

The card doesn't actually transmit anything. The chip changes how much energy it absorbs from the field, and the phone reads those small dips as data. That's how it gets my link.

The light show

An ATtiny816 microcontroller runs at 1 MHz to save power and drives the LEDs with Charlieplexing. Since LEDs only conduct one way, x pins can control x(x − 1) LEDs, so 6 pins can handle up to 30. Only one LED is on at a time, but the code switches between them fast enough that it looks like one smooth animation.

Power management

The chip's datasheet limits the capacitance on its output to 220 nF, but the LEDs need a 10 µF capacitor to smooth out their current. The circuit charges that capacitor slowly through a resistor at startup, then after about 120 ms the microcontroller turns on a MOSFET that bypasses the resistor so the capacitor can do its job.

Board layout

There's no ground plane, because solid copper would block the magnetic field from reaching the antenna. The LED traces also leave a gap instead of forming a full loop, so they don't pick up the field and create eddy currents that would fight the antenna.

What I learned

• How resonance works in a real circuit: the antenna only picks up useful energy at its tuned frequency, so the coil's shape and size are part of the electrical design
• How NFC uses one magnetic field to carry both power and data, through energy harvesting and load modulation
• How Charlieplexing trades lighting LEDs at the same time for using fewer pins, and why fast switching makes it look smooth
• Why datasheet limits matter, like the 220 nF output limit, and how a MOSFET and resistor can work around one safely
• How RF changes the layout rules: no ground plane near the antenna and no closed trace loops
• How to prepare a board for professional assembly, including the BOM, placement files, part numbers and checking that every part matches its footprint

Built with

KiCad, Arduino (megaTinyCore), NXP TagWriter

SARP, the Society for Advanced Rocket Propulsion, is UW's student rocketry team. I joined in fall 2026 and work on the Structures team, where I'm designing the rocket's nose cone in SolidWorks. My main goal is making it lighter while keeping the strength it needs to survive the flight.

Club

SARP, UW

Team

Structures

Part

Nose cone

Software

SolidWorks

Goal

Reduce mass

Joined

Fall 2026

Why weight matters

Every bit of mass in the structure is mass the motor has to carry, so a lighter airframe means a higher apogee and more margin for the payload. The nose cone is a tricky place to cut weight, though. It sits at the very front of the rocket, so its mass helps keep the center of gravity ahead of the center of pressure, which is what keeps the rocket flying straight. Any weight I take out has to be checked against the rocket's stability.

What the nose cone has to handle

• Aerodynamic pressure and heating at the tip, where the airflow hits first
• Axial loads during launch and motor burn
• Shock loads during recovery, when the parachutes deploy

What I'm learning

Rocket architecture, how loads travel through the structure in flight, and how to design a part around strength, weight and manufacturability all at once.

A friend and I started this project to design and build a mid size flying wing drone from scratch, similar in size to the Skywalker X8. Instead of buying a kit, we're making our own frame with a fiberglass skin. We're now expanding the team and registering it as a Registered Student Organization (RSO) at UW.

Layout

Flying wing

Size

Mid size

Skin

Fiberglass

Team

Started by 2, growing

Flight

Autonomous (planned)

Mission

Area surveillance

The problem

When a drone is flown manually to watch over an area, the operator has to pilot it and stare at the camera feed the entire time, scanning everything it sees. That takes a lot of time and attention, and it's easy to miss something important.

Our goal

We want the drone to fly itself over a field or area in a set search pattern and use AI on its camera feed to flag anything unusual, like unexpected movement. Instead of watching the whole flight, soldiers or anyone else using it would only be directed to review the moments the drone flagged, saving their time and attention for what actually matters.

How it would work

1. Choose the area to cover on a map
2. The drone flies a back and forth search pattern on its own
3. Onboard computer vision watches the camera feed for movement or anything out of place
4. Each detection gets flagged with its location and an image
5. The operator reviews only the flagged detections

For the Technology Student Association (TSA) Flight Endurance event, I designed and built a rubber band powered model aircraft with one goal: stay in the air as long as possible. I competed with it two years in a row and placed 4th at the Washington State TSA conference both times.

Event

TSA Flight Endurance

Power

Rubber band

Goal

Maximum time aloft

Result

4th at State, 2024 and 2025

Process

CAD and flight testing

Submitted

Plane and design portfolio

How the event works

Every plane is powered only by a wound up rubber band, so everyone starts with about the same small amount of stored energy. The plane that stays in the air the longest wins. Along with the flights, each competitor turns in a documentation portfolio that explains the design and testing behind their plane.

The engineering challenge

Since the energy in the rubber band is fixed, flight time comes down to how little power the plane needs to stay up. That means keeping it as light as possible and getting as much lift as possible for as little drag. In other words, the whole design is about maximizing the lift to drag ratio while keeping weight down.

What I worked on

• Airfoil selection: choosing a wing shape that makes lift efficiently at the slow speeds these planes fly
• Wing loading: balancing wing area against total weight so the plane can fly slowly without stalling
• Center of gravity: placing it so the plane flies stable and smooth instead of stalling or diving

Design process

I designed the plane in CAD, built it, and then flight tested it. Based on how it flew, I adjusted the design and tested again, repeating that loop to keep improving its time in the air.

What I learned

• How a fixed energy budget drives every design decision
• How small changes in weight and drag make a big difference in performance
• Why flight testing matters as much as the design on paper
• How to document design choices and test results clearly