Smart LED Hourglass

02 · Project Information

• Project Duration: November 2024 - December 2024

• Role: Team Leader

• Advisor: Prof. Joseph Skovira, Cornell University

• Project Type: Embedded Systems Design Project

03 · Project Overview

This project involved the design and development of an interactive Smart LED Hourglass.

The system used a 64×64 RGB LED matrix to display animated sand particles, a Raspberry Pi as the central processing platform, and a LIS3DH accelerometer to measure device orientation and motion.

When the user rotated or flipped the device, the system recalculated the direction of gravity from the accelerometer data. The virtual sand particles then moved toward the new lower area of the display, reproducing the interactive behavior of a physical hourglass.

In addition to the traditional hourglass simulation, the system included a Cornell logo animation and multiple countdown modes.

04 · Project Objectives

The primary objectives were to:

1. Simulate physical sand movement on an RGB LED matrix

2. Detect device orientation through an accelerometer

3. Adjust sand movement according to the direction of gravity

4. Produce smooth and continuous particle animations

5. Support button-based visual mode switching

6. Reset timers through device movement

7. Minimize the delay between user input and visual response

8. Combine embedded control with interactive visual design

05 · Hardware Components

The system consisted of:

• Raspberry Pi

• 64×64 RGB LED matrix

• LIS3DH three-axis accelerometer

• Mode-selection button

• GPIO interfaces

• I²C communication

• LED matrix driver and power components

• Structural and mounting components

The Raspberry Pi read the sensor data, processed user inputs, calculated sand-particle positions, and refreshed the LED matrix.

The LIS3DH accelerometer measured acceleration along three axes, enabling orientation detection and motion-triggered interaction.

06 · System Architecture

The system contained the following modules:

• Hardware initialization

• Accelerometer data acquisition

• Device-orientation detection

• Virtual sand simulation

• LED matrix rendering

• Button interaction

• Countdown control

• Visual mode management

• Motion-triggered reset

These modules were integrated into a real-time loop so that the visual direction of the sand changed immediately when the user flipped the device.

07 · System Workflow

The system followed the process below:

1. Start the Raspberry Pi

2. Initialize the RGB LED matrix

3. Initialize the LIS3DH accelerometer

4. Configure the button and communication interfaces

5. Read three-axis acceleration data

6. Determine the current device orientation

7. Calculate the virtual gravity direction

8. Update the sand positions and timer state

9. Render the new frame on the LED matrix

10. Check for visual mode changes

11. Check whether the timer should be reset

12. Repeat the loop and continuously refresh the animation

08 · Core Functions

1. Traditional Hourglass Simulation

The traditional mode represented sand particles using LED pixels and simulated their movement according to accelerometer data.

When the user flipped the device:

• The accelerometer detected the orientation change

• The system updated the virtual gravity direction

• Sand particles began moving toward the new lower region

• The LED matrix continuously refreshed particle positions

• The particles gradually moved from the upper region to the lower region

This mode digitally reproduced the interaction of a traditional hourglass.

2. Real-Time Motion Sensing

The LIS3DH accelerometer continuously measured motion along three axes.

The system used the data to determine:

• Whether the device had been flipped

• The current gravity direction

• Whether significant movement had occurred

• Whether the timer should restart

• The direction in which the particles should move

Motion sensing allowed the animation to respond dynamically to the physical orientation of the device.

3. Sand-Particle Animation

The sand simulation was the primary visual component of the project.

The software continuously processed:

• Initial particle positions

• Current gravity direction

• The next position of each particle

• Particle collisions

• LED matrix boundaries

• Occupied pixel locations

• Particle colors

• Animation refresh rate

Updating the pixel states over time created a continuous and smooth sand-flow effect.

4. Cornell Logo Sand Mode

The project also included a Cornell logo sand animation.

This mode combined university visual identity with the particle simulation, allowing the Cornell logo to be dynamically represented through LED pixels and moving sand.

It strengthened the project's demonstration value and explored the combination of embedded systems with digital visual design.

5. Countdown Timers

The system provided three countdown modes:

• 30 seconds

• 1 minute

• 5 minutes

The user selected a timer using the control button. During the countdown, the LED matrix displayed the remaining time through particle flow or corresponding visual progress.

A completion state was displayed when the timer reached zero.

6. Button-Based Mode Switching

A physical button allowed the user to switch between the demonstration and timer modes.

The available modes included:

• Traditional hourglass

• Cornell logo animation

• 30-second countdown

• 1-minute countdown

• 5-minute countdown

The software detected the button input and updated the active operating mode and display content.

7. Motion-Triggered Reset

The system allowed the countdown to be reset through device movement.

When the accelerometer detected a qualifying flip or motion event, the software:

1. Checked whether the movement exceeded the trigger threshold

2. Confirmed the new device orientation

3. Cleared the previous particle state

4. Reset the countdown

5. Regenerated the sand particles

6. Restarted the animation in the new direction

This interaction allowed users to restart the timer by flipping the device like a physical hourglass.

09 · Response Performance

The original interaction-response target was 100 ms.

After optimizing sensor acquisition, state evaluation, and display refresh, the final system achieved a 50 ms response time, exceeding the original target.

The lower latency reduced the delay between device movement, button input, and visual updates.

10 · Software Design

The software was organized around real-time input, state evaluation, physical simulation, and visual rendering.

The primary modules included:

• LED matrix initialization

• Accelerometer initialization

• I²C data acquisition

• Button-state detection

• Device-orientation calculation

• Particle-position updates

• Collision and boundary evaluation

• Countdown logic

• Mode switching

• Frame buffering and display refresh

The visual modes and low-level hardware control were separated into relatively independent modules to improve maintainability.

11 · Real-Time Loop

The system processed sensor input and animation updates within a continuous real-time loop.

Each cycle included:

1. Reading accelerometer data

2. Calculating the gravity direction

3. Checking the button state

4. Determining whether the mode had changed

5. Updating the countdown

6. Calculating the next particle positions

7. Processing boundaries and occupied pixels

8. Updating the frame buffer

9. Refreshing the LED matrix

The software controlled the computational load of each cycle to prevent sensor or button processing from interrupting the animation.

12 · Development Process

The project was developed through modular implementation and incremental integration.

The development process included:

• Defining the project concept and interactions

• Connecting the Raspberry Pi and LED matrix

• Testing individual pixels and basic graphics

• Configuring the LIS3DH accelerometer

• Reading and calibrating three-axis acceleration data

• Implementing device-orientation detection

• Developing the initial sand animation

• Adding boundary and collision logic

• Developing countdown functionality

• Adding button-based mode switching

• Implementing motion-triggered reset

• Developing the Cornell logo animation

• Optimizing system response time

• Completing the final demonstration and testing

13 · Challenges and Solutions

14 · My Responsibilities

As the team leader, I was responsible for:

• Defining the project concept and feature scope

• Designing the overall hardware and software architecture

• Dividing tasks and coordinating project progress

• Supporting Raspberry Pi and LED matrix integration

• Participating in LIS3DH data acquisition

• Designing the sand simulation and gravity-direction logic

• Supporting timer and mode-switching development

• Organizing system integration and performance testing

• Optimizing real-time response

• Participating in the Cornell logo visual-mode design

• Preparing the project results and final demonstration

15 · Key Contributions and Outcomes

• Developed a digital hourglass using a 64×64 RGB LED matrix

• Used a Raspberry Pi for real-time animation control

• Integrated a LIS3DH accelerometer for orientation sensing

• Implemented motion-based sand-flow simulation

• Implemented button-controlled mode switching

• Implemented motion-triggered timer resets

• Developed a traditional hourglass visual mode

• Developed a Cornell logo sand animation

• Implemented 30-second, 1-minute, and 5-minute timers

• Achieved a 50 ms interaction response time

• Exceeded the original 100 ms response target

• Produced smooth animations across multiple visual modes

16 · Conclusion

This project combined embedded systems, motion sensing, pixel animation, and interaction design to create a Smart LED Hourglass that dynamically adjusted particle movement according to device orientation.

Through the project, I strengthened my skills in Raspberry Pi development, RGB LED matrix control, I²C sensor communication, real-time input processing, and animation-state management.

The project also improved my experience in system architecture, hardware-software integration, team coordination, performance optimization, and interactive product design.