Magnetic Levitation Control for a Steel Ball

ISEF Category: Robotics and Intelligent Machines

Subcategory: Control Theory  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: Full Year

The Hook

A steel ball can float in midair, but only if your controller reacts faster than gravity pulls it down. That makes magnetic levitation a perfect test of feedback control. Small errors can send the ball crashing or sticking to the magnet. If you can keep it stable, you are doing real systems engineering.

What Is It?

Magnetic levitation, or maglev, means holding an object in the air with a magnetic field. In this project, a coil acts like an adjustable magnet, a Hall-effect sensor measures the ball’s position, and a controller changes the coil current to keep the ball centered. Think of it like balancing a broom on your hand, except your hand is a sensor and the broom is a steel ball.

The hard part is that the system is unstable. If the ball moves a little too far up or down, the magnetic force changes in the wrong direction unless your controller corrects it fast enough. A phase-lead compensator gives the control signal an early push, which helps the system react sooner. You can also model the plant, which means the part of the system you are trying to control, by using step-response data and seeing how the ball responds when you change the input suddenly.

Why This Is a Good Topic

This is a strong science fair topic because you can test real control ideas with measurable results. You can compare stability, settling time, and disturbance rejection across different controller settings, so your project is more than a cool demo. It also connects to real systems used in robotics, precision positioning, and active suspension. You can learn sensing, feedback, modeling, tuning, and data analysis in one build.

Research Questions

• How does phase-lead compensation change the settling time of a magnetic levitation system?
• What is the effect of different Hall-effect sensor placements on position stability?
• Does adding derivative action improve disturbance rejection when the rig is tapped?
• To what extent does coil resistance change affect the maximum stable levitation gap?
• Which controller gain settings produce the smallest steady-state error for a steel ball?
• How does the identified plant model predict the real step response across different ball positions?

Basic Materials

• Hand-wound coil with insulated copper wire.
• Steel ball bearings of known size and mass.
• Hall-effect sensor module.
• Microcontroller board, such as Arduino or similar.
• Logic-level MOSFET or transistor driver.
• Flyback diode.
• DC power supply with current limit.
• Breadboard or prototyping board.
• Jumper wires and alligator clips.
• Multimeter.
• Rigid frame or mount for the coil and sensor.
• Tape measure or ruler.
• Notebook for data logging.

Advanced Materials

• Oscilloscope or USB data acquisition system.
• Precision current sensor.
• Bench power supply with low-noise output.
• 3D-printed or machined coil mount.
• Laser distance sensor or linear position sensor for cross-checking.
• Function generator for disturbance tests.
• Different coil geometries for comparison.
• Multiple Hall-effect sensors for calibration studies.
• Thermal probe for coil heating measurements.
• Current shunt resistor with known tolerance.
• Computer for model fitting and controller design.

Software & Tools

• Excel: Organizes step-response data, plots stability metrics, and compares controller settings.
• Google Sheets: Tracks trials, computes averages, and makes quick graphs during early testing.
• Python: Fits transfer functions, estimates model parameters, and compares control responses.
• ImageJ: Measures ball position from video frames when you use a camera-based backup method.
• MATLAB Onramp: Helps you learn basic control analysis and plotting if your school has access.

Experiment Steps

1. Define the control goal, then choose one measurable output such as levitation height or position error.
2. Map the open-loop behavior, so you can identify where the system becomes unstable and what the step response looks like.
3. Build a calibration curve that converts sensor output into ball position, then check whether that curve stays consistent across trials.
4. Choose the controller structure, then decide which parameters you will tune first, such as proportional, integral, or phase-lead terms.
5. Plan disturbance tests that stress the system in the same way each time, so you can compare rejection performance fairly.
6. Set your analysis metrics, then decide how you will report settling time, overshoot, steady-state error, and repeatability.

Common Pitfalls

• Mounting the Hall-effect sensor too far from the ball, which makes the signal too weak to support stable feedback.
• Ignoring coil heating, which changes resistance and shifts the controller behavior during longer runs.
• Tuning gains only until the ball floats, which can hide poor disturbance rejection and large overshoot.
• Skipping calibration of sensor output to position, which makes your error measurements hard to trust.
• Using inconsistent tap strength during disturbance tests, which makes your comparison between controller settings meaningless.

What Makes This Competitive

A stronger project goes beyond making the ball hover. You need a clear model, a repeatable way to identify the unstable plant, and a fair comparison between controller designs. Strong entries also measure disturbance rejection with the same test method every time, then use real metrics instead of just video evidence. If you compare predicted response against actual response and explain the gaps, your project starts to look like engineering research.

Project Variations

• Use a different levitated object, such as a smaller steel sphere, and compare how mass changes controller tuning.
• Replace the Hall-effect sensor with an optical sensor and test whether position estimation becomes more stable.
• Compare phase-lead control against PID control, then analyze which one handles taps and drift better.

Learn More

• MIT OpenCourseWare Signals and Systems: Search MIT OpenCourseWare for lecture notes and assignments on feedback, transfer functions, and stability.
• NIH PubMed: Search for review articles on magnetic levitation control, sensor feedback, and system identification.
• NASA NTRS: Search the NASA Technical Reports Server for control papers on actuator response, stability, and disturbance rejection.
• IEEE Xplore abstracts: Search for recent magnetic levitation and control theory papers, then read abstracts and accessible figures.
• NIST Engineering Statistics Handbook: Use this for guidance on uncertainty, repeatability, and comparing trial results.