Observations on Arduino Resonance Capture

Experiment Setup

The first physical experiment uses an Arduino Mega 2560 connected to a set of piezoelectric sensors mounted along a steel beam of length 0.5m. The beam is clamped at one end and free at the other (cantilever configuration).

Excitation is provided by a small electromagnetic shaker driven by a swept-sine signal from 20 Hz to 2 kHz. The Arduino samples the piezo outputs at 10 kHz using its built-in ADC (10-bit resolution).

Initial Observations

The raw time-domain signal shows clear resonance peaks when the excitation frequency matches the beam's natural modes. The first three bending modes are clearly visible at approximately 47 Hz, 294 Hz, and 823 Hz - consistent with Euler-Bernoulli beam theory predictions for this geometry.

The 10-bit ADC resolution is a limiting factor. Signal-to-noise ratio degrades significantly above 1.5 kHz. An external 16-bit ADC module is being ordered for the next iteration.

Spectral Analysis

A 4096-point FFT applied to windowed segments of the captured signal clearly resolves the first three modes. The frequency resolution at the current sampling rate is approximately 2.44 Hz, which is sufficient for the lower modes but may be inadequate for closely spaced higher modes.

Hanning windowing reduces spectral leakage compared to the rectangular window, and the peak frequencies agree with the theoretical predictions to within 3%.

Issues and Next Steps

• ADC resolution limits dynamic range; upgrade to 16-bit external ADC.
• Electromagnetic interference from the shaker coupling into sensor wires; shielded cables needed.
• Sampling rate constrains upper frequency bound; consider dedicated DAQ board for higher modes.
• Temperature sensitivity of piezo sensors needs characterization.