The characteristic waveform that train wheels emit determines how you analyze and filter the waveform signal point-by-point. A train wheel in motion emits a signal that contains low- and high-frequency components. If you mount a strain gauge in a railroad track, you detect a noisy signal similar to a bell curve. The following illustration shows the low- and high-frequency components of this curve.

The low-frequency component of train wheel movement represents the normal noise of operation. Defective and normal wheels generate the same low-frequency component in the signal. The peak of the curve represents the moment when the wheel moves directly above the strain gauge. The lowest points of the bell curve represent the beginning and end of the wheel, respectively, as the wheel passes over the strain gauge.

The signal for a train wheel also contains a high-frequency component that reflects the quality of the wheel. In operation, a defective train wheel generates more energy than a normal train wheel. In other words, the high-frequency component for a defective wheel has greater amplitude.