What is Data Aliasing?
Aliasing, in the recording of dynamometer data, is the inducement of anomalies in the captured signals as a result of using too long a sampling interval. In video capture, using too slow a frame rate, you see aliasing as the slowing or reversing of a spinning wheel’s spokes.
During engine testing, if the selected sampling interval is too long, high-frequency events may induce low-frequency humps and bumps into the recorded data. Aliasing effects can vary with things like the test RPM!
Consider the torque channel, it is subjected to a high/low oscillation every time a cylinder fires. So, it’s data inherently contains a ripple – the frequency of which varies with RPM. That is why you see the magnitude of aliasing errors change as the RPM frequency comes in and out of phase with a data-acquisition system using some fixed sampling interval.
Accurate capture theoretically requires a sampling interval less than one-half the inverse of the highest Fourier frequency component within the sampled channel. This is referred to as the Nyquist interval.
The effects of aliasing in the recorded data may be avoided by filtering (e.g. electronically dampening with capacitors) the source signal to remove any frequency components that are higher than those required in the captured data.