Measurement Uncertainty

General Instructions

This spectral simulation is an interactive Java applet. You can change parameters by clicking on the vertical arrow keys. The five control buttons at the lower right are used to start (triangle) and pause (square) the simulation, to skip forward or back one section at a time (double triangles), and to change speed (+ and -).

After the simulation is complete, the start button takes you back to the beginning of the simulation. You may experience a delay at this point.

What happens when we try to measure a signal in the presence of an unknown interfering signal (noise)? How badly does the measurement degrade? For example, if we have two sinusoids of the same frequency, with the interfering signal 6 dB below the signal of interest (i.e., about 50% of its amplitude), what will the resulting measurement be?...

  • No change?
  • 1.5 times the amplitude of the signal of interest?
  • 0.5 times the amplitude of the signal of interest?

You can see the result in yellow when you hit the Play button.

 


This experiment requires a Java-enabled Web Browser.

 

The answer depends upon the phase of the interfering signal. If the two are directly in phase or directly out of phase, the impact on the amplitude will be the greatest, but the phase error will be small. conversely, if the signals are 90 degrees out of phase, the amplitude measurement is relatively accurate but the phase error is large. The worst case "uncertainty" caused by the interfering signal is displayed numerically on the top of the simulation.

In a measurement system, this is an especially annoying problem. If you are trying to measure an AC signal, the interfering signal can cause the unknown to "rack" up and down. If the "racking" is at a very low rate, you might be fooled into thinking that the measurement is stable, but indeed it is not only unstable, but often incorrect.