Civil Engineering Reference
In-Depth Information
called
chirp
or
sinusoidal sweep
to give some names found in the literature. This is a
signal where the frequency varies continually, linearly or nonlinearly, from a selected
starting frequency to a selected stop frequency.
Figure 1.20
gives an example where the
frequency varies linearly with time. There are several reasons for the relatively recent
popularity of these test signals.
0
0.25
0.5
0.75
0
0.05
0.1
0.15
0.2
0.25
Figure 1.20
Example on a swept sinusoidal signal (“chirp”). The frequency increases linearly from 0-100 Hz in
1 second.
Time (s)
One may shape the spectrum by varying the sweep in different ways. A linear
sweep with constant amplitude will give a white spectrum while increasing the frequency
exponentially with time will give a pink spectrum. However, there are other properties
just as important, especially when comparing with the MLS signals described below.
One advantage that should be mentioned is the reduced sensitivity to time variance
e.g. changes in the propagating medium during measurement, for sound waves caused by
air movements and temperature changes. The reason is that information is collected for
one frequency at the time, whereas for noise signals the information is collected for all
frequencies during the whole measurement.
Figure 1.21
shows an example of spectrum and autocorrelation function for a swept
sine signal where we have chosen a starting frequency of 100 Hz and made it sweep
linearly to 300 Hz. As expected we get a reasonably flat band-like spectrum but as seen
from the diagram at b) one certainly needs to be aware of the side lobes in the spectrum.
The second type of deterministic signals extensively used in modern measurement
technique is the MLS signals (maximum length sequences). The advantages of using this
type of noise-like signals will be clear when treating transfer function measurements in
the following chapter. Here we shall give an overview of the most important properties
only.
