NF FRA5097, Instruction Manual

Page: 82 / 192

4.3 Integration
FRA5097 4-8
4.3 Integration
The methodology of the FRA5097 measurement uses Fourier integral operation of waveforms to
be measured for each cycle (or one (1) time period) so that noise and harmonics components will
essentially be excluded from the measurement result. The accuracy of measurement can be
augmented by increasing the number of integrations (note: the integral for the period of one cycle
is counted as one time of integration), especially when signal levels are low compared with noise
levels or when high accuracy measurement is required.
The number of integrations (or the integration time period)that is set for the FRA5097
measurement corresponds to averaging operation of input signal waveforms before Fourier
integration. The Fourier integral operation by its nature suppresses harmonics component by
more than 60 dB, irrespective of the number of times of Fourier integrations.
White noise component is suppressed by Fourier integration by the amount approximately
proportional to the square root of the number of Fourier integrations. Noise components outside
of analysis frequencies are also suppressed by increasing the number of Fourier integrations.
Therefore, the larger the number of Fourier integrations, the higher the accuracy of measurement.
The time required for measurement is, needless to say, proportional to the number of times of
Fourier integrations.
The FRA5097 has two (2) ways of integration operation methods available: manual integration
and automatic integration.
a)
Manual integration
Measurement is done at each analysis frequency in the sweep frequency range for the number
of times of Fourier integrations (or the integration time period) that has been preset. The
number of integrations is set by the menu [Measure][BASIC FUNCTION][INTEGRATION
TIME]. The unit to be used for setting this integration parameter is either time duration (e.g.,
ms) or period (cycle). When you set the parameter by time duration, this will be converted
into the number of cycles within the instrument.
The time period required for one time of integration depends upon the analysis frequency f;
approximate time period required for one time of integration is shown in the following for each
analysis frequency range:
・ f ≦ 54 Hz (approx.) : Time period (duration) of one (1) cycle
・ 54 Hz (approx.)
<
f
<
3 kHz : In the range of 18.2 ms to 54.6 ms
・ 3 kHz ≦ f : 18.2 ms (approx.)
“Fig. 4-5 Effect of integration”
illustrates an example of effect of the number of times
of integrations by comparing the numbers 100 versus 1. The figure shows that the noise
suppression effect is approximately 10 (20 dB), which is the square root of
100
.