15A.10 Swept-Sine Measurements

Back

This Appendix discusses using the swept-sine VIs located on the Swept Sine palette (see Chapter 17). The

swept-sine measurements include dynamic measurements for stimulus level, response level, frequency

response (gain and phase), THD, and individual harmonic distortion.

15A.10.1 Swept-Sine Overview

Swept sine is a technique for characterizing the frequency response of the DUT. Two techniques are

commonly used in swept-sine measurements. The first technique slowly sweeps through a range of

FIGURE 15A.54 STFT VI block diagram.

FIGURE 15A.55 STFT waterfall display.

FIGURE 15A.56 Block diagram for VI displaying octave spectra.

Virtual Instrumentation for Data Acquisition, Analysis, and Presentation 15-105

© 2005 by Taylor & Francis Group, LLC

frequencies in a manner similar to a chirp. Figure

15A.58 shows an example of the excitation signal

for this form of swept-sine measurement.

The second technique steps through a range of

frequencies. Figure 15A.59 shows an example of

the excitation signal for this form of swept-sine

measurement. The swept sine implemented in the

Sound and Vibration Toolkit generates an excitation

signal that steps through a range of test

frequencies, similar to the signal in Figure 15A.59.

Both techniques can yield similar results. However,

they require very different measurement analysis

processes.

Swept-sine frequency-response measurements

compare a response signal to the stimulus tone in order to compute the FRF of the DUT. The

magnitude of the FRF is equivalent to gain and represents the ratio of the output level to the input level

for each test frequency. The phase of the FRF is equivalent to the phase lag introduced by the DUT for

each test frequency.

Swept-sine measurements require a signal source. The stimulus signal is always a single tone that

excites the DUT at the test frequency. Since the stimulus is a single tone, swept-sine analysis can measure

the harmonic distortion while simultaneously measuring the linear response.

15A.10.2 Choosing Swept-Sine vs. FFT Measurements

The frequency response of the DUT is a useful tool. The Sound and Vibration Toolkit provides two

distinct techniques to measure the frequency response. The swept-sine technique performs single – tone

FIGURE 15A.58 Sweeping swept sine example.

FIGURE 15A.59 Stepping swept sine example.

FIGURE 15A.57 Waterfall display for octave spectra

analysis.

15-106 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

measurements at each test frequency. The FFTbased

technique measures the response over the

entire acquisition bandwidth. Table 15A.8 lists the

basic differences between swept-sine and FFTbased

techniques for measuring frequency

response.

Swept-sine measurements offer superior

dynamic range over FFT-based measurements

because you can optimize the signal level and

input ranges at each test frequency. FFT-based

techniques must specify a signal level and input

ranges appropriate for the maximum broadband

response.

Figure 15A.60 shows the simulated frequency response function for a four-DoF system. The peak at

17.6 Hz has a magnitude roughly 1000 times larger than the peak at 5.8 Hz. To use an FFT-based

technique, use broadband excitation to excite the entire frequency range of interest, to measure the

frequency response. This situation forces one to set the input range so that the overall response does not

overload the DUT or the acquisition device. Therefore, when you measure the response at 5.8 Hz, you

lose 60 dB of measurement dynamic range. The swept-sine technique allows you to tailor the excitation

amplitude to the specific test frequency, preserving the full measurement dynamic range.

FFT-based measurements are limited to a linearly spaced frequency resolution determined by the

sample rate and the block size. Refer to Appendix 15A.7 for more information on FFT-based

measurements. When the response changes rapidly, this frequency resolution may not yield enough

information about the dynamic response. Also, a linear resolution may yield an excessive amount of

information in frequency regions where the dynamic response is relatively constant. Swept-sine analysis

has the ability to test arbitrary frequency resolutions that are linear, logarithmic, or adapted to the

dynamic response of the DUT. When the frequency resolution is adapted to the DUT dynamic response,

you can test more frequencies in regions where the dynamic response is of interest to your application

and fewer where it is not.

The main benefit of swept-sine analysis is the ability to measure harmonic distortion simultaneously

with linear response. FFT-based analysis offers a speed advantage for broadband measurements with

many test frequencies.

15A.10.3 Taking a Swept-Sine Measurement

Use the SVT Initialize Swept Sine VI to create a new swept-sine task for the designated device, source

channel settings, and acquisition channel settings. Swept sine in the Sound and Vibration Toolkit only

supports measurements on a single device with output and input capabilities.

Use configure swept sine VIs in the configure swept sine palette to configure the scaling, test

frequencies, averaging, delays, and other measurement settings. These configuration VIs allow control

over basic and advanced measurement parameters. The order in which you place the configuration VIs is

important, as it allows you to customize a swept-sine measurement. For example, you can easily generate

TABLE 15A.8 Swept Sine and FFT Differences

Swept-Sine Frequency Response FFT-Based Frequency Response

Single-tone excitation Broadband excitation

Can measure harmonics Cannot measure harmonics

Arbitrary test frequencies Linearly spaced frequency resolution

Longer test time for many test frequencies

Better dynamic range possibility

FIGURE 15A.60 Swept-sine and FFT measurements.

Virtual Instrumentation for Data Acquisition, Analysis, and Presentation 15-107

© 2005 by Taylor & Francis Group, LLC

100 logarithmically spaced test frequencies in the audio range, then apply inverse A-weighted scaling to

the excitation level by adding code similar to that in Figure 15A.61 into your swept-sine application.

You can use the swept-sine configuration VIs to customize your swept-sine application. For example,

to speed up a swept-sine measurement, reduce the settling or integration time specified by the SVT Set

Swept Sine Averaging VI. You also can configure the device IEPE with the SVT Set Swept Sine Coupling

and IEPE Excitation (DAQmx) VI. You also can reduce the block duration input to SVT Set Swept Sine

Block Duration VI.

Note: The minimum block duration is limited by the capabilities of the computer processing the

measurement. A very small block duration can result in a loss of continuous processing, causing the

swept-sine measurement to stop and return an error.

Use the SVT Start Swept Sine VI to begin the generation and acquisition. The VI fills the device output

buffer with zeros before writing the first test frequency excitation.

The SVT Swept Sine Engine VI continually acquires data and processes it to remove samples acquired

during delays, transitions, and settling periods. The SVT Swept Sine Engine VI performs measurement

analysis on samples acquired during integration periods. The SVT Swept Sine Engine VI updates the

excitation to excite the DUT at the next test frequency after it integrates sufficient data at the current test

frequency.

Note: The transition to the next excitation tone, both frequency and amplitude, always occurs at a zero

crossing to minimize transients introduced to the DUT.

Use the Read Swept Sine Measurements VIs in the Read Swept Sine Measurements palette to read the

raw measurements, scale the measurements, and perform additional conversions to display and report

the data in the desired format.

Use the SVT Close Swept Sine VI to stop the generation and acquisition, and clear the sweptsine

task.

15A.10.4 Swept-Sine Measurement Example

This example of a swept-sine application measures

the frequency response and harmonic distortion of

a notch filter. In this example, a NI PXI-4461

generates the excitation signal and acquires the

stimulus and response signals.

Figure 15A.62 illustrates the connection scheme

used in this example to measure the dynamic

response of the DUT using a swept-sine measurement.

The acquired stimulus signal on the analog

input channel 0, the AI0, is the generated excitation

signal from the analog output channel 0, AO0.

The NI PXI-4461 converts the desired stimulus

signal from digital data to an analog signal and

outputs that signal on AO0. The excitation signal is

FIGURE 15A.62 Swept-sine measurement connection

diagram.

FIGURE 15A.61 Customizing a swept-sine measurement.

15-108 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

connected to both the stimulus input channel AI0 and the input terminal of the DUT. The response

signal is connected from the output terminal of the DUT to the response input channel AI1.

The DUT for this example is a notch filter centered at 1 kHz.

Figure 15A.63 shows the block diagram of the example SVXMPL_swept sine FRF DAQmx VI, which

ships with the Sound and Vibration Toolkit.

FIGURE 15A.63 Block diagram of SVXMPL_swept sine FRF DAQmx VI.

TABLE 15A.9 Swept Sine Measurement Steps

Step Number Description Optional or Required

1 Initialize a swept-sine measurement by specifying the hardware device and

channel settings

Required

2 Specify the scaling that will be applied to the acquired stimulus and

response data.

Optional

3 Configure the source by specifying the test frequencies, amplitude, and

whether or not the sweep automatically restarts after completion

Required

4 Set the settling and integration parameters to allow sufficient time for the

DUT to settle before the measurement is performed at the new test

frequency and that there is sufficient integration time to achieve the

desired level of accuracy

Required

5 Set the block duration input terminal for the measurement to be small

enough to give a reasonable test time and large enough so that it does not

put the test computer at risk of being unable to continuously generate

and read the signals. The smaller the block size, the faster the swept sine

can transition from one test frequency to the next

Optional

6 Explicitly set the sample rate for the measurement. The rate is automatically

selected if this VI is not used. The same rate is used for input and output

channels

Optional

7 Specify the propagation time terminal input specific to the DAQ device

being used for the measurement. You can measure the device

propagation time using the SVL Measure Propagation Delay VI.

Refer to Appendix 3, Scaling and Calibration, for more information

Optional

8 Configure the harmonic distortion measurement by specifying the

maximum harmonic to use in the computation of the THD.

Only those harmonics specified in the harmonics to visualize

array return individual harmonic components

Required if

performing distortion

measurements

9 Start the swept sine to perform the hardware configuration and start the

output and input tasks. Channel synchronization is performed

internally in this VI

Required

10 Generate the excitation and acquire the stimulus and response data at

each test frequency

Required

11 Convert the raw data to the specified format in order to display and

report measurement results

Required

12 Stop the swept-sine measurement and clear the output and input tasks

to release the device

Required

Virtual Instrumentation for Data Acquisition, Analysis, and Presentation 15-109

© 2005 by Taylor & Francis Group, LLC

FIGURE 15A.65 Magnitude and phase response of a 1 kHz notch filter.

FIGURE 15A.64 Time-domain results.

FIGURE 15A.66 THD vs. frequency.

15-110 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

Table 15A.9 documents the actions performed by the VIs in Figure 15A.63. Some steps are required

and must be done for the VI to function correctly. The optional steps allow you to customize your

measurement.

The while loop in Figure 15A.63 controls the synchronized generation and acquisition. Display

controls and measurement indicators are updated inside the while loop. This loop allows for the

monitoring of intermediate results.

Many of the steps in Table 15A.9 are configuration steps. Through the Sound and Vibration Toolkit

swept-sine configuration VIs, you can specify numerous configuration parameters to achieve fine control

of the swept-sine measurement parameters. For many applications two or three configuration VIs are

sufficient.

It is important to allow for the propagation delay of the DAQ or DSA device. This delay is specific to

the device used to perform the measurement. To determine the device propagation delay, refer to the

device documentation or measure the delay with the SVL Measure Propagation Delay VI.

Figures 15A.64 to 15A.67 display measurement results obtained with the SVXMPL_swept sine FRF

DAQmx VI example program. Figure 15A.64 shows the time-domain stimulus and response signals for

the 138.49 Hz test frequency. From the time-domain data, you can see that the notch filter has attenuated

the signal and introduced a phase shift.

Figure 15A.65 shows the magnitude and phase responses of the notch filter at all the test frequencies in

the magnitude and phase spectra in the Bode plot.

In addition to measuring the frequency response, this example simultaneously measures

the harmonic distortion at each test frequency. Figure 15A.66 shows the graph of THD vs.

frequency.

You expect to see a peak in the THD at the notch frequency. The peak occurs because the

fundamental frequency is attenuated at the notch frequency. However, the graph indicates that this

measurement has failed to accurately identify the power in the harmonic distortion components. For the

example in Figure 15A.66, the number of integration cycles is two. More integration cycles must be

specified to perform accurate harmonic distortion measurements. If you change the number of

integration cycles to ten and rerun the example, you obtain the THD vs. frequency results displayed

in Figure 15A.67.

Now, with a sufficient number of integration cycles specified, you can see the characteristic peak in the

THD at the center frequency of the notch filter.

FIGURE 15A.67 THD vs. frequency results.

Virtual Instrumentation for Data Acquisition, Analysis, and Presentation 15-111

© 2005 by Taylor & Francis Group, LLC

Bibliography

Crocker and Malcolm, J. 1998. Handbook of Acoustics, Wiley, New York.

Design Response of Weighting Networks for Acoustical Measurements, ANSI S1.42-2001, American National

Standards Institute, Washington, 1986.

Electroacoustics — Sound Level Meters, International Standard IEC 61672, 1st ed., 2002 – 2005,

International Electrotechnical Commission, Geneva, Switzerland, 2002.

Hassall, J.R. and Zaveri, K. 1988. Acoustic Noise Measurements, Bruel & Kjær, Nærum, Denmark.

Measurement of Audio-Frequency Noise Voltage Level in Sound Broadcasting, ITU-R Recommendation

468-4, 1986.

Octave-Band and Fractional Octave-Band Filters, International Standard IEC 1260, 1st ed., 1995-07.

International Electrotechnical Commission, Geneva, Switzerland, 1995.

Preferred Frequencies for Measurements, International Standard IEC 266, 1st ed., 1975-07-15,

International Electrotechnical Commission, Geneva, Switzerland, 1975.

Psophometer for Use on Telephone-Type Circuits, ITU-T Recommendation O.41, Revised, 1993 – 1996.

Telecommunication Standardization Sector of the International Telecommunication Union, 1995.

Randall, R.B. 1987. Frequency Analysis, BrY¨ el & Kjær, Nærum, Denmark.

Smallwood, D., An improved recursive formula for calculating shock response spectra, Shock Vib. Bull.,

51, Pt 2, 211 – 217, 1981, May 1981.

Specification for Octave-Band and Fractional Octave-Band Analog and Digital Filters, ANSI S1.11-1986,

Acoustical Society of America, New York, 1986b.

Specification for Sound Level Meters, ANSI S1.4-1983, American National Standards Institute,

Washington, 1983.

15-112 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

Навигация