JP2015114294A - Inspection apparatus of acoustic device and inspection method of acoustic device and inspection program of acoustic device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reliably perform a test of an acoustic device even when the type of a product and the environment of a production line change.SOLUTION: The buzzer sound outputted from a buzzer 5 is collected by a microphone 6, and converted into a frequency spectrum by a first Fourier converter 12, and normalized by a first normalization part so that the maximum peak value becomes 1. At the same time, the vibration of a workpiece 2 is obtained by a vibration sensor 7, and converted into the frequency spectrum by a second Fourier converter 22, and normalized by a second normalization part so that the maximum peak value becomes 1. The normalized frequency spectrum of two types are compared by a minimum value selection part 25, and the frequency spectrum which selects the minimum value is prepared, and the presence or absence of abnormal noise generated due to the buzzer 5 from the peak of the frequency spectrum is determined.

Description

  The present invention relates to an inspection apparatus for an acoustic device, an inspection method for an acoustic apparatus, and an inspection program for an acoustic apparatus.

  In a manufacturing process in which a buzzer is attached to an acoustic device that outputs sound to a work such as a substrate, the operation of the buzzer is checked after the buzzer is attached to the work. At this time, if the caulking of the buzzer housing is weak or the buzzer and the work are not fixed properly, resonance will occur between the housing and the work when the buzzer sound is output, or the friction between the housing and the work will be reduced. May cause abnormal noise.

  Here, as a method for examining the occurrence of abnormal noise, it is possible that an operator directly determines the presence or absence of abnormal noise in the production line testing process. However, this method may lack objectivity because the presence or absence of abnormal noise is determined by the operator's sense. Therefore, conventionally, the buzzer sound can be accurately measured by simultaneously measuring the sound pressure and vibration when the buzzer is sounded. For example, in a product that generates vibration at the same time as the output of sound, approximating the sound radiation coefficient of other frequencies using the sound radiation coefficient calculated at a frequency with a high vibration level, the sound generated by the buzzer using the sound radiation coefficient It becomes possible to acquire with high accuracy.

  However, in a production line that produces a large amount of products, when the products are conveyed by the conveyor device, the vibration of the conveyor device may be transmitted to the product and the product may vibrate. Further, factory noise, for example, air blow operation sound used when removing dust adhering to a product may be collected as sound pressure data. These sounds and vibrations are factors that reduce the measurement accuracy of the sound generated by the buzzer. Therefore, when the sound pressure and vibration from the product are measured and the magnitude of vibration exceeds the threshold value, or when it is clear that vibration has occurred due to driving a part of the product, The sound generated at that time is removed from the time-series sound pressure data and analyzed.

JP-A-1-69921 JP 2003-279401 A JP 2004-317248 A

However, in order to perform data analysis by removing sound pressure data under a certain condition, it is necessary to set a vibration threshold value for each type of product in advance. Further, when there are individual differences in products, when the environment of the production line changes, or when there are a plurality of production lines, a deviation occurs in the vibration for each production line. For this reason, conventionally, it is necessary to set the threshold value by measuring the sound pressure and vibration for each production line, and the work efficiency is poor.
The present invention has been made in view of such circumstances, and it is an object of the present invention to ensure that a test of an acoustic device can be performed even if the type of product or the environment of a production line changes.

  According to one aspect of the embodiment, the first normalization unit that normalizes the frequency spectrum of the sound output from the inspection object using the maximum value of the amplitude, and the vibration generated around the inspection object A second normalization unit that normalizes the frequency spectrum of the two using the maximum amplitude value, a selection unit that compares the normalized amplitudes of the two frequency spectra for each frequency, and selects a smaller amplitude value; A spectrum creation unit that creates a test spectrum consisting of a frequency spectrum of sound pressure using the data selected by the selection unit, and the presence or absence of abnormal noise generated from the test object from the peak height of the test spectrum An inspection device for an acoustic device, characterized by comprising:

  Further, according to another aspect of the embodiment, the sound pressure data of the sound output from the inspection object is acquired, and vibration data of vibration generated around the inspection object is acquired simultaneously with the sound pressure data. The frequency spectrum of the sound pressure data is normalized using the maximum amplitude value, the frequency spectrum of the vibration data is normalized using the maximum amplitude value, and the amplitudes of the two normalized frequency spectra are As compared with the above, a method for inspecting an acoustic device is obtained, in which a value having a smaller amplitude is acquired and a frequency spectrum of sound pressure for inspecting abnormal noise generated from the inspection object is created.

  Furthermore, according to another aspect of the embodiment, the sound pressure data of the sound output from the inspection object is acquired, and vibration data of vibration generated around the inspection object is acquired simultaneously with the sound pressure data. The frequency spectrum of the sound pressure data is normalized using the maximum amplitude value, the frequency spectrum of the vibration data is normalized using the maximum amplitude value, and the amplitudes of the two normalized frequency spectra are Compared to the above, an acoustic apparatus inspection program is realized, wherein the computer executes a process of acquiring a value having a smaller amplitude and creating a frequency spectrum of sound pressure for inspecting the abnormal sound generated from the inspection object. Is done.

  It is possible to directly remove the noise of disturbance by directly comparing the sound pressure data and the vibration data, and it is possible to inspect the acoustic apparatus from which the influence of the external noise is removed without using the threshold value.

FIG. 1 is a block diagram showing an example of a schematic configuration of an inspection apparatus for an acoustic device according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating an example of an inspection method for an acoustic device according to an embodiment of the present invention. FIG. 3 is a diagram illustrating an example of a normalized frequency spectrum of sound pressure created by the acoustic device inspection apparatus according to the embodiment of the present invention. FIG. 4 is a diagram illustrating an example of a normalized frequency spectrum of vibration created by the inspection apparatus for an acoustic device according to the embodiment of the present invention. FIG. 5 is an enlarged view of a part of the normalized frequency spectrum of sound pressure and vibration created by the inspection apparatus for an acoustic device according to the embodiment of the present invention. FIG. 6 is a diagram for explaining an example of processing for comparing the amplitudes of normalized frequency spectra of sound pressure and vibration for each frequency in the inspection apparatus for an acoustic device according to the embodiment of the present invention. FIG. 7 is a partially enlarged view showing an example of a comparison result of a normalized frequency spectrum of sound pressure and vibration created by the inspection apparatus for an acoustic device according to the embodiment of the present invention. FIG. 8 is a diagram illustrating an example of a frequency spectrum of sound pressure for inspection of an acoustic device created by the inspection device for an acoustic device according to an embodiment of the present invention.

The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.
The foregoing general description and the following detailed description are exemplary and explanatory only and are not intended to limit the invention.

FIG. 1 shows a schematic configuration of the production line.
The production line 1 includes a conveyor device 3 that conveys the workpiece 2, and an acoustic device inspection device 4 (hereinafter simply referred to as an inspection device). The inspection device 4 includes a microphone 6 that is a sound collecting device that collects sound generated by an acoustic device such as a buzzer 5 attached to the workpiece 2, and a vibration sensor 7 that detects vibration generated in the workpiece 2. The vibration sensor 7 is preferably of a type that can measure vibration without contact with the workpiece 2.

  The inspection device 4 includes a data processing device 8 that analyzes data acquired by the microphone 6 and the vibration sensor 7. The data processing device 8 has a first input circuit 11 to which sound pressure data is input. The first input circuit 11 includes, for example, an A / D (Analog / digital) converter that digitally converts analog sound pressure data output from the microphone 6. The output of the first input circuit 11 is connected to the first Fourier transformer 12. The first Fourier transformer 12 creates a frequency spectrum by performing fast Fourier transform (FFT) on time-series sound pressure data. The output of the first Fourier transformer 12 is connected to the first normalization unit 13. The first normalization unit 13 normalizes the frequency spectrum of the sound pressure data so that the maximum value of the amplitude of the frequency spectrum of the sound pressure data becomes “1”.

  In addition, the data processing device 8 includes a second input circuit 21 to which vibration data is input. The second input circuit 21 includes, for example, an A / D converter that digitally converts analog vibration data output from the vibration sensor 7. The output of the second input circuit 21 is connected to the second Fourier transformer 22. The second Fourier transformer 22 performs a fast Fourier transform on the time-series vibration data to create a frequency spectrum. The output of the second Fourier transformer 22 is connected to the second normalization unit 23. The second normalization unit 23 normalizes the frequency spectrum of the vibration data so that the maximum value of the amplitude of the frequency spectrum of the vibration data becomes “1”.

  The outputs of the two normalization units 13 and 23 are connected to the minimum value selection unit 25. The minimum value selection unit 25 selects the minimum value of the amplitudes of the two normalized frequency spectra for each predetermined frequency bin, and constructs a frequency spectrum (reconstructed spectrum) composed of the minimum values of amplitude. The output of the minimum value selection unit 25 is connected to a multiplier 26 which is a spectrum creation unit. The multiplier 26 receives the reconstructed spectrum and the normalized frequency spectrum of the sound pressure data, and creates a sound pressure spectrum. The multiplier 26 is connected to the determination unit 27. The determination unit 27 determines whether the buzzer 5 is operating normally on the basis of the peak position and peak height of the sound pressure spectrum, and determines whether the work 2 is good or bad.

  Note that the data processing device 8 of the inspection device 4 may be configured to realize the units 12, 13, 22, 23, and 25 to 27 by causing a computer to execute an inspection program for an acoustic device. In this case, the data processing device 8 develops the inspection program for the acoustic device stored in a storage device (not shown) on a CPU (Central Processing Unit), thereby performing the processing of each unit 12, 13, 22, 23, 25-27. Let it run. Further, the inspection program for the acoustic device may be stored in a storage device (not shown) arranged outside the data processing device 8 and read into the data processing device 8 as necessary.

  Examples of the work 2 include various products such as information terminals, portable devices, and home appliances. The sound device is not limited to the buzzer 5 as long as it can output a predetermined sound by signal input. The buzzer 5 has, for example, a housing in which a plurality of parts are combined, and is attached to the work 2 by fixing the housing to the work 2. Moreover, you may manufacture the housing | casing of the buzzer 5 from one part. Attachment of the buzzer 5 to the work 2 is performed in a previous process (not shown). Further, near the inspection device 1, a factory noise source 31 such as an air blow device or operation sound of other devices is arranged.

Next, an abnormal sound inspection method for the workpiece 2 using the inspection apparatus 4 will be described.
First, the buzzer 5 is attached to the work 2 in the previous process, and then is carried into the conveyor device 3. When the workpiece 2 on the conveyor device 3 reaches the inspection position, the buzzer 5 is sounded based on a command from the data processing device 8 or another control device (not shown). At this time, vibration is generated in the work 2 and the buzzer 5 together with the buzzer sound. In the following, it is assumed that the buzzer sound has a frequency of 1 kHz, for example. Further, it is assumed that factory noise such as air blow noise is generated at 3 kHz. Moreover, the vibration which has a bad influence on a test to the conveyor apparatus 3 etc. has not generate | occur | produced.

  As shown in step S <b> 101 </ b> A of FIG. 2, the inspection device 4 collects sound pressure data of a buzzer sound using the microphone 6. At this time, in addition to the buzzer sound output from the buzzer 5, other sounds generated from the buzzer 5 and the work 2, and abnormal sounds such as the factory noise source 31 are also taken into the microphone 6. In step S102A, the sound pressure data is digitally converted by the first input circuit 11 and fast Fourier transformed by the first Fourier transformer 12 to be transformed into a frequency spectrum. Furthermore, in step S103A, the frequency spectrum is normalized by the first normalization unit 13 so that the maximum value of the amplitude becomes “1”. Thereby, for example, a normalized spectrum FS1 of sound pressure data as shown in FIG. 3 is obtained. In FIG. 3, the horizontal axis represents frequency and the vertical axis represents amplitude, and shows a change in frequency of sound pressure collected by the microphone 6. The normalized spectrum FS1 of the sound pressure amplitude has a peak at 1 kHz of the buzzer sound. The amplitude of the normalized spectrum FS1 is normalized so that the maximum peak having a frequency of 1 kHz has a maximum value “1”. Furthermore, peaks are also generated at 2 kHz and 3 kHz corresponding to harmonics of the buzzer sound. The amplitude of the 3 kHz peak is increased by overlapping the peak of the abnormal sound with the harmonic peak of the buzzer sound.

  Simultaneously with the acquisition of the sound pressure data in step S101A, the vibration sensor 7 acquires the vibration generated in the work 2 and the buzzer 5 in step S101B. In step S102B, the vibration data is digitally converted by the second input circuit 21 and fast Fourier transformed by the second Fourier transformer 22 to be transformed into a frequency spectrum. Further, in step S103B, the vibration data is normalized by the second normalization unit 23 so that the maximum value of the amplitude becomes “1”. Thereby, for example, a normalized spectrum FS2 of vibration data as shown in FIG. 4 is obtained. In FIG. 4, the horizontal axis represents frequency and the vertical axis represents amplitude, and shows a frequency change in the amplitude of vibration measured by the vibration sensor 7. The vibration normalization spectrum FS2 has a peak at 1 kHz of the buzzer sound. In the normalized spectrum FS2, the amplitude is normalized so that the peak at the frequency of 1 kHz, which is the maximum peak, has the maximum value “1”. Furthermore, peaks are also generated at 2 kHz and 3 kHz corresponding to harmonics of the buzzer sound.

Here, the two normalized spectra FS1 and FS2 will be described.
If the casing of the buzzer 5 vibrates or the like when the buzzer 5 is sounded, abnormal noise is generated in addition to the original buzzer sound. Since the housing of the buzzer 5 vibrates, a vibration waveform also appears in the vibration sensor 7. That is, when an abnormal sound is generated as the buzzer 5 sounds, a similar reaction appears in the sound pressure data and the vibration data. For example, in the normalized spectra FS1 and FS2 of FIGS. 3 and 4, the buzzer 5 rings at 1 kHz, and the accompanying harmonics are generated at 2 kHz and 3 kHz. Peaks indicating harmonics of 2 kHz and 3 kHz are generated in each of the two normalized spectra FS1 and FS2.

  Further, when factory noise such as air blow noise occurs, the factory noise is taken into the microphone 6 and a peak occurs in the frequency spectrum. When the air blow sound is a noise having a frequency of 3 kHz, the peak of the normalized spectrum of the sound pressure in FIG. 3 becomes higher than the sound pressure corresponding to the factory noise, for example, about 15 db. Since the abnormal sound of the buzzer 5 also has a peak at 3 kHz, the peak in which the factory noise and the abnormal noise are superimposed is large. On the other hand, the air vibration accompanying air blow has no power to vibrate an individual such as the workpiece 2. Therefore, the vibration data is not changed by the air blow sound. Therefore, in the vibration normalization spectrum FS1 in FIG. 4, only a peak due to the buzzer sound appears at 3 kHz. From this, the peak height of 3 kHz of the vibration normalization spectrum FS2 is lower than the normalization spectrum FS1 of the sound pressure data.

  Here, when vibration is generated in the conveyor device 3, the vibration may be detected by the vibration sensor 7. In this case, a peak due to the vibration of the conveyor device 3 appears in the vibration normalization spectrum FS2. On the other hand, if no sound is generated even when the conveyor device 3 vibrates, the sound pressure data collected by the microphone 6 is not affected. For this reason, the peak resulting from the vibration of the conveyor apparatus 3 does not generate | occur | produce in the normalization spectrum FS1 of sound pressure data. As described above, in the normalization spectrum FS1 of sound pressure and the normalization spectrum FS2 of vibration, when data from the same measurement object are acquired at the same time, one of the normalization spectra FS1 depending on the type of the external noise source. , Fluctuation occurs only in FS2.

  Next, in step S104 of FIG. 2, the minimum value selection unit 25 compares the two normalized spectra FS1 and FS2, extracts only the minimum value, and creates a reconstructed spectrum. Here, the two normalized spectra FS1 and FS2 are normalized so that the frequency of the maximum amplitude is 1 kHz, which is the frequency of the buzzer sound, and the peak value of 1 kHz, which is the maximum amplitude, is “1”. Can be directly compared. The minimum value of the two normalized spectra FS1 and FS2 is adopted because the smaller one of the two spectra is not affected by the external noise at all or is hardly affected by the external noise. It is because it is considered. In the example of the normalized spectra FS1 and FS2 in FIGS. 3 and 4, the influence of the factory noise appears only in the sound pressure normalized spectrum FS1, and does not appear in the vibration normalized spectrum FS2. Therefore, if the peak value of the normalized spectrum FS2 having a relatively small peak value among the two normalized spectra FS1 and FS2 is employed, the influence of factory noise can be eliminated. Further, when external noise due to vibration of the conveyor device 3 is generated, the influence of the external noise due to vibration appears only in the frequency spectrum FS2 of vibration and does not appear in the frequency spectrum FS1 of sound pressure. Therefore, if the peak value of the normalized spectrum FS1 having a relatively small peak value of the two normalized spectra FS1 and FS2 is employed, the influence of vibration of the conveyor device 3 can be eliminated.

  Further, in step S105, the minimum value selection unit 25 constructs a spectrum (reconstructed spectrum) in which only the minimum values of the two normalized spectra FS1 and FS2 are selected. Thereafter, in step S106, the vertical axis of the reconstructed spectrum is converted into a sound pressure level using the sound pressure normalized spectrum FS1. As a result, a frequency spectrum of sound pressure from which noise caused by factory noise or external vibration is removed is obtained. Thereafter, in step S107, the determination unit 27 uses THD (Total Harmonic Distortion: Total High Frequency Distortion) and dBc (Relative ratio of high frequency to the fundamental wave), and the buzzer 5 and the work to which the buzzer 5 is attached. 2. Pass / fail judgment is performed.

  Here, a specific example of the minimum value selection process in step S104 and the spectrum reconstruction process in step S105 will be described with reference to the graph of FIG. 5, the table of FIG. 6, and FIG. FIG. 5 shows the two normalized spectra FS1 and FS2 superimposed on each other. Furthermore, the waveform from 2700 Hz to 3300 Hz on the horizontal axis is extracted and enlarged. FIG. 6 shows the result of extracting the minimum value for each frequency bin of the two normalized spectra FS1, FS2. FIG. 7 shows the reconstructed spectrum RF1, and the lower part shows an enlarged part of the frequency region of the upper part. 5 and 7, the horizontal axis indicates the frequency, and the vertical axis indicates the amplitude.

  As shown in FIG. 5, when the two normalized spectra FS1 and FS2 are displayed in an overlapping manner, depending on the frequency, there are a place where the amplitude of the normalized spectrum FS1 is large and a place where the amplitude of the normalized spectrum FS2 is large. In particular, at a frequency of 3000 Hz, the amplitude of the sound pressure normalized spectrum FS1 is larger than the amplitude of the vibration normalized spectrum FS2. This is because the factory noise is collected through the microphone 6 in addition to the buzzer sound in the sound pressure normalized spectrum FS1.

  As shown in FIG. 6, the minimum value selection unit 25 sets the frequency bin to 10 Hz, extracts the minimum value of the sound pressure and the amplitude of vibration within this range, and compares the two. For example, at 2700 Hz, the amplitude of sound pressure is -70 and the amplitude of vibration is -55. Therefore, at 2700 Hz, a sound pressure of -70 is selected. Similarly, vibration of −30 is selected at 3000 Hz, and vibration of −35 is selected at 3010 Hz. Then, when the normalized spectrum is reconstructed by adopting only the minimum values of the two normalized spectra FS1 and FS2, a reconstructed spectrum RF1 as shown in FIG. 7 is created. In the reconstructed spectrum RF1, the value of the normalized spectrum FS2 of vibration having a small amplitude is adopted for 3 kHz, so that the factory noise component included in the peak is removed and only the sound component from the buzzer 5 remains.

Further, a specific example of the sound pressure spectrum creation process in step S106 and the determination process in step S107 will be described with reference to FIG. FIG. 8 shows an example of a frequency spectrum FS0 of sound pressure that is a spectrum for inspection. The horizontal axis in FIG. 8 indicates the frequency, and the vertical axis indicates the sound pressure level.
When the multiplier 26 creates the sound pressure frequency spectrum FS0 as shown in FIG. 8 based on the reconstructed spectrum RF1 and the sound pressure normalization spectrum FS1, the determination unit 27 generates a buzzer based on the sound pressure frequency spectrum FS0. Evaluate 5. For example, THD or dBc is used for the determination process. For example, THD can be calculated as follows using the effective value V1 of the fundamental wave, the effective value V2 of the second harmonic, and the effective value V3 of the third harmonic.

  The smaller the THD value, the smaller the distortion. Therefore, if the THD value is equal to or less than a predetermined threshold, it is determined that the buzzer 5 and the work 2 are normal. On the other hand, when the THD value exceeds the threshold value, it is determined that the buzzer 5 or the work 2 is abnormal.

  As described above, since the inspection device 4 acquires vibration data simultaneously with sound pressure data and normalizes the frequency spectra of the two data, it is possible to directly compare the two types of frequency spectra. This makes it possible to create a frequency spectrum for inspection by adopting the minimum value of the two frequency spectra, and to remove external noise. It is possible to inspect an audio device that eliminates the influence of external noise without using a threshold value. In addition, since the minimum value is selected for each predetermined frequency bin, external noise can be removed while reducing the amount of information processing.

  All examples and conditional expressions given here are intended to help the reader understand the inventions and concepts that have contributed to the promotion of technology, and such examples and It is to be construed without being limited to the conditions, and the organization of such examples in the specification is not related to showing the superiority or inferiority of the present invention. While embodiments of the present invention have been described in detail, various changes, substitutions and variations can be made thereto without departing from the spirit and scope of the present invention.

2 Work 5 Buzzer 4 Inspection device 8 Data processing device 12 1st normalization part 22 2nd normalization part 25 Minimum value selection part 26 Multiplier (spectrum creation part)
27 Determination part FS0 Frequency spectrum of sound pressure (spectrum for inspection)
FS1, FS2 normalized spectrum

Claims (5)

A first normalization unit that normalizes the frequency spectrum of the sound output from the inspection object using the maximum value of amplitude;
A second normalization unit for normalizing a frequency spectrum of vibration generated around the inspection object using a maximum value of amplitude;
A selection unit that compares the normalized amplitudes of the two frequency spectra for each frequency and selects a smaller value of the amplitude;
A spectrum creation unit that creates a test spectrum consisting of a frequency spectrum of sound pressure using the data selected by the selection unit;
A determination unit for determining the presence or absence of abnormal noise generated from the inspection object from the peak height of the inspection spectrum;
An inspection apparatus for an acoustic device, comprising:   2. The inspection apparatus for an acoustic device according to claim 1, wherein the selection unit is configured to compare the amplitudes of two frequency spectra normalized for each frequency bin having a predetermined width for each frequency. Obtain the sound pressure data of the sound output from the inspection object,
Obtaining vibration data of vibrations generated around the inspection object simultaneously with the sound pressure data;
Normalize the frequency spectrum of the sound pressure data using the maximum amplitude,
Normalize the frequency spectrum of the vibration data using the maximum amplitude,
Compare the amplitudes of two normalized frequency spectra for each frequency, acquire the value with the smaller amplitude, and create a frequency spectrum of sound pressure that inspects the abnormal sound generated from the inspection object Device inspection method.   4. The inspection apparatus for an acoustic device according to claim 3, wherein amplitudes of the two normalized frequency spectra are compared for each frequency bin having a predetermined width. Obtain the sound pressure data of the sound output from the inspection object,
Obtaining vibration data of vibrations generated around the inspection object simultaneously with the sound pressure data;
Normalize the frequency spectrum of the sound pressure data using the maximum amplitude,
Normalize the frequency spectrum of the vibration data using the maximum amplitude,
Comparing the amplitudes of the two normalized frequency spectra for each frequency, obtaining a value with the smaller amplitude, and causing the computer to execute a process of creating a frequency spectrum of sound pressure for inspecting the abnormal sound generated from the inspection object An inspection program for an acoustic device.

Images