JPH08297044A - Weight sorting system - Google Patents

Abstract

PURPOSE: To obtain a weight sorting system wherein the noise, which can not be removed easily by conventional digital filter, is removed. CONSTITUTION: The weight sorting system comprises an A/D converter 26 for converting an analog weight signal into a digital weight signal at a variable sampling period, a frequency analysis means 32 for detecting the frequency component of digital weight signal, and means 33 for setting the sampling period of A/D converter 26, wherein the sampling period of A/D converter 26 can be set based on the results of frequency analysis.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a weight sorting apparatus for weighing articles during transportation and sorting the articles based on the weight value.

[0002]

2. Description of the Related Art Conventionally, a weight sorter has been used to measure the weight of an article manufactured by a production line while the article is being conveyed and inspect the weight of the article. FIG. 3 is a block diagram of a conventional weight sorting apparatus. In FIG. 3, reference numeral 1 denotes a carry-in conveyor, which conveys the article W to the weighing conveyor 2.
The weighing conveyor 2 is equipped with a weight detector 3,
An analog weight signal corresponding to the weight change of the article W is output to the analog filter 4.

The analog filter 4 is a low-pass filter composed of an OP amplifier, a resistor and a capacitor, and attenuates a noise component contained in the analog weight signal and outputs it to the amplifier 5. The amplifier 5 is a non-inverting amplifier circuit composed of an OP amplifier and a resistor, amplifies a minute analog weight signal, and outputs it to the A / D converter 6.

The A / D converter 6 converts the analog weight signal into a digital weight signal at a constant cycle (sampling cycle) and outputs it to the digital filter means 7. The digital filter means 7 removes the noise component of the digital weight signal received by the calculation method described later, and the weight calculation means 8
Output a digital weight signal to. In the weight calculation means 8,
A weight value is calculated based on the digital weight signal and output to the determination means 9.

The judging means 9 judges whether or not the calculated weight value is within a predetermined range, and outputs a selection signal according to the result. This sorting signal is output to the sorting machine 10 provided on the carry-out conveyor 11, and the weighed articles W carried out from the weighing conveyor 2 are responsive to this sorting signal.
They are transported to different transport routes and sorted.

The digital filter means 7 is a communication FI.
It is an R type low-pass filter, and a time series digital weight signal input at a constant cycle is calculated by the following equation (1).

[0007]

[Equation 1] The filter coefficient h [n] in Expression (1) is a factor that determines the filter characteristic, and is calculated and set in advance by Expression (2).

[0008]

[Equation 2] The cut-off frequency fc of the digital filter is usually set to be ½ or less of the noise frequency to be removed.

N is the time when data can be sampled,
That is, it is limited by the time that the article is on the weighing conveyor. That is, as shown in FIG.
The time Ts = (Lc−Lw) / v from the state in which the article W having the length Lw is completely transferred onto the weighing conveyor 2 of c to the time when the tip of the article reaches the discharge side end of the weighing conveyor 2 is measured. This is a possible time, and it is necessary to perform a predetermined number of samplings during this time. Therefore, n must be set so that nT ≦ Ts.

[0010]

FIG. 2 shows the filter characteristics of the digital filter set at the cutoff frequency f0. In FIG. 2, the frequency of noise is f
When it is a, the effect of the filter is exerted and the noise component can be sufficiently attenuated, but when the noise frequency is fb, the noise component cannot be sufficiently attenuated. That is, there is a problem that a sufficient filter effect cannot be obtained depending on the noise frequency. The present invention
Based on such a current situation, it is an object of the present invention to provide a weight selection device capable of removing noise components more accurately than ever with respect to any noise.

[0011]

In order to solve the above-mentioned problems, in the weight sorting apparatus of the present invention, the weight of the article carried from the pre-stage conveyor to the weighing conveyor is carried by the article on the weighing conveyor. A weight detector that detects the weight of an article and outputs an analog weight signal in a weight sorting device that measures in-between and performs sorting determination of the article based on the measured weight, and carries out the sorting-determined article to a post-stage conveyor. An A / D converter with a variable sampling period for converting the analog weight signal into a digital weight signal, weight value calculating means for calculating the weight of the article from the digital weight signal, and detecting a frequency component of the digital weight signal. And a sampling cycle setting means for setting the sampling cycle of the A / D converter.

[0012]

With this configuration, noise generated in the weight measuring system can be clearly known, and noise of any frequency can be accurately removed by changing the sampling cycle of the A / D converter. .

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an embodiment of the present invention, in which reference numeral 21 denotes a weighing belt conveyor which is supported by a weight detector 23. The analog weight signal output from the weight detector 23 is used as the analog filter 2
4, input to the A / D converter 26 via the amplifier 25.

The A / D converter 26 converts the analog weight signal into a digital weight signal at a cycle preset by the sampling cycle setting means 33, which will be described later, and outputs it to the weight value calculating means 27. The weight value calculating means 27 converts the input digital weight signal into the sampling cycle setting means 3
The number of points set by 3 is stored in the data storage means 28. Then, the weight value is calculated from the average value of the digital weight signal and output to the determination means 29.

The judging means 29 judges whether or not the inputted weight value is within a preset weight range, and outputs a selection signal according to the result. This sorting signal is output to the sorting machine 31, and the weighed articles W carried out from the weighing conveyor 22 are transported to different transport routes and sorted according to the sorting signal.

The input means 30 is composed of a keyboard, an external interface, etc., and sets the weight information used by the determination means 29 in the data storage means 28 and switches between the weight value calculation mode and the frequency analysis mode. There is. When the frequency analysis means 32 is switched to the frequency analysis mode, the number of digital weight signals preset by the input means 30 (N) is stored in the data storage means 28, and the frequency component is analyzed by the following method. . First, N
When the average value bar X of the individual digital values (X [0], ..., X [N-1]) is calculated and the average value is subtracted from the individual values, the formula (3) is obtained.

[0017]

(Equation 3) Next, for y [0], y [1], ..., Y [N-1],
Equation (4) is summarized and Fourier transform is performed.

[0018]

[Equation 4] k is an index indicating the frequency, and the A / D converter 2
When the sampling cycle of 6 is T, Y [k] is a frequency component at frequency k / NT. Where j ² =-
Since equation (4) is a complex number because of the relation of 1, the frequency component of equation (5) can be expressed as Re {Y [k]} for the real part and Im {Y [k]} for the imaginary part. Similarly, it can be expressed as the absolute value of Y [k].

[0019]

(Equation 5) FIG. 4 shows the result of frequency analysis by this method. There is a peak at the position of the frequency f1, and the position of this peak represents the frequency of noise. The frequency analysis unit 32 detects this peak and outputs the frequency f1 as a noise frequency to the sampling period setting unit 33. The sampling period setting means 33 executes the calculation of equation (6) in order to remove the noise of the frequency f1, and A / A
The sampling period of the D converter 26 is calculated. The magnification m in the equation (6) is an even number of 2 or more, and this value is preset by the input means 30.

[0020]

(Equation 6) Here, the reason why the sampling period is set as in Expression (6) will be described below. FIG. 5 is a waveform of noise of frequency f1, and the instantaneous value Z (t) at time t is expressed by equation (7).

[0021]

(Equation 7) Now, if the noise is sampled for one cycle from t = 0 in the cycle of the equation (6), that is, if the cycle of the noise is equally divided into m in FIG. 5 and the instantaneous value at each time divided into m is obtained. It is represented by Expression (8) to Expression (12).

[0022]

(Equation 8)

[0023]

[Equation 9]

From the above, the noise component becomes 0 when the noise frequency f1 is multiplied by an even number (m times), sampling is performed at a period (1 / mf1) which is the reciprocal thereof, and the sum total is obtained for one period. Therefore, if the analog weight signal including the frequency f1 is A / D converted in the cycle 1 / mf1 and the obtained digital weight signals are summed and averaged in one cycle, the noise component is removed and only the DC component is extracted. You can do it.

Now, the operation of the weight sorting apparatus having the above-mentioned structure will be described. It is assumed that the number N of digital weight signals captured and the magnification m in the frequency analysis mode are preset by the input means 30. When the weight value calculation mode is switched to the frequency analysis mode by the input means 30, the frequency analysis means 32 operates to store N digital weight signals in the data storage means 28.
Then, frequency analysis is performed on the N digital weight signals, the peak of the frequency component is detected, and the sampling period setting means 3 uses the frequency at this position as the frequency of noise.
Output to 3.

The sampling period setting means 33 multiplies the noise frequency by m, calculates the reciprocal thereof, and sets it as the sampling period of the A / D converter 26. When the sampling cycle of the A / D converter 26 is set, the mode automatically returns to the weight value calculation mode. In the weight value calculation mode, the A / D converter 26 performs A / D conversion at the cycle set by the sampling cycle setting means 33 and outputs a digital weight signal to the weight value calculation means 27. The weight value calculation means 27 stores m digital weight signals in the data storage means 28, and calculates the weight value from the average value of the digital weight signals.

When the weight value is calculated, the judging means 29 judges whether the quantity is an excessive quantity, a positive quantity, or an appropriate quantity, and outputs a selection signal according to the result. The sorting signal is output to the sorting machine 31, and the articles are transported to different transporting paths according to the sorting signal to perform sorting.

[0028]

As described above, in the weight sorting apparatus of the present invention, the weight of the article carried into the weighing conveyor from the pre-stage conveyor is measured while the article is being conveyed on the weighing conveyor. In the weight sorting device that carries out the sorting determination of the articles based on the measured weight, and carries out the sorting-determined articles to the subsequent conveyor, a weight detector that detects the weight of the articles and outputs an analog weight signal, and the analog weight signal. A with variable sampling period
/ D converter, weight value calculation means for calculating the weight of the article from the digital weight signal, frequency analysis means for detecting the frequency component of the digital weight signal, and the A / D conversion based on the frequency analysis result The sampling cycle setting means for setting the sampling cycle of the converter is provided, and the sampling cycle of the A / D converter is set based on the frequency analysis result. This allows
The digital filter according to the conventional technique can also remove noise that cannot be sufficiently attenuated due to its nature. Further, the data processing method is simpler than the noise removal method using the conventional digital filter.

[Brief description of drawings]

FIG. 1 is a block diagram showing a circuit configuration of a weight sorting device according to the present invention.

FIG. 2 is a diagram showing frequency characteristics of a digital filter.

FIG. 3 is a block diagram showing a conventional weight selection circuit configuration.

FIG. 4 is a diagram showing a waveform of noise of frequency f1.

FIG. 5 is a diagram in which a noise cycle is divided into m equal parts.

FIG. 6 is a diagram in which articles are conveyed on a weight sorting device.

[Explanation of symbols]

 1 Carry-in conveyor w Items to be weighed 2,21 Weighing conveyor 11,22 Carry-out conveyor

Claims (1)

[Claims] 1. The weight of an article carried into the weighing conveyor from the preceding conveyor is measured while the article is being conveyed on the weighing conveyor, and the sorting judgment of the article is performed based on the measured weight, and the sorting is performed. In a weight sorting device for delivering judged goods to a downstream conveyor, a weight detector for detecting the weight of the goods and outputting an analog weight signal, and a sampling cycle variable A / D for converting the analog weight signal into a digital weight signal A converter, a weight value calculation means for calculating the weight of the article from the digital weight signal, a frequency analysis means for detecting a frequency component of the digital weight signal, and the A based on the frequency analysis result.
A weight selecting apparatus comprising: sampling period setting means for setting a sampling period of the A / D converter, and setting the sampling period of the A / D converter based on the frequency analysis result.