JPH07280630A - Weight sorting machine - Google Patents

Weight sorting machine

Info

Publication number
JPH07280630A
JPH07280630A JP9582794A JP9582794A JPH07280630A JP H07280630 A JPH07280630 A JP H07280630A JP 9582794 A JP9582794 A JP 9582794A JP 9582794 A JP9582794 A JP 9582794A JP H07280630 A JPH07280630 A JP H07280630A
Authority
JP
Japan
Prior art keywords
weight
weighing
correction
value
step
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
JP9582794A
Other languages
Japanese (ja)
Other versions
JP3427946B2 (en
Inventor
Toru Takahashi
孝橋  徹
Original Assignee
Yamato Scale Co Ltd
大和製衡株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yamato Scale Co Ltd, 大和製衡株式会社 filed Critical Yamato Scale Co Ltd
Priority to JP09582794A priority Critical patent/JP3427946B2/en
Publication of JPH07280630A publication Critical patent/JPH07280630A/en
Application granted granted Critical
Publication of JP3427946B2 publication Critical patent/JP3427946B2/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Abstract

PURPOSE:To eliminate the weight measurement error due to drift of zero point and span and to prevent occurrence of a defective product by calculating a correction value using a newest corrected weight measurement signal at all times and determining a corrected weight measurement signal using an updated correction value thereby obtaining a highly accurate weight measurement signal. CONSTITUTION:Tare weight (Wi) and zero point variation Wz are prestored in a memory 18 and a counter (n) is reset along with six stage registers 20, 22. When articles 4 are fed sequentially from a conveyor belt 2 onto a weighing belt 6, a load cell 8 weighs the article 4 and delivers a weight measurement signal Wx to a CPU 16. The CPU 16 extracts the tare weight Wi from the weight measurement signal Wx to produce a weight measurement signal Wb1 which is then corrected by a primary correction value Wf1 determined based on six preceding weight measurement signals Wb1, to produce a primary corrected weight measurement signal Wb2'. A decision is then made whether the Wb2' comes within a prescribed range and if it comes within the prescribed range, a secondary correction value Wf2 is calculated including the current Wbl which is subsequently corrected by the Wf2 to produce a secondary corrected weight measurement signal Wb2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a weight sorter for weighing an article and determining whether the article is a non-defective article based on the weighted value.

[0002]

2. Description of the Related Art Generally, a weight sorter has a conveyor equipped with a weight sensor, that is, a weighing conveyor in order to weigh articles. Double signals can drift. In addition, articles may bring dust or the like on the weighing conveyor. As a result, the zero point or span of the weighing signal shifts during operation of the weight sorter. Regarding the zero point drift, the so-called automatic zero point correction function automatically detects that the supply of articles to the weighing conveyor is interrupted, and the weighing signal at that time (that is, in the unloaded state) is zeroed. The weight value is stored as a weight value and automatically corrected by subtracting the zero-point weight value from the weight signal at the time of weighing the article.

[0003]

By the way, in the weighing conveyor of the weight sorter, the conveyor is supported on the weight sensor, and the weight WI is a relatively large part of the capacity range of the weight sensor. is occupying. That is, a relatively large tare load is applied on the weight sensor.
The relationship between the input weight and the indicator weight value in such a case is shown as a straight line m in FIG. The indicator weight value DWI obtained for the tare weight WI is used as the initial zero point weight value.

In this state, if the ambient temperature of the weight sorter changes, especially in the transitional period of temperature change, even if the weight sensor is temperature-compensated, it is compensated due to the difference in temperature time constant between the compensation characteristic and the sensor output characteristic. The balance is lost and, for example, as shown by a straight line m1 in the figure, a drift occurs in the zero point and the span of the indicator weighing value. In such a case, for the same tare weight WI, the indicator weight value is DWI1 instead of DWI1.
Therefore, the zero point has drifted by DWI1-DWI. This is done by measuring DWI1 when the article is unloaded and using this as a new zero weighing value, or
The zero point is corrected by measuring I1-DWI, storing it as a zero point drift amount, and calculating a new zero point weight value by calculating DWI + (DWI1-DWI). As a result, the straight line m1 is shifted by (DWI1-DWI) as shown by the dotted line m2 in the figure.

With the zero correction thus performed, when an article of weight Wx is carried onto the weighing conveyor, if the drift does not occur, the weighing signal is DWIx on the straight line m.
However, if there is a drift, DWIx1 on the straight line m1
Becomes However, since the zero point correction is performed, DWIx
1- (DWI1-DWI). However, as is clear from the figure, this does not become a value equal to DWIx. This is because the drift of the zero point is corrected by the zero point correction, but the drift of the span is not corrected. Therefore, a weighing error occurs.

The above description is an example of the case where the zero point correction is performed. However, due to the device for feeding the article to the weighing conveyor, the automatic zero point correction is performed when the article is continuously carried in for a long time. DW not done for a long time
Even the error of I1-DWI is not corrected, which causes more and more weight error.

The drift of such zero point and span is
It will be resolved as the ambient temperature and the temperature of the weight sensor stabilize, but if the error continues for a fairly long time, or if a large error suddenly occurs for a short time, some measures are required. Is. In addition, the zero point drift due to dust attached to the weighing conveyor is
Unless the zero-correction function works, it will not be resolved, and some countermeasures are necessary against this.

An object of the present invention is to eliminate the weighing error due to the drift of the zero point and the span as described above, and the article to be weighed is composed of a plurality of necessary articles. If there is a missing item in the above necessary items, a weight difference of a certain value or more occurs, and this weight difference is
It is sufficiently large in view of the variation range of the weight of the article itself, in which all the necessary items are included. Such articles include, for example, a tablet sheet of a pharmaceutical product, a plurality of tablet sheets placed in a box, and an instruction sheet placed in such a box. The case where there is a tablet missing from the tablet sheet, the case where a part of the tablet sheet comes out of a plurality of boxed tablet sheets, and the case where the manual comes out of the box correspond to the above-mentioned missing item.

Then, if such an article is sent to a weight sorter and there is a weight difference of a certain value or more, it is judged that there is a stockout. However, at the time of warming up during the weight sorting period, a weight sensor and a measuring circuit are provided. If the amount of drift that occurs when the room temperature is undergoing a sudden change after the start of cooling and heating during the transition period of internal temperature rise of N becomes a size that cannot be ignored for the weight difference above the certain value , Even though the accurate weight is within the range of the weight difference of the above constant value, if the weighing signal is affected by the drift and the weight difference exceeds the above certain value, a defective product occurs. To do. The present invention aims to prevent this.

[0010]

In order to achieve the above object, the present invention provides that the weight value of a non-defective product is within a predetermined allowable range, and the permissible range is more than the variation of the weight of the non-defective product. Large articles are supplied one after another, and a weighing means for generating a weighing signal representing the weight of the articles and a primary correction value based on the plurality of weighing signals input up to the previous time are used to input the above A first correcting means for outputting a primary correction weighing signal obtained by correcting the weighing signal; and when the primary correction weighing signal is within a predetermined specified value range, the weighing signal inputted this time and the above A correction value calculating means for calculating a secondary correction value based on a plurality of weighing signals input up to the previous time, and a secondary correction weighing signal obtained by correcting the weighing signal input this time with a secondary correction value. And a second correction weight signal or a second correction weight signal. No. is what the determining means for determining whether or not there is within the allowable range comprises.

Further, the first-order and second-order correction values can be obtained based on the least-squares method, and the first-order and second-order correction values can also be used as the average value of the respective weighing signals. It should be noted that after the weighing signal is corrected with the primary correction value, it may be judged by the judging means whether or not it is within the allowable range. in this case,
It is desirable to obtain the primary correction value based on the least squares method.

[0012]

According to the present invention, the current weighing signal of the article generated by the weighing means is corrected by the first correcting means on the basis of the primary correction value to become a primary corrected weighing signal. If this primary correction weighing signal is within the specified value range, a new 2 is added including the weighing signal of the current article, that is, including the latest data.
The correction value calculation means calculates the next correction value. And this 2
The weight signal of this time is corrected again using the next correction value.
The next correction weight signal is used. Since this secondary correction weighing signal is corrected using the latest data, the accuracy of correction is higher than that of the primary correction weighing signal. Then, the determining unit determines whether the primary correction weighing signal or the secondary correction weighing signal is within the allowable value range, and performs selection.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment is shown in FIGS. The weight sorter of this embodiment, as described above, weighs a tablet sheet of a pharmaceutical product, a plurality of tablet sheets put in a box, or a box in which instructions are put in such a box, and a weight value is obtained. Is an allowable range W defined by the standard weight Ws and the allowable value We.
It is determined whether or not there is a stockout depending on whether or not it is within s ± We. Of course, the variation of the weight value itself is
It is smaller than ± We.

In this weight sorter, as shown in FIG. 2, the above-mentioned article 4 is carried into the weighing conveyor 6 by the carry-in conveyor 2 almost every predetermined time. In order to detect this carry-in, an article detector, for example, a photoelectric detector 7 is provided between the carry-in conveyor 2 and the weighing conveyor 6. The photoelectric detector 7 is used to determine whether it is the zero point correction timing. In this weighing conveyor 6,
A weighing unit, for example, a load cell 8 is provided, and each time an article 4 is carried into the weighing conveyor 6, an analog weighing signal representing the weight of the conveyor itself of the weighing conveyor 6 and the article 4 is generated. To do. Although not shown, the load cell 8 is provided with a temperature compensation circuit.

The analog weighing signal of the load cell 8 is amplified by the amplifier 10, converted into a digital weighing signal by the A / D converter 12, and supplied to the CPU 16 through the input / output port 14. The amplifier 10 and the A / D converter 12 constitute a measurement circuit attached to the load cell 8, and a temperature compensation circuit is also provided to these. However, for the reasons described above, the analog weighing signal may include drift amounts such as zero and span variations.

A memory 18 is attached to the CPU 16, and a weight shift register 20, a weight count shift register 22, a weight count counter n, and the like, which will be described later, are included in the memory 18. In addition to this, the CPU 16 uses the input / output port 14
The numeric keypad 26 and the display device 28 are connected via the numeric keypad 26, and the numeric keypad 26 is, for example, the standard weight Ws or the allowable deviation W described above.
It is used to set e, and the display device 28 displays the selection result,
The estimated weight value described later is displayed.

The processing performed by the CPU 16 will be briefly described. Now, when power is applied to this weight sorter, the CPU 16 performs the calculation represented by Formula 1 in order to represent the weight of the article on the weight conveyor 6. That is, the tare subtraction is performed in consideration of the zero point variation.

## EQU1 ## Wb1 = K (Wx-Wi-Wz) Wb1 is an indicator weight value, K is a proportional coefficient, Wx is a weight signal,
Wi is a stored value in the tare register, and Wz is a stored value in the zero point drift amount register. Wi and Wz are set to zero when the power is applied, and immediately after that, the weighing signal Wx is read and stored in Wi. As a result, Wi = Wx, Wz =
It becomes 0 and Wb1 becomes 0. After that, when calculation is performed based on Equation 1, Wb1 represents the weight of the article on the weight conveyor 6 unless there is a drift. If the zero point drifts here, Wi = Wx is not established and a weighing error occurs. Therefore, when the photoelectric detector 7 does not detect an article, Wx
The value of −Wi is stored in Wz, and the fluctuation of the zero point is corrected by performing the calculation of Expression 1. Even with this, the span drift cannot be corrected as described above.

Therefore, in this weight sorter, the tendency of the drift of the weighing signal is judged and corrected. The correction will be described with reference to FIG. It is assumed that the indicator weight values W1, W2, ... Are calculated based on the equation 1 for the plurality of articles 4 sequentially conveyed to the weight conveyor 6. It is assumed that the indicator weight value tends to increase due to the above-mentioned drift, for example. In order to determine the tendency of this drift, when the number of samples for that purpose is set to 6, for example, a minimum of two is obtained from the number of times of weighing (1, 2, ...) And the indicator weighing value (W1, W2 ... The regression line L1 is obtained by the multiplication method. If there is no variation in the weight of the article and there is no random noise during measurement,
It is considered that the indicator weight value changes along the regression line L1 only by the drift.

Therefore, if there is no variation in the weight of the article, the indicator weight value in the seventh measurement drifts to W71 determined by the regression line L1. Therefore, the first-order estimated correction value, that is, the drift amount Wf1 is calculated by the equation 2.

## EQU00002 ## Wf1 = W71-Ws By using this Wf1, a correct weight value Wb1 is obtained by correcting the influence of the drift of the seventh designated weighing value by the equation 3.

## EQU3 ## Wb2 = Wb1-Wf1

Then, the regression line L2 is obtained again by using the present measured weight value W7 and the measured weight values W2, W3 ... W6 so far, that is, the latest sample data. However, since the weight fluctuation of the article due to the lack of goods should not affect the determination of the drift tendency, whether Wb2 is within the range of the allowable deviation ± We from the reference weight Ws, that is, WW.
It is determined whether b2 satisfies the equation 4, and only when it is satisfied,
A regression line L2 is obtained.

[Formula 4] Ws−We ≦ Wb2 ≦ Ws + We

An estimated weight value W72 based on the regression line L2 is obtained, and a second-order estimated correction value Wf2 is obtained from equation (5).

## EQU00005 ## Wf2 = Ws7-Ws Then, the correct indicator weight value Wb is obtained by the equation 6.

## EQU6 ## Wb = W7-Wf2

As described above, every time one article 4 is weighed, the trend of the N (for example, 6) indicator weight values up to the previous time is found by the regression line L1 and is found from this. One of the features of the present invention is that the first-order estimated weight value Wb1 at the time of the current weighing is obtained using the first-order estimated correction value Wf1 and an appropriate amount is determined. , The current weight value is added to the sample for newly obtaining the correction value, the regression line L2 is obtained again with the latest data, and the quadratic estimated correction value Wf2 obtained from this is finally used to finally obtain the correct value of this time. Another feature of the present invention is that the weight value is obtained.

The least squares method is used because it is possible to obtain an estimated reference value that follows the change trend of the weighing signal without delay. Further, since the regression line is obtained by the least squares method, that is, the change trend of the weighing signal is estimated by a linear expression, but n (n is a positive integer of 2 or more) It is also possible to estimate the change trend of the weighing signal. However, in general, a linear equation is sufficiently practical.

The processing executed by the CPU 16 for that purpose is shown in the flow chart of FIG. The tare weight is previously stored in Wi and the zero point variation is stored in Wz, the counter n for counting the number of times of weighing is reset, and the weight shift used to obtain the regression lines L1 and L2. The register 20 and the weight count shift register 22 each have 6
It's a tier and it's reset. It is assumed that the article 4 is continuously fed to the weighing conveyor 6 at every predetermined time, and zero correction is not performed during this period.

First, it is determined whether or not it is the dynamic weighing mode, that is, the mode in which the weighing conveyor 6 is weighed in a running state (step S2). If not, the process proceeds to another processing routine, and the dynamic weighing mode is set. Then, the weight data Wx is input (step S4), and the indicator weight value Wb1 from which the influence of the zero point fluctuation is removed is calculated based on the equation 1 (step S6).

Then, it is judged whether or not the regression line L1 exists by the weights up to the previous time (step S8). Since it does not exist at the beginning, Wb1 is regarded as corrected by Wf1 and stored in Wb2 (step S1).
0), the value of the weight counter n is advanced by 1 (step S
12).

Next, it is judged whether or not this Wb2 exists within the permissible ranges Ws-We and Ws + We (step S14). If not, it is processed as a defective product (step S).
16) and returns to step S2. If Wb2 is within the allowable range, it is treated as a good product (step S
18), the weight count of the weight counter n, the indicator weight value W
b1 is input to the weight count shift register 22 and the weight shift register 20, respectively (step S2).
0). The non-defective product processing and the defective product processing are performed, for example, by a sorting device (not shown) provided at the subsequent stage of the weight conveyor 6 in the direction of the non-defective product and in the direction of the non-defective product. It means to distribute the articles 4 in different directions.

Then, it is judged whether or not each of the registers 20 and 22 stores data for six times (step S22). If the data for six times are not available, the process returns to step S2. Therefore, if all of the first to sixth indicator weight values fall within the allowable range, steps S2, S4, S6, S8, S10, S12, S14,
The loop of S16, S18, S20 and S22 is repeated 6 times.

At this time, in step S22, it is determined that 6 pieces of data are stored in each of the registers 20 and 22, and a regression line is obtained from these 6 pieces of data (step S24). Next, the estimated weight value W6 for this time, that is, the sixth time, is obtained from the values of the regression line and the weight counter n, and the estimated correction value Wf2 is obtained (step S26). Further, using this estimated correction value Wf2, the final final weight value is obtained (step S2).
8). Then, the process returns to step S2 through steps S30 to S36 described later.

In the seventh weighing, steps S2, S4, S6 and S8 are executed in the same manner as described above,
In step S8, since the regression line has been obtained the previous time (sixth time), the primary estimated weight value is calculated from this and the current weight count obtained by adding 1 to the value (6) of the weight count counter n. (W71 in this case) is calculated, and the correction value Wf1 is calculated using this (step S38). And this one
The corrected primary indicator weight value Wb2 is obtained using the next estimated correction value Wf1 (step S39). Next, in step S12, the value of the weighing counter n is incremented by 1, and in step S12
14 is executed. Here, when it is determined that the product is non-defective, a regression line is obtained in step S24 through steps S18, S20, and S22, and in step S26,
Using this and the number of times of weighing (7 in this case), a second-order estimated weighing value (W72 in this case) is further obtained, and a second-order estimated correction value Wf2 is further obtained, and in step S28, the seventh weighing is performed. A second order estimate of the value is made. This secondary estimated weight value is
It is displayed on the display device 28. The same operation is performed thereafter.

In the above example, in order to quickly judge whether the product is a good product or a defective product, the primary estimated weight value is set to the allowable range Ws ± We.
However, in the case of more accurately determining whether the product is a good product or a defective product, after executing step S28, the product falls within the allowable weight range defined by the allowable lower limit weight Ws-WL and the allowable upper limit weight Ws + WU. It is determined whether or not Wb is included (step S40). If it is included, non-defective product processing is performed (step S42). If it is not included, defective product processing is performed (step S44), and then step S32.
You only have to execute the following. In this case, step S18 is omitted, and if it is determined in step S22 that the data for six times is not complete, non-defective processing is performed and then the process returns to step S2. In this case, step S
The weight range defined by ± We used in 14 may be the same as or different from the weight range defined by WL and WU.

Next, the reason why steps S30 to S36 are provided will be described with reference to FIGS. 3 (b) to 3 (d). If all of the indicated weighing values W1 to W6 are within the permissible range, each weighing number is stored in each stage of the weighing number register 22 as shown in FIG. When the weight is measured on the screen and it also falls within the allowable range, the same figure (c)
As shown in, the final stage value changes from 1 to 2. After that,
If this is left as it is, the value of the final stage will gradually increase and eventually overflow. In order to avoid this, in step S30, it is determined whether or not the value P of the final stage is 1, and if it is 1, the process returns to step S2, and if it is not 1, the value of the final stage is set to 1. Required value Q (=
P-1) is calculated (step S32), Q is subtracted from the value of each stage of the weight count shift register 22, and the value of the final stage is corrected to 1 as shown in FIG. Step S34). In response to this, Q is also subtracted from the value of the weighing counter n (step S36). Then, the process returns to step S2. This prevents overflow of the weight count shift register 22.

FIG. 4 shows a flowchart of the second embodiment. In the first example, it was assumed that the articles 4 were constantly fed to the weight conveyor 6 and zero adjustment was not performed. However, in practice, zero correction may be performed on the way. For example, the zero fluctuation is set to WZ0 at the start of weighing,
Even if there is a zero variation WZ1 until time t1, this is corrected by the first and second-order estimated correction values Wf1 and Wf2 at time t1, so that the weight of the article is immediately measured at time t1. However, the indicator weight is K (Wx-Wi-
There is no problem because it is performed by calculating WZ0) -Wf1 (or Wf2).

However, between the time t1 and the time t2, the zero point fluctuates even though the weighing of the article is not performed, and if the zero point variation amount at the time t2 is WZ2, the weighing of the article remains as it is. Moving to, the estimated correction value Wf1 or Wf2 estimated at time t2 does not include the correction for the change from WZ1 to WZ2. Therefore, K (Wx-Wi-WZ0) -Wf1 (or Wf
If the calculation of 2) is performed, the change from WZ1 to WZ2 cannot be corrected. Therefore, Wf1 is set to K (WZ
Only 1-WZ0) is corrected.

Therefore, as shown in FIG. 4, step S2
After determining whether it is the dynamic mode in the zero point timing,
That is, it is determined whether the photoelectric detector 7 detects the article 4 (step S46). If it is the zero point correction timing, W
Wi is subtracted from x to obtain the fluctuation amount Wz of the zero point (step S48). Then, it is determined whether the flag F, which is set to 1 when the zero point adjustment is performed, is 1 (step S50). If flag F is not 1, Wz is set to Wz0
Store in register (step S52) and set flag F to 1
Then (step S54), the process returns to step S4 to determine whether it is the zero point timing. If the zero point timing remains, the zero point variation amount is obtained again in step S48, and it is determined in step S50 whether the flag F is 1, but since this time is 1, the zero point variation amount Wz is set to W.
It is stored in the zn register (step S56), and the process returns to step S46. Therefore, the zero point variation amount when the zero point timing is started is stored in the Wz0 register, and the zero point variation amount at the end of the zero point timing is stored in the Wzn register.

When the zero point timing ends, W
It is determined whether z0 and Wzn are equal (step S5).
8), if they are equal to each other, it is not necessary to correct the zero point variation, and thus the process proceeds to step S6 shown in FIG. If they are not equal, the flag F is set to 0 (step S60), Wb1,
Wf1 is calculated (step S62). However, the zero point variation amount Wz used to calculate Wb1 is the zero point variation amount immediately before Wz0.

Next, K (Wzn-Wz) is added to the obtained Wf1.
The value obtained by adding 0) is set as a new Wf1 (step S
64), and Wb1 is corrected with a new Wf1 (step S
66). Then, it is determined whether or not this is included in the range of Ws ± We (step S68). If it is included, non-defective product processing is performed (step S70), and if it is not included, defective product processing is performed (step S72). Following step S70 or 72, both shift registers are reset, the weighing counter n is set to 1 (step S74), n is stored in the weighing shift register 22, and the weighing value wb1 is stored in the weight register 20, respectively. (Step S76). Then, the process returns to step S2. Note that Wb1 in step S76 is newly obtained by using Wzn in step S56 as the zero point variation amount. This is because it is possible to determine the latest fluctuation tendency of the zero point by starting over from the beginning.

FIG. 5 shows a third embodiment. First and second
The example of Example 1 used the number of times of weighing to determine the regression line. This is because it can be considered that the weighing of the article 4 is performed almost every time a fixed time elapses. In the third embodiment, not the number of times of weighing but the time at which the weighing is actually performed is measured, and the regression line is obtained using this.

Therefore, instead of the weight count shift register, a weight time shift register is used, and instead of the weight count counter n, a clock counter tx that counts, for example, every 1 msec. Is used. It should be noted that both the weighting time shift register and the weight shift register have six stages as in the first embodiment.

First, it is judged whether or not it is the dynamic weighing mode (step S80). If it is not the dynamic weighing mode, the process shifts to another processing routine, and if it is the dynamic weighing mode, the weight shift register and the weight time shift. The register is reset and the flag F1 for determining whether or not weighing is started is set to 0 (step S82).

Then, it is judged whether it is the weighing timing (step S84), and if it is the weighing timing, it is judged whether the F1 flag is 1, that is, whether the weighing of the article has already started (step S86). ). The processing when the timing is not the weighing timing, steps S85 and S87, will be described later. If the flag F1 is not 1, the value of the clock counter tx is set to 0 and the flag F1 is set to 1 (step S8).
8) Then, the indicator weight value Wx is input (step S90), and Wb1 is obtained by calculating K (Wx-Wi-Wz) as in the first embodiment (step S92). Then, the value of the clock counter tx is added to the value m0 of the first stage of the clock time shift register at this time to obtain the clock time Tx (step S94).

Next, it is judged whether or not the regression line already exists (step S96). If it does not exist, Wb1 is set to Wb2 (step S98), and Wb2 exists within the allowable range of Ws ± We. Judgment (Step S10
0), if it does not exist, defective product processing is performed (step S102), and the process returns to step S80. If Wb2 is within the allowable range of Ws ± We, non-defective processing is performed (step S104), Wb1 is input to the weight shift register, and Tx is input to the time shift register (step S106), and the time shift is performed. It is determined whether or not there is 0 other than the final stage of the register, that is, whether a total of 6 times are stored in the time shift register (step S108),
If six are not stored, the process returns to step S84.

When steps S80 to S108 are repeated and six pieces of weighing time and weight are stored in the weighing time shift register and the weight shift register, respectively, from step S108 to step S11 described later.
After 0 and 112, a regression line is calculated from these six weights and the weighted time (step S114), and the correction value Wf2 is obtained in the same manner as in the first embodiment (step S1).
16), Wb1 is corrected using this correction value (step S118), and the process returns to step S84.

Then, steps S84, 86, 90, 9
Step S96 is executed through steps 2 and 94. At this time, since the regression line already exists, the regression line and the current weighted time Tx are used to perform 1 as in the first embodiment. Obtain the next estimated correction value Wf1 (step S12
0), the primary estimated weight Wb2 is obtained using the primary estimated correction value Wf1 (step S122). Then, in step S100, it is determined whether the first-order estimated weight value Wb2 is within the range of Ws ± We. If it is not within the range, defective product processing is performed in step S102, and the process returns to step S84. If the first-order estimated weight value Wb2 is within the range of Ws ± We, after the non-defective product processing is performed in step S104, Wb1 and the weighting time Tx are set to the weight shift register and the weighting time shift register respectively in step S106. Then, step S108 is executed.

Thereafter, steps S110, 112, 11
4 and 116 are executed, but the regression line obtained here uses the current weighing value and the weighing time at this time, and is based on the latest data, so it is highly accurate and inevitable. As a result, the second-order estimated weight value is also highly accurate. Thereafter, the same operation is performed. Note that, similarly to the first embodiment, the secondary estimated weight value may be used to determine whether the product is a good product or a defective product.

If it is determined in step S84 that the timing is not the weighing timing, it is determined in step S85 whether the count value of the timing counter tx is equal to or greater than the predetermined upper limit value Tu. If it is not equal to Tu, the process returns to step S84. If it is Tu or more, in step S87, the weight shift register, the time shift register, and the time counter t
x is reset, the flag F1 is also set to 0, and step S84 is performed.
Return to. This is to prevent the time counter tx from overflowing if the articles are not weighed for a too long period.

Similarly, in order to prevent the time shift register from overflowing, steps S110 and S11 are performed.
Two are provided. That is, if the weighted time Tx is stored in the time shift register as it is, it overflows. Therefore, if the value of the final stage of the time register is other than 0 in step S110, the value of each stage of the time shift register is decreased by the value of the final stage in step S112.

FIG. 6 shows a fourth embodiment. First to third
In this embodiment, the least squares method is used to determine the drift tendency. This is best because there is no time delay. However, in some cases, even if there is a certain time delay, it can be used sufficiently. The fourth embodiment is used in such a case, and an average value, particularly a moving average is used instead of the least squares method.

Since the average value is obtained, the weight count shift register is not used. First, as in the first embodiment, steps S2, S4, and S6 are executed, and it is determined whether there is an average value based on the weights up to the previous time (step S).
8a), if there is no average value, Wb obtained in step S6
1 is set to Wb2 (step S10a), and it is determined whether or not this Wb2 is within the allowable range of Ws ± We (step S14a). If it is not within the allowable range, defective product processing (not shown) is performed. After that, the process returns to step S2. If it is within the allowable range, although not shown, Wb1 is put into the weight shift register after performing non-defective processing (step S20a). Then, it is judged whether or not there are six pieces of data in the weight shift register (step S22a), and if not, the process returns to step S4. If you have 6 pieces of data,
The average value Wav is obtained (step S24a), the difference between the average value Wav and the standard weight Ws is obtained, the correction value Wf2 is obtained (step S26a), the correction value Wf2 is used to perform the correction, and the process proceeds to step S2. Return (step S28
a).

When step S8a is executed again,
If the average value is calculated by the weight value up to the last time,
Since the previous correction value Wf2 is used as the primary correction value Wf1 for performing the primary correction, Wf2 is stored in Wf1 (step S38a), and Wb1 obtained in step S6.
Is subtracted from Wf1 (step S39a) to obtain a primary corrected weight value Wb2. Then, step S14a is executed, and if it is within the allowable range, step S20a is executed, and then step S22a is executed. However, since there are 6 pieces of data, steps S24a and S26a are executed to execute the secondary correction value Wf2. Is calculated, and the secondary corrected weight Wb is calculated. In this case, among the six weights stored in the weight shift register, the oldest one is discarded instead of adding the latest one weighed this time, and the moving average is taken as a result. There is. Thereafter, the same operation is performed. It should be noted that in this embodiment as well, the secondary corrected weight Wb may be used to determine whether the product is a good product or a defective product.

[0051]

As described above, according to the first aspect of the present invention, the weighting signal input this time is corrected by the primary correction value based on the plurality of weighting signals input up to the previous time. When the primary correction weighed signal is within the predetermined specified value range, the secondary correction value is calculated based on the weighing signal input this time and the plurality of weighing signals input up to the previous time. A secondary correction weighing signal calculated and corrected by this secondary correction value is obtained. Therefore, the secondary correction weight signal is obtained by using the latest data than the primary correction weight signal.
The estimation accuracy is high, and a highly accurate weighing signal can be obtained. In particular, this is effective in a case where the drift changes drastically, such as immediately after the weight sorter is operated, and the state of the drift change is significantly different between the previous time and this time. According to the second aspect of the present invention, since the above estimation is performed using the least squares method, there is no time delay in obtaining the estimated correction value, and the correction timing may be delayed. Absent. Further, according to the invention described in claim 3, by using a plurality of weighing signals input up to the previous time,
A correction value is obtained by the least squares method, and the current weighing signal is corrected by this correction value to obtain a corrected weighing signal.
Therefore, the corrected weighing signal can eliminate the influence of drift and the like, and can perform highly accurate selection.

[Brief description of drawings]

1 is a flow chart of a first embodiment of a weight sorter according to the present invention.

FIG. 2 is a block diagram of the first embodiment.

FIG. 3 is an operation explanatory diagram of the first embodiment.

FIG. 4 is a flowchart of the second embodiment.

FIG. 5 is a flowchart of the third embodiment.

FIG. 6 is a flowchart of the fourth embodiment.

FIG. 7 is a diagram showing a relationship between an input weight and an indicator weight value in a conventional weight sorter.

[Explanation of symbols]

8 load cell (weighing means) 16 CPU (first and second correction means, correction value calculation means, determination means)

Claims (3)

[Claims]
1. The weight value of a good product is within a predetermined permissible range, and the permissible range is sufficiently larger than the variation of the weight of the non-defective product. And a primary correction value based on a plurality of the weighing signals input up to the previous time, and a primary correction weighing signal obtained by correcting the weighing signal input this time are output. When the primary correction weight signal is within a prescribed value range determined in advance by the correction means, the secondary based on the weight signal input this time and the plurality of weight signals input up to the previous time. A correction value calculating means for calculating a correction value, and the above-mentioned weighing signal inputted this time are corrected by a secondary correction value.
A weight sorter provided with a second correction means for outputting a secondary correction weight signal and a determination means for determining whether the primary correction weight signal or the secondary correction weight signal is within the allowable range. .
2. The weight sorter according to claim 1, wherein 1
A weight sorter characterized in that the second and second correction values are obtained based on the least squares method.
3. The weight value of a non-defective product is within a predetermined allowable range, the allowable range being sufficiently larger than the variation of the weight of the non-defective product is supplied one after another, and a weighing signal representing the weight of the product. And a correction means for outputting a corrected weighing signal obtained by correcting the weighing signal input this time with a correction value based on the plurality of weighing signals input up to the previous time. A weight sorter, comprising: a determination unit that determines whether or not the weight signal is within the allowable range, and obtains the correction value based on the least squares method.
JP09582794A 1994-04-08 1994-04-08 Weight sorter Expired - Fee Related JP3427946B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP09582794A JP3427946B2 (en) 1994-04-08 1994-04-08 Weight sorter

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP09582794A JP3427946B2 (en) 1994-04-08 1994-04-08 Weight sorter

Publications (2)

Publication Number Publication Date
JPH07280630A true JPH07280630A (en) 1995-10-27
JP3427946B2 JP3427946B2 (en) 2003-07-22

Family

ID=14148242

Family Applications (1)

Application Number Title Priority Date Filing Date
JP09582794A Expired - Fee Related JP3427946B2 (en) 1994-04-08 1994-04-08 Weight sorter

Country Status (1)

Country Link
JP (1) JP3427946B2 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006170846A (en) * 2004-12-16 2006-06-29 Yamato Scale Co Ltd Rotating type weighting device
JP2008296201A (en) * 2007-06-04 2008-12-11 Anritsu Sanki System Co Ltd Weight sorting apparatus
WO2011044609A1 (en) * 2009-10-12 2011-04-21 Leonard Ian Burrell Belt image zero tracking system
CN102700739A (en) * 2012-06-20 2012-10-03 上海大和衡器有限公司 Quantitative packaging scale for sub-package according to amount of grains and quantitative weighing method of quantitative packaging scale
JP2016102716A (en) * 2014-11-28 2016-06-02 大和製衡株式会社 Weighting device and article conveyance system

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102615710B (en) * 2012-03-28 2013-12-04 中联重科股份有限公司 Proportioning material metering method, proportioning weight controller, system and concrete mixing plant

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006170846A (en) * 2004-12-16 2006-06-29 Yamato Scale Co Ltd Rotating type weighting device
JP2008296201A (en) * 2007-06-04 2008-12-11 Anritsu Sanki System Co Ltd Weight sorting apparatus
WO2011044609A1 (en) * 2009-10-12 2011-04-21 Leonard Ian Burrell Belt image zero tracking system
US8352214B2 (en) 2009-10-12 2013-01-08 Leonard Ian Burrell Belt image zero tracking system
CN102700739A (en) * 2012-06-20 2012-10-03 上海大和衡器有限公司 Quantitative packaging scale for sub-package according to amount of grains and quantitative weighing method of quantitative packaging scale
JP2016102716A (en) * 2014-11-28 2016-06-02 大和製衡株式会社 Weighting device and article conveyance system

Also Published As

Publication number Publication date
JP3427946B2 (en) 2003-07-22

Similar Documents

Publication Publication Date Title
US7214954B2 (en) Method for operating optical sensors
EP0171237B2 (en) Creep-compensated weighing apparatus
US7407323B2 (en) Methods and systems for determining temperature of an object
EP0040539B1 (en) Method of measuring by weighing
US4858145A (en) Force measuring device
CN1021369C (en) Method for correcting sensitivity and linearity of weigher of electronic balance and electronic balance
EP2246675B1 (en) Weighing method and apparatus
EP0403200B1 (en) Method for scale calibration and weighing
US5990422A (en) Apparatus for measuring weight and length of article
US4709770A (en) Device for correcting dynamic weight signal from conveyer weigher
US5832417A (en) Apparatus and method for an automatic self-calibrating scale
US5321634A (en) Automatic thin-film measuring apparatus
US20040015302A1 (en) Electromagnetic flowmeter
US4660667A (en) Multi-range load cell scales
US8063320B2 (en) Counting scale with a moving average start calculating section, for determining whether or not to calculate a moving average
EP1736745B1 (en) Method for adaptively correcting drift conditions in a force measuring device and force measuring device for carrying out the method.
US20050023047A1 (en) Weight measuring device
JP3088751B2 (en) Combination weighing device with tare weight reduction
GB2221039A (en) Load cell device with integral data processing/output means
US6636820B2 (en) Method and apparatus for measuring weight using uncalibrated load cells
US20080141755A1 (en) Weight scale fault detection
EP0124356B1 (en) Span adjusting method and apparatus
DE3340512C2 (en)
US4796212A (en) Load cell type, weight-measuring device
DE112010003522B4 (en) Device and method for controlling a weighing device for electronic belt scales

Legal Events

Date Code Title Description
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20030422

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20120516

Year of fee payment: 9

LAPS Cancellation because of no payment of annual fees