JPH0321826A - Weighing device - Google Patents

Abstract

PURPOSE:To provide a low-pass filter and eliminate the need to remove the influence of a shock and to prevent response delay from being generated by providing a weighing means, a digital variable band removal filter, a 1st and a 2nd constant setting means, and a control means. CONSTITUTION:The output signal of a load cell 12 as the weighing means is amplified 24, digitized by an A/D converter 26, and supplied to a CPU 30 through an interface 28. Similarly, the output of a sensor 20 is supplied to the CPU 30 through the interface 28. The CPU 30 is further supplied with a signal indicating reference weight generated by operating a keyboard 32, i.e. the standard weight of each article supplied to a weighing part through an interface 28. The CPU 30 processes the weighing signal after the A/D conversion according to a control program stored on a memory 34 by using the output signal of the sensor 20, the signal indicating an initial load, etc., and functions as the digital variable band removal filter, 1st and 2nd constant setting means, and control means.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a measuring device, and particularly to one equipped with a digital filter for removing noise contained in a measuring signal.

[Prior Art] Conventionally, as a weighing device designed to reduce noise contained in a weighing signal, there is, for example, one shown in FIG. 13 (Japanese Unexamined Patent Publication No. 62-238423). In the figure, ■ is a weight detector, 2 is an amplifier, 3 is a switched capacitor filter that functions as a notch filter, 4 is a low-pass filter that attenuates noise caused by external vibration, and 5 is a low-pass filter 4. connected to the output side,
A level shift circuit that sets the initial load output of the weight detector l in an unloaded state to almost zero level, 6 is an A/D converter, 7 is a microcomputer, 8 and 9 are cuts of the switched capacitor filter 3 This is a V/F converter for changing the off frequency according to the output of the level shift circuit 5.

In such a weighing device, in an unloaded state, the V/F converters 8 and 9 change the cutoff frequency of the switched capacitor filter 3 according to the output of the level shift circuit 5 to the weight detector 1 in the unloaded state. It is set to almost match the natural vibration frequency of the object to be measured or the weight detector l.
, the V/F converters 8 and 9 convert the switched capacitor 3 according to the output of the level shift circuit 5 at that time.
The purpose is to change the cutoff frequency of the weight detector l and remove the natural vibration frequency of the weight detector l at that time.

[Problems to be Solved by the Invention] This weighing device can determine the outline of the weight value by knowing the output of the level shift circuit 5, so the natural vibration frequency of the weight detector 1 can also be determined, and this frequency can be determined by the switched capacitor filter. However, a problem arises when this weighing device is used, for example, in a weight sorting machine in which objects to be measured are conveyed by a conveyor.

That is, in such a case, the weight detector 1 is installed on the weighing conveyor, but the output of the weight detector 1 becomes extremely low due to the shock generated when the object to be measured that has been conveyed by the conveyor gets onto the weighing conveyor. This becomes a large value, which also occurs in the output of the level shift circuit 5, which controls the cutoff frequency of the switched capacitor filter 3, and reduces the natural vibration actually occurring in the output of the weight detector 1. The cutoff frequency of the switched capacitor filter 3 becomes a value far different from the frequency, and the natural vibration of the weight detector l cannot be effectively removed. this is,
Is it possible to solve the problem by increasing the time constant of the low-pass filter 4 to remove large values caused by impact before supplying the output of the weight detector 1 to the level shift circuit 5? Another problem arises in that increasing the time constant of the bus filter 4 causes a response delay.

The present invention provides that, in a weight sorter, etc., (1) the weight value of the article to be weighed falls within a certain range from a certain reference weight, and the weight value of the article to be weighed falls within a certain range from a certain reference weight, and accurate Weighing is necessary, but when weighing an item whose weight is outside of this range, it is only necessary to know that it is outside this range, and there is no need for precise weighing. (2)
In addition, even when weighing items with weights within or near the above range, there is no need for precise measurement until a weighing signal is generated within or near the range; (3) In order to perform precise measurements, It is necessary to remove the natural vibration even when the object to be measured is not loaded, and also when the object to be measured is loaded, but it is necessary to remove the natural vibration even when the object to be measured is loaded. Taking advantage of the fact that it can be easily detected,
It is an object of the present invention to provide a measuring device that solves the above problems.

[Means and effects for solving the problem] In order to achieve the above object, the measuring device according to the present invention mainly includes a measuring means in which the weight of the article to be weighed is close to a predetermined reference weight, and A measuring signal is supplied, and a digital variable band rejection filter exhibiting a band rejection characteristic according to the input characteristic constant, and a measurement characteristic constant determined based on the initial load of the weighing means and the above reference weight are set. The first
a second constant setting means that is set with a characteristic constant for zero point measurement determined based on the initial load, and a second constant setting means that indicates that the article is completely loaded on the weighing means or that the article will be loaded soon. When this is detected, the weighing signal is filtered by a filter that inputs the measurement characteristic constant from the first constant setting means, and when it is detected that no article is loaded on the weighing means, zero point measurement is performed from the second constant setting means. and a control means for filtering the measurement signal with a filter into which a characteristic constant for the measurement is input.

In the weighing device configured as described above, when the control means detects that no article is loaded, the characteristic constant for zero point measurement is supplied to the filter from the second constant setting means. Therefore, from the filter, a weighing signal from which the natural vibration of the article in the unloaded state is removed can be obtained. Further, when the control means detects that the article has been loaded or will be loaded, that is, when the article is loaded onto a weighing machine etc., the first constant setting means sets characteristic constants for measurement (reference weight and initial weight). (a characteristic constant determined based on the load) is supplied to the filter. At this time, if an article of standard weight is loaded at the measurement timing, the natural vibration in this state can be completely eliminated. At this time, if the value of the loaded article is within a certain range from the reference weight, the natural vibration can be removed with sufficient accuracy for practical use, since the filter is a digital band elimination filter. Furthermore, if an article whose weight is outside a certain range or vicinity from the reference weight is loaded, it is only necessary to know that it is outside the range, so it is not necessary to remove the natural vibration with high precision. Furthermore, by subtracting the output of the digital filter in the unloaded state from the output of the digital filter in the loaded state of an object whose weight is within a certain range or in the vicinity of the reference weight, the weight of the loaded object can be calculated with high accuracy. can be measured.

In the above invention, the characteristic constant for measurement is always constant;
If the weight of the item being actually weighed becomes uneven within a certain range, for example, the average of the filtered weighing signals of multiple items output from the filter can be used to achieve more accurate measurement. It is also possible to provide means for calculating the value and means for correcting the measurement constant based on the average value and the initial load. In this way, the characteristic constant for measurement is corrected in accordance with the weight of the article being actually weighed, so that it is possible to perform weighing with higher precision.

Furthermore, a means for correcting a measurement characteristic constant based on the reversely waved weighing signal of one article output from the filter and the initial load, and a means for supplying the corrected measurement characteristic constant to the filter, and means for refiltering the unfiltered metering signal of said one article. In this way, the measurement characteristic constant is corrected by the weight of the article once weighed, and the corrected constant is filtered again by the supplied digital filter, resulting in extremely high precision weighing.

Furthermore, the present invention can also be applied to cases where it is only necessary to precisely measure the switching weight for switching the filling amount, such as in a fixed-quantity filling machine. In this case, the weighing means changes the weight of the loaded article over time, and the first constant setting means may have at least one measurement characteristic constant. The at least one measuring characteristic constant is determined by the at least one switching weight and the initial load.

In this case, of course, if there is a plurality of switching weights, the first constant setting means has a plurality of measurement characteristic constants each determined based on a plurality of reference weights having different values and the above-mentioned initial load. Alternatively, the control means may supply a measurement characteristic constant to the filter according to the measurement signal.

[Example] A first example is shown in FIGS. 1 to 4. In this embodiment, the present invention is implemented in an automatic weight sorting machine as shown in FIG. 2, where IO is a weighing section, and a weighing means such as a load cell 12 is connected to the lower part of the weighing section. 14 is an incoming conveyor, 16 is an outgoing conveyor, and l8 is a chain that constitutes these conveyors. Reference numeral 20 denotes a sensor that detects that the article 22, which is the object to be measured, has arrived just in front of the measuring section IO.

The output signal of the load cell I2 is amplified by an amplifier 24 as shown in FIG.
30. Similarly, the output signal of the sensor 20 is also supplied to the CPU 30 via the interface 28.
A signal representing the standard weight of each article that is sequentially supplied to the machine 0 is also supplied via the interface 28. C.P.
The U30 processes the digitally converted weighing signal based on the output signal of the sensor 20, the signal representing the initial load, etc. according to the control program stored in the memory 34, which includes, for example, RAM and ROM. The digital variable band-rejection filter functions as the first and second constant setting means and the control means, and furthermore, although it is not directly related to the present invention, whether the natural vibration or the weighing signal removed by the present invention is within an allowable range. It also functions as a means of selecting whether or not to use it. Note that 36 is a display unit that displays the value of the weighing signal from which the natural vibration has been removed and the selection results.

FIG. 1 shows a routine related to the present invention in the control program of the CPU 30, and this routine mainly functions as a digital band-removal filter. Special public school 1986-5z
As disclosed in Publication No. 6a, s, the natural vibration frequency can be determined by sampling the measurement signal containing the natural vibration frequency for at least one period of the natural vibration and finding the average value of these sampling values. It removes prestige. This natural vibration frequency component is removed when T1 time elapses after the article 22 passes the sensor 2' position, as shown in FIG.
When the article 22 passes the position of the sensor 20 and T2 time has elapsed, that is, the article 2 passes over the measuring section 10.
This is done after 2 has passed. In the former case, the measured value is obtained by removing the natural vibration frequency frN in the state where the article of reference weight WT is loaded, and in the latter case, the natural vibration frequency frequency fN in the state where the article is not loaded is obtained. The metric value of the zero point from which is removed is obtained. Therefore, by first operating the keyboard 32, the measuring section 1
The reference weight WT of each article sequentially sent to o is set (step S2). And this reference weight WT,
Based on the initial load stored in advance in the memory 34, for example, the weight Wl of the weighing section 10, the measurement natural vibration frequency fTN is determined, which is the natural vibration frequency when an article of reference weight WT is loaded on the weighing gflIO. is calculated (step S4). This calculation is done by. however,
g is the gravitational acceleration, and K is the spring constant of the load cell 12. From this measurement natural vibration frequency fTH and the sampling period T (m seconds) of the A/D converter 26, the natural vibration frequency distribution fTN when the article of reference weight WT is loaded
Calculate the number of samplings Nn corresponding to a period that is an integral multiple of the natural vibration frequency component fTH necessary to remove
It is stored in the memory 34 (step S6). This calculation can be done by This N1 corresponds to the measurement characteristic constant in the claims.

Next, in the case of the initial load Wl, that is, the natural vibration frequency f for zero point measurement, which is the natural vibration frequency when the article is not loaded. is calculated in the same manner as step S4 (step S8
), and the natural vibration frequency f for zero point measurement. and the sampling period T of the A/D converter 25, the natural vibration frequency f. The sampling number N corresponds to a cycle that is an integral multiple of the weight. is calculated in the same manner as in step S6 and stored in the memory 34 (step 310). corresponds to the characteristic constant for zero point measurement in the claims. Steps S2 to S10 up to this point correspond to so-called initial settings. Although not shown in the flowchart, upper and lower allowable weight limits are also set for sorting each article into good and bad items.

Next, it is determined whether the sensor 20 is off, that is, whether the article 22 has reached the installation position of the sensor 20 shown in FIG. 2 (step Sll). If the answer is YES, a zero measurement timer (not shown) is reset and activated, and the measurement timer is also activated (step S12).F1 flag is then set to 1, and the measurement timer is activated. Then, it is determined whether the measurement timer has timed up (step S14). If the answer is NO, the process returns to step Sll. Step Sll
The answer is. This time, since the article is already on the weighing section 101, the answer is No, and it is determined whether the Fl flag is 1 (step S15). Since the answer is YES, step S
14. Thereafter, until the answer to step Sl4 becomes YES, that is, the measurement timer starts counting, and T.
As time elapses, until the article 22 reaches the weighing section 10 as indicated by 22a in FIG. 2, step SI4,
Repeat loops 11 and 15.

If the answer to step S14 is YES, the Fl flag is set to O.
(step S16), and then the A/D converter 26
Nn digital measurement signals obtained by digitally converting the output of the amplifier 24 every cycle T are input, and their average value wn is determined (step S17). In addition, No
The data may be obtained by digitally converting the output of the amplifier 24 from the time when step S12 is started. As a result, a weighed value of the article with the natural vibration frequency component fTn removed when the article of reference weight WT is loaded is obtained. In FIG. 4, the period T is 6.71 seconds, the natural vibration frequency fTn is 18.6Hz when an article with a weight equal to the reference weight WT is loaded, and the number of samples sampled at that time is 1000/(18. 6X 6.7) This shows the amount of attenuation for the frequency when #8 is used.As is clear from this, the amount of attenuation of -40dB or more is 18.6Hz.
It is obtained within a range of ±0.582 around . In other words, even if the weight of the incoming article fluctuates from the reference weight WT, the fluctuation of the rigid vibration frequency will be ±0.5.
}1, if it is within the range, approximately 40 dB of attenuation can be obtained. Since the weights of the articles sequentially fed into the weight sorter vary significantly from the standard weight WT, this is sufficient for practical use. Even if products with a large variation in weight from the standard weight WT are sent and the natural vibration frequency component cannot be removed sufficiently, these will definitely be determined as defective products during the sorting process described later. Therefore, there is no problem even if the measured value cannot be measured precisely.

The previously obtained weighing signal Wo at the zero point is subtracted from the weighing signal Wn from which the natural vibration frequency a component has been removed in this way, and the net weight of the article is calculated (step 318). It is determined whether this net weight is within the allowable weight range (step S19). After that, the measurement timer is reset (step S20), and then the process returns to step Sll. Here, if the sensor 2o detects a new article, the answer to step Sll becomes YES, the zero measurement timer that has been continuing measurement is reset, and new measurement is performed from 0.
Thereafter, the operation is as described above. On the other hand, step Sll
If the answer is No, step s15 is executed, but at this time, the weight measurement has been completed, so the Fl flag is set to O. Therefore, the answer to step S15 is No, and it is determined whether the zero measurement timer is up (step S21).

If the answer is No, the article 22 is shown at 2 in FIG.
The loop of steps S21 and 11.15 is repeated until the position shown by 2b is reached and there is no article 22 on the weighing section 10. If the answer to step S21 is YES, A
A digital measurement signal N obtained by digitally converting the output of the amplifier 24 every cycle T by the /D converter 26. The average value Wo is calculated and stored in the memory 34 (step S22). As a result, the natural vibration force f when the article is not loaded. The metric value of the zero point from which is removed is obtained.

Note that in the above embodiment, the detection of the article 22 is performed by the weighing section l.
This was done just before the article 22 was placed on the weighing section IO, but the state in which the article was completely placed on the weighing section IO was detected, and then the average value of the N0 digital weight values was calculated. good.

A second embodiment is shown in FIGS. 5 and 6. In this embodiment, the number of samples for zero point measurement is N. And, the sampling number Nn for measurement is fixed, and instead, the sampling period T is changed according to the natural vibration frequency fTn for measurement and the natural vibration frequency f■ for zero point measurement. Therefore, as shown in FIG. 5, a programmable timer 38 controlled by the CPU 30 is provided, which allows the A/D
It is configured to give an A/D conversion command to the converter 26. As shown in FIG. 6, first, the reference weight WT. Number of samples for zero point measurement N. and the measurement sampling number N7 is set (step S24). In addition, No. N
n may be stored in the memory 34 in advance, or both may be the same number. Next, in the same manner as in the first embodiment, the fixed vibration frequency f■ for zero point measurement and the natural vibration frequency fTn for measurement are calculated (step S26), and the sampling period Tn for measurement and the sampling period T for zero point measurement are calculated. ．． is calculated (step 328). For example, in the case of Tn, this calculation is ? Therefore, it is done. T. The same applies to the case of .

Thereafter, processing is performed in the same manner as in the first embodiment, but when executing step S17, the CPU 30 sets the cycle T■ in the programmable timer 38. The programmable timer 38 generates a conversion command signal to the A/D converter 26 every time period T, and at the same time generates a conversion command signal to the CPU 30.
A signal indicating that /D conversion has been started is supplied. Similarly, in the case of step S22, the CPU 30 sets the period Tn to the programmable timer 38, and the programmable timer 38 controls the A/D converter 26 every time the period Tn elapses.
A conversion command signal is supplied to the CPU 30, and at the same time a signal indicating that A/D conversion has started is supplied to the CPU 30. By changing the sampling period in this way, it is easier to obtain sampling values for an integer period of the natural vibration frequency. In other words, if you fix the sampling period and try to obtain sampling values for an integer period of the natural vibration frequency by changing the number of samples, the condition is that the natural vibration frequency is an integer multiple of the sampling period. It is not always possible to satisfy this requirement.For example, in the example shown in Figure 4, the number of samples is 1000/(18.6 x 6.7
) = 8.02..., which is not 8. Therefore,
When the number of samples is 8, the removed frequency weight is 18.6H. There is a slight deviation from However,
If the number of samples is fixed at 8, the sampling period is set to 6.72 double seconds to obtain 18.61-1
．． The natural vibration of can be removed. Furthermore, setting the sampling period to 6.721 seconds can be easily achieved by controlling the programmable timer 38.

A third embodiment is shown in FIG. This embodiment differs from the first embodiment in the method of determining when to perform zero point measurement. That is, in the first embodiment, zero point measurement was performed when the sensor 20 did not detect an article for a predetermined period of time, but in this embodiment, when the sensor 20 did not detect an article,
A predetermined number of consecutive weighing signal values below a predetermined threshold value Th, for example, the sampling number N for zero point measurement.
Zero point measurement is performed when only o is reached.

That is, when the answer to step S15 is No, it is determined whether the digital weighing signal is below the threshold Th (step S30), and if the answer is YES, the threshold Th
The value n of the software counter that counts the number of the following weighing signals is increased by one (step S32), N
．． Input the weighing signal to the shift register configured in stages (
Step S34), and the value n of the software counter is N. (step S36), and if the answer is No, the process returns to step S30. Hereinafter, steps S30 to S36 are repeated, but if the weighing signal becomes larger than the threshold value Th during the repeating because an article is loaded on the weighing section 10, etc., the answer to step S30 becomes No. Therefore, the value n of the software counter is set to 0, and the shift register is cleared (step 538), and step S
Return to ll. Therefore, the answer to step S36 becomes YES only when No measurement signals are continuously below the threshold value Th. At this time, each stage of the shift register stores No weighing signals below the threshold Th, so these are added and divided by No to obtain the average value, and the natural vibration Obtain the zero point weight from which frequency components have been removed (step S40), Figures 8 and 9
The figure shows a fourth embodiment. In this embodiment, a digital band rejection filter is installed as the first filter so that even if the weighed article is somewhat away from the reference weight WT and the natural vibration frequency is far from fTn, the natural vibration frequency can be sufficiently removed.
Rather than simply determining the average value as in the third embodiment, multiple moving average values are determined, for example, as disclosed in Japanese Patent Application Laid-open No. 62-280625. That is, if the sampling number Nn determined in the same manner as in the first or second embodiment is 4, first, as shown in FIG. The first moving average value Wl. W2, W3, W
4, ------- Find W8 (step S42),
Then, these moving average values Wl, W2, W3, W4, .
...The second moving average value Wl' for W8 is Nn, 4 each. W2', W3'. W4'
, W5' are determined (step S44), and a third moving average value W' is determined using five (N.+1) moving average values (step S46). In this example, when calculating the second moving average value, the moving average value was calculated for Nn pieces at a time, but it is also possible to calculate the moving average value for Nn pieces at a time when calculating the first moving average value. ．． In that case, the second moving average is taken by each Nn+1 moving average,
For the third moving average, just take the moving average of each N1+2 number. If you want to take the third moving average of N,, the first time you should take the moving average of Nn-2 pieces at a time, and the second time you take the moving average of N-1 pieces each. In this example, a triple moving average was taken, but a double moving average, or a quadruple or quintuple moving average may be taken. Also, in zero point measurement, multiple moving averages may be taken in the same way as above. A fifth embodiment is shown in FIG. In each of the above embodiments, the sampling number Nn and the sampling period fTrl are determined at the time of initial setting and are continued thereafter, but in this embodiment, N. , or fTn. That is, first perform initial settings (step S
48). This corresponds to steps S2 to S10 in FIG. Following this, it is determined whether the article is to be measured (step S50). For example, step S11 in FIG.
, 12, 13, l4, 15, l6, and 21. If it is not an item measurement, zero point measurement is performed (
Step S52). This is, for example, step S2 in FIG.
This corresponds to 2.

In the case of measuring an article, the weight of the article is measured (step S
54). This is, for example, step S17 in the first or second embodiment or steps S42 to S4 in the fourth embodiment.
This corresponds to 6. Thereafter, in the same manner as in the first embodiment, the net weight is calculated (step S18), a judgment is made (step S19), and the weight of the article judged to be good is memorized. When a predetermined m number of samples have been collected, their average value is calculated, fTn is recalculated using this average value and the initial load, and the sampling number Nn or the sampling period T1 is modified based on this (step S56). Return to step S50.

Figure 11 shows the sixth embodiment. This embodiment is applied when the weight of the article is slightly widely distributed around the reference weight WT. First, the natural vibration frequency component is removed from the weighing signal as in the first or fourth embodiment. However, as mentioned above, if the weight of the article is slightly widely distributed from the reference weight WT, there is a possibility that the natural vibration frequency distribution has not been sufficiently removed. The number of samplings necessary to remove the natural vibration frequency and component is recalculated based on the initial load and the weighing signal from which the oscillation frequency has been removed, and the natural vibration frequency component is removed again using the recalculated number of samplings. It is. That is, first, initial settings are performed as in the sixth embodiment (step S48), and then, as in the sixth embodiment, it is determined whether the article is to be measured or not. Step SSO), if the answer to this judgment is YES, input the A/D conversion data sequentially,
It is stored (step S60). If the answer to step S50 is No, zero point measurement is performed in step S52. However, the number input in step S60 is set to be larger than the sampling number determined in the initial settings.
This is done in consideration of the possibility that the recalculated number of samples will be larger than the number of samples determined in the initial settings. Then, as in the first or fourth embodiment, the metric value W is obtained by removing the natural vibration frequency component. is obtained (step S62).

Note that steps S60 and S62 may be performed in parallel. Next, based on this Wn and the initial load W1, the natural vibration frequency fa. is recalculated (step S64), and this f
The sampling number N is based on the sampling period T
,,1 is recalculated (step s66), and this sampling number i? The measurement signal previously stored in step S60 is processed based on I[N■, and the natural vibration frequency component is removed to obtain a measurement value Wn. Calculate? Step 368), and thereafter operates in the same manner as in the first embodiment. According to this embodiment, since the calculation is performed twice to remove the natural vibration frequency component, the measured value can be obtained with high accuracy. However, since the calculation is performed twice, the calculation time becomes longer, but since each A/D conversion data is stored first, the second calculation is performed while the article is on the weighing section 10. There is no need to slow down the conveyor, as there is no need to do so, and there is no need to miss the optimal measurement timing.

FIG. 12 shows a seventh embodiment. In this embodiment, the present invention is implemented in a quantitative filling machine, and includes a switching weight W■ for switching from large to medium charging, a switching weight W2 for switching from medium to small charging, and a switching weight W3 for switching from small charging to stop charging. The aim is to detect with high precision whether or not the weighing signal has reached each point. First, each switching weight W. ,W
2, W, is set (step S70), and then the natural vibration frequency f? at the zero point is determined based on the initial load stored in advance in the memory 34. . Calculate the initial load of Sarajiko? Based on the switching weight W1, the fixed vibration frequency f7■ is calculated when the article is weighed by the switching weight W1, and similarly, the natural vibration frequencies fA2 and fT3 when the switching weight W2 and the switching weight W3 are the same. (Step S
72). And these natural vibration frequencies fa. , f■,
f? 2. Number of samples N based on fT3 and A/D conversion period T. ,N. , N2, and N3 are calculated (step S74). Then, it is determined whether to start feeding (step S76), and the answer is N.
N if o. Input A/D data and average value WA
vo is calculated and stored (step S78). If the answer to step S76 is YES, that is, the input is started, the large input is started (step 880), and the shift register configured in N stages is Input A/D data into
The average value WAvl of the stored values of each stage is calculated (step S82), and then WAvo is subtracted from this average value WA to calculate the filling weight W1 at that time (step S8
4). This filling weight W. It is determined whether the switching weight is greater than or equal to the switching weight W (step S86). If the answer is No, step S
Return to 80. Therefore, the average value WAv, is calculated every time A/D data is input. The answer to step 386 is Y
If it is ES, it is switched to medium loading (step S88).
). In addition, the filling weight W. is far apart from the switching weight W1, the natural vibration frequency component cannot be removed sufficiently, but the filling weight W. The natural vibration frequency component begins to be removed as it approaches the switching weight W, and is completely removed when it becomes equal to the switching weight WI. When the medium injection starts, the A/N2 stage shift register is
D data is input, and the average value WAv■ of the values stored in each stage of this shift register is calculated (step S90). W AVO is subtracted from the average value W AV2 to obtain the net weight W
M is calculated (step S92), and it is determined whether the net weight WM is greater than or equal to the switching weight W2 (step S94). If the answer is No, the process returns to step 388, and if the answer is YES, the mode is switched to small input (step S96). When switched to small input, the A/D data is input to the N3-stage shift register. Average value W of memory values for each stage
AVI is calculated (step 398). This average value W
W AVO is subtracted from AV:l to calculate the net weight WL (step S100). Then, it is determined whether this net weight WL is greater than or equal to the switching weight W3 (step S102). If the answer is No, the process returns to step S96, and if the answer is YES, the loading is stopped (step S1
04), In this example, three stages of charging were performed: large charging, medium charging, and small charging, but the present invention can also be implemented in two stages of large charging and small charging, and one stage charging of only large charging. be able to.

In addition, in this embodiment, the present invention was implemented in the case where the weighing signal increases as the article is filled, but it can also be applied in the case of so-called residual weighing, in which the weighing signal decreases as the article is filled. The invention can be practiced.

[Effects of the Invention] As described above, according to the invention described in claim 1, the characteristic constant for zero point measurement that gives the digital band-rejection filter a characteristic capable of removing the natural vibration frequency when the article is not loaded, and the reference weight. A measurement characteristic constant that gives the digital band rejection filter a characteristic that eliminates the natural vibration frequency when an article is loaded on the weighing means is prepared, and the influence of impact when the article is completely loaded on the weighing means and when boarding is prepared. Since the weighing signal is filtered by the digital band rejection filter into which the measurement characteristic constant is input when the shock is not being applied, there is no need to install a low-pass filter to remove the impact, which causes a delay in response. Never.

According to the invention recited in claim 2, similarly to the invention recited in claim 1, the weight of the plurality of articles is calculated by removing the natural vibration frequency contribution in a state in which the articles having a weight approximately equal to the reference weight are loaded. Since the measurement characteristic constants are corrected based on
It is possible to measure with high precision. According to the invention set forth in claim 3, the measurement characteristic constant is corrected based on the weight value of the article from which the natural vibration frequency component has been removed in the same way as the invention set forth in claim 1, and the measurement characteristic constant is corrected based on the weight value of the same article again. Since the natural vibration frequency component is removed, highly accurate measurement is possible.

According to the inventions set forth in claims 4 and 5, high-accuracy metering without response delay can be performed in a quantitative filling machine as well, as in the invention set forth in claim 1.

[Brief explanation of the drawing]

1 is a block diagram of a first embodiment of the present invention, FIG. 2 is a schematic diagram of the first embodiment, FIG. 3 is a block diagram of the first embodiment, and FIG. 4 is a block diagram of the first embodiment of the present invention. 5 is a frequency characteristic diagram of the digital band-rejection filter used in the first embodiment.
The figure is a block diagram of the pulling part of the second embodiment, FIG. 6 is a flowchart of a part of the second embodiment, FIG. 7 is a flowchart of a part of the third embodiment, and FIG. is a partial flowchart of the fourth embodiment, FIG. 9 is a diagram showing the basic principle of the digital band-rejection filter of the fourth embodiment, FIG. 10 is a flowchart of the fifth embodiment, and FIG. 12 is a flowchart of a part of the sixth embodiment, FIG. 12 is a flowchart of the seventh embodiment, and FIG. 13 is a block diagram of an example of a conventional weighing device. l2... Roth cell (measuring means) 30...● CPU (constant setting means, digital band rejection filter, control means)

Claims (5)

[Claims] (1) A weighing means whose weight of the article to be weighed is close to a predetermined reference weight, and a digital variable band-elimination filter that is supplied with a weighing signal from this weighing means and exhibits band-elimination characteristics according to input characteristic constants. and a first constant setting means in which a characteristic constant for measurement determined based on the initial load of the measuring means and the reference weight is set, and a characteristic constant for zero point measurement determined based on the initial load is set. inputting the measurement characteristic constant from the first constant setting means when it is detected that the article is completely loaded on the measuring means or when the article is being loaded. The weighing signal is filtered by the filter, and when it is detected that no article is loaded on the weighing means, the weighing signal is filtered by the filter into which the zero point measurement characteristic constant from the second constant setting means is input. A measuring device comprising a control means. (2) The weighing device according to claim 1, further comprising means for calculating an average value of filtered weighing signals of a plurality of articles output from the filter, and a means for calculating the average value of the filtered weighing signals of the plurality of articles output from the filter, A measuring device comprising means for correcting characteristic constants. (3) The weighing device according to claim 1, further comprising means for correcting the measurement characteristic constant based on the filtered weighing signal of one article outputted from the filter and the initial load; and means for supplying said measuring characteristic constant to said filter and causing said filter to refilter the unfiltered weighing signal of said one article. (4) a measuring means in which the weight of the loaded article changes with time; a digital variable band-rejection filter that is supplied with a measuring signal from the measuring means and exhibits a band-rejection characteristic according to an input characteristic constant; a first constant setting means having a measurement characteristic constant determined based on the initial load of the measuring means and at least one predetermined reference weight; and a zero point measurement characteristic determined based on the initial load. a second constant setting means in which a constant is set, and the measurement characteristic constant from the first constant setting means when it is detected that an article is loaded on the measuring means or when it is about to be loaded. When it is detected that no article is loaded on the weighing means, the filter inputting the characteristic constant for zero point measurement from the second constant setting means filters the weighing signal. and control means for filtering. (5) The measuring device according to claim 4, wherein the first constant setting means has a plurality of measurement characteristic constants each determined based on a plurality of reference weights having different values and the initial load, and A weighing device, wherein the control means supplies the filter with a measurement characteristic constant according to the weighing signal.