JP4558334B2 - Weight measuring apparatus and weight measuring method - Google Patents

Description

  The present invention relates to a weight measuring device and a weight measuring method, and more particularly to a weight measuring device and a weight measuring method for measuring the weight of an object to be measured based on a weighing signal output from a weighing means such as a load cell.

  In this type of weight measuring device and weight measuring method, the natural vibration of the weighing means itself becomes noise and greatly affects the measurement accuracy. The frequency of the natural vibration varies depending on the weight applied to the measuring means. However, as a technique for reducing the influence of such natural vibration, there has been conventionally disclosed in Patent Document 1, for example. According to this prior art, an electric signal (measurement signal) output from a measuring unit as a measuring unit is integrated over a certain integration time width by an integrator. Then, the integration output of the integrator is divided by the integration time width by a divider, and weight data of the object to be measured is generated based on the division output of the divider. Further, based on this weight data, a time width of one period or an integral multiple of the natural vibration period of the measuring unit is calculated, and the calculated time width is equal to the integration time width by the integrator. The integration time width is reset and updated. Therefore, after the integration time width is updated, the measurement signal is averaged over the updated integration time width, that is, over a time width corresponding to one period or an integral multiple of the natural vibration period of the measurement unit. . As a result, the natural vibration component included in the measurement signal is removed, and accurate weight data is obtained.

Japanese Patent Laid-Open No. 62-261021

  However, in the above-described prior art, after the integration time width is updated, the measurement signal is integrated again over the updated integration time width. Then, after the integration process ends, the integration time width is updated again. That is, it takes at least the integration time width, that is, one cycle of the natural vibration period of the measuring unit or an integral multiple of the time to update the integration time width. As described above, in the conventional technique, the update speed of the integration time width, that is, the response (transient response) depends on the natural vibration period, and therefore, an application requiring higher response than this, for example, a weight-type filling device, It cannot be used sufficiently for applications as a so-called dynamic balance that needs to measure the weight of an object to be weighed from a weighing signal that is constantly changing (relatively fluctuating) like a weighing conveyor. There is a problem.

  Accordingly, an object of the present invention is to provide a weight measuring apparatus and a weight measuring method with high responsiveness that can sufficiently cope with a dynamic balance.

The first invention is an invention relating to a weight measuring apparatus, wherein sampling means for sampling a weighing signal obtained from the weighing means at a period shorter than the natural vibration period of the weighing means, and discrete weighing after sampling by the sampling means. Filter means for applying a filtering process to the signal according to a certain filter constant, and an updating means for updating the filter constant based on the processed weighing signal processed by the filter means. In such a configuration, the updating unit updates the filter constant so that the frequency characteristic of the filter unit becomes a characteristic of attenuating a component in a frequency band including the natural vibration frequency of the weighing unit in the discrete measurement signal . Further, the updating means updates the filter constant every time sampling is performed by the sampling means.

That is, in the first invention, the weighing signal obtained from the weighing means is sampled by the sampling means at a period shorter than the natural vibration period of the weighing means. The discrete metric signal sampled by the sampling means is subjected to a filtering process according to a certain filter constant by the filtering means. Further, the updating means updates the filter constant based on the processed weighing signal after processing by the filter means. Specifically, the update unit updates the filter constant so that the frequency characteristic of the filter unit becomes a characteristic of attenuating a frequency band component including the natural vibration frequency of the measurement unit in the discrete measurement signal . By updating the filter constant in this manner, the natural vibration component is effectively removed from the discrete measurement signal, in other words, a processed measurement signal from which the natural vibration component is well removed is obtained. In addition, the filter constant is updated every time sampling is performed by the sampling means, that is, at a time interval shorter than the natural vibration period.

  It is desirable that the sampling period by the sampling means is shorter than 1/2 of the natural vibration period of the weighing means. This is to satisfy the sampling theorem in order to accurately capture the natural vibration period.

  Moreover, as a filter means, there exists a moving average filter, for example. In this case, the number of discrete metric signals to be subjected to moving average processing by the moving average filter can be used as a filter constant.

  Further, the updating means is a known vibration frequency different from the first weight when the known first weight is applied to the weighing object placing portion of the weighing means, for example, the placing table. The filter constant may be updated based on an arithmetic expression including the natural vibration frequency of the weighing means when the second weight is applied. In other words, the natural vibration frequency of the weighing means is related to a so-called initial load applied to the weighing means such as a mounting table from the beginning. The natural vibration frequency is also related to the spring constant of the measuring means itself. Therefore, in order to update the filter constant so that the frequency characteristic of the filter means becomes the above-described characteristic, it is necessary to accurately obtain these initial load and spring constant. Therefore, the updating means uses the initial load based on the natural vibration frequency actually measured when the first weight is applied to the mounting table and the natural vibration frequency actually measured when the second weight is applied. And determine the spring constant. More specifically, the initial load and the spring constant are obtained based on a binary simultaneous equation representing the relationship between each natural vibration frequency and the initial load and the spring constant which are unknown here. An arithmetic expression including the initial load and the spring constant, in other words, an arithmetic operation indirectly including the natural vibration frequency when the first weight is applied and the natural vibration frequency when the second weight is applied. Based on the expression, an appropriate filter constant is obtained and updated. Note that the first weight (or the second weight) may be zero. That is, the time when the first weight (or the second weight) is applied may be when nothing is placed on the object placing portion.

The second invention is an invention relating to a weight measuring method, a sampling process for sampling a weighing signal obtained from the weighing means at a cycle shorter than the natural vibration period of the weighing means, and a discrete weighing after sampling by this sampling process. A filtering process for applying a filtering process to the signal according to a certain filter constant; and an updating process for updating the filter constant based on the processed metric signal after the process by the filtering process. In such a configuration, in the updating process, the filter constant is updated so that the frequency characteristic in the filtering process becomes a characteristic of attenuating the frequency band component including the natural vibration frequency of the measuring means in the discrete metric signal . Further, in the updating process, the filter constant is updated every time sampling is performed in the sampling process.

  That is, the second invention is a method invention corresponding to the first invention, and thus has the same effect as the first invention.

  According to the present invention, since the filter constant is updated at a time interval shorter than the natural vibration period of the measuring means, the time for at least one natural vibration period or an integral multiple of the natural vibration period is required to update the integral time width. High responsiveness can be obtained compared to the above-described conventional technology in which the above is applied. Therefore, there is an effect that it is possible to sufficiently cope with the use as a dynamic balance such as a weight-type filling device and a weighing conveyor which could not be handled by the prior art.

  An embodiment of a weight type filling apparatus to which the present invention is applied will be described with reference to FIGS.

  As shown in FIG. 1, the weight-type filling device 10 according to this embodiment is a so-called quantitative method in which a predetermined container 12 such as a bottle is filled with a predetermined liquid, for example, a beverage, as an object to be weighed, at a constant weight. It is for filling and is provided with N (N: plural) units 14, 14,.

  Each unit 14 includes a mounting table 16 serving as an object mounting unit on which the container 12 is mounted, a weighing machine 18 serving as a weighing unit to which the mounting table 16 is coupled, and an upper side of the weighing machine 18. A valve 20 serving as a supply means for supplying a beverage to the container 12 is provided. In addition, a stand 22 for supporting the container 12 placed on the placement table 16 is provided in the vicinity of the placement table 16. In this embodiment, N = 36, that is, 36 units 14, 14,... Are provided. Each unit 14, 14,... Is assigned an individual identification number n of 1 to N.

  As shown in FIG. 2, each unit 14, 14,... Is disposed at a portion near the periphery of the disc-shaped turntable 24 at equal intervals along the periphery, that is, every 10 degrees (= 360 degrees / N). Have been placed. The turntable 24 rotates about the rotation shaft 26 in the direction indicated by the arrow 100 in FIG. 2 (clockwise in FIG. 2) together with the rotation shaft 26. As the turntable 24 rotates, the units 14, 14,. The above-mentioned identification numbers n (= 1 to N) are assigned to the units 14, 14,... So that the values sequentially increase in the counterclockwise direction in FIG.

  Returning to FIG. 1, the upper end portion of the rotating shaft 26 is located at a position higher than each unit 14, 14,... (Each valve 20, 20,...). A shaped storage tank 28 is fixed. This storage tank 28 is for temporarily storing beverages, and N pipes 30, 30,... Are radially connected to the lower part of the side wall thereof. Are individually coupled to the ends of the pipes 30, 30,... Therefore, the beverage in the storage tank 28 flows through the pipes 30, 30,... To the valves 20, 20,. And a drink is supplied to each container 12,12, ... via these valves 20,20, .... In addition, when the storage amount of the beverage in the storage tank 28 becomes a certain amount or less, the storage tank 28 is automatically replenished with the beverage from a replenishment tank (not shown).

  On the other hand, the lower end of the rotary shaft 26 is positioned below the turntable 24 in a state of passing through the center of the turntable 24. A motor 34 (strictly, the rotating shaft of the motor 34) is connected to the lower end portion of the rotating shaft 26 via a gear mechanism 32. The motor 34 is controlled by a controller 36 described later, and the rotating shaft 26 rotates through the gear mechanism 32 when the motor 34 is driven. The gear mechanism 32 generates one pulse every time the rotating shaft 26 makes one rotation. Specifically, every time a rotation reference (origin) provided at one position on the periphery of the rotating shaft 26 is detected. An encoder 38 for generating a single pulse, that is, a rotation reference detection is combined. In addition to the encoder 38, the gear mechanism 32 is also coupled with an encoder 40 for detecting rotation angle, which generates α × N (α: natural number) pulses each time the rotating shaft 26 makes one rotation. Has been. The output pulses of these encoders 38 and 40 are input to the controller 36. The controller 36 includes a data output terminal 42 for outputting raw load data Wy [n] to be described later, and the raw load data Wy [n] is observed at the data output terminal 42. For example, a digital recorder 44 can be connected.

  The controller 36 will be described in more detail. As shown in FIG. 3, the controller 36 includes a CPU (Central Processing Unit) 360. The data output terminal 42 described above is connected to the CPU 360 via a data output circuit 362 as an interface circuit. The controller 36 also has a pulse input circuit 364 as an interface circuit between the encoders 38 and 40, and output pulses of the encoders 38 and 40 are sent to the CPU 360 via the pulse input circuit 364. Entered. The CPU 360 recognizes the rotation angle of the rotary shaft 26 from these output pulses, and consequently recognizes the current position of each unit 14, 14. And based on this recognition result, the positional data mentioned later for every unit 14 is produced | generated.

  Further, the controller 36 includes a motor control circuit 366 as an interface circuit with the motor 34 and an external control circuit 368 as an interface circuit with an external device such as a carry-in conveyor 50 (see FIG. 2) described later. These are also connected to the CPU 360. The CPU 360 is also connected with an operation key 370 as an operation unit, a liquid crystal display 372 as a display unit, and a memory 374 in which a so-called control program for controlling the operation of the CPU 360 is stored. The memory 374 also functions as a first storage unit that stores various parameters such as a zero point measurement time Tz described later.

  Further, the controller 36 is capable of communicating individually and mutually with each weighing machine 18, 18,... Of each unit 14, 14,... .. Has a communication circuit 376 as an interface circuit between the machines 18, 18,. This communication circuit 376 is also connected to the CPU 360.

  On the other hand, as shown in FIG. 4, each weighing machine 18 has, for example, a strain gauge type load cell 180, and the mounting table 16 is coupled to the load cell 180. The load cell 180 generates a measurement signal having a voltage corresponding to the magnitude of the load Wy [n] applied to the load cell 180 (hereinafter, this measurement signal is also represented by a symbol Wy [n]). The load Wy [n] includes not only the weight Wx [n] of the object placed on the mounting table 16 such as the container 12 or the beverage described above, but also the load cell such as the weight of the mounting table 16 itself. The so-called initial load Wi [n] applied to 180 from the beginning is also included.

  The weighing signal Wy [n] generated by the load cell 180 is amplified by the amplifier circuit 182 and then input to the A / D conversion circuit 184 as sampling means. The A / D conversion circuit 184 is, for example, of the Δ-Σ type, samples the input measurement signal at a predetermined sampling period Tc, and generates raw load data (hereinafter, this raw measurement data) as a discrete (digital) measurement signal. The load data is also converted to Wy [n]. The converted raw load data Wy [n] is input to the CPU 186. Note that the sampling period Tc of the A / D conversion circuit 184 is, for example, about 1 ms.

  The CPU 186 performs moving average processing as filter processing on the raw load data Wy [n] input from the A / D conversion circuit 184. Therefore, the CPU 186 has a built-in adaptive filter 188 responsible for the moving average process, and strictly has a function as the adaptive filter 188. Then, the CPU 186 calculates the total load Wy [n] currently applied to the load cell 180 based on the so-called processed load data Wy [n] ′ after the processing by the adaptive filter 188, and finally the final measurement described later. The value Ws [n] is calculated. The calculated final measured value Ws [n] (specifically, data representing the final measured value Ws [n]) is transmitted to the controller 36 via the communication circuit 190 serving as an interface circuit with the controller 36. Is done.

  Further, the weighing machine 18 also functions as a control means for controlling the valve 20 corresponding to itself. In order to realize this, the weighing machine 18 has a valve control circuit 192 as an interface circuit with the valve 20. The valve control circuit 192 is also connected to the CPU 186. The CPU 186 operates according to a control program stored in the memory 194. The memory 194 also functions as a second storage unit that stores various parameters such as the zero point measurement time Tz described above.

  The weight type filling device 10 configured as described above operates as follows, thereby realizing quantitative filling. That is, when an operation for starting operation is performed by the operation key 370 of the controller 36, the motor 34 is activated. As a result, the turntable 24 rotates around the rotation shaft 26 in the direction indicated by the arrow 100 in FIG. 2 at a constant rotation speed, for example, at a rate of one rotation every several seconds (about 3 seconds to 5 seconds). As the turntable 26 rotates, the units 14, 14,..., The storage tank 28, and the pipes 30, 30,. Further, in accordance with the activation of the motor 34, the external device described above, specifically, the carry-in conveyor 50, the carry-in star wheel 52, the carry-out star wheel 54, the carry-out conveyor 56 shown in FIG. The machine also starts.

  Then, as shown by an arrow 102 in FIG. 2, empty containers 12, 12,... Are first transported to the loading star wheel 52 by the loading conveyor 50. The carry-in star wheel 52 rotates in the direction indicated by the arrow 104 in FIG. 2 (counterclockwise in FIG. 2), and the containers 12, 12,... One by one is conveyed to the turntable 24. As a result, the containers 12, 12,... Are placed one by one on the placement tables 16, 16,. The containers 12 set in the respective units 14 in this way naturally rotate in the direction indicated by the arrow 100 together with the units 14. During this rotation, each unit 14 fills the container 12 set therein with a beverage.

  Specifically, referring to FIG. 2 and FIG. 5, an arbitrary unit 14 having an identification number “n” (hereinafter referred to as unit n for convenience of description) has its own container 12. Zero point measurement is started at time t0 when reaching a predetermined position Pz before being set. In this way, whether the unit n has reached the zero point measurement start position Pz is recognized based on the position data provided from the controller 36. That is, as described above, the controller 36 (CPU 362) recognizes the current positions of all the units 14, 14,... Including the unit n based on the output pulses of the encoders 38 and 40. And based on this recognition result, the position data showing the said present position are produced | generated, and the produced | generated position data are transmitted to all the units 14, 14, .... Each unit 14 (CPU 180) recognizes its current position including whether or not it has reached the zero point measurement start position Pz based on this position data (see FIG. 5B).

  The zero point measurement started at the time point t0 when the zero point measurement start position Pz is reached is performed from the time point t0 until a predetermined zero point measurement time Tz elapses. Specifically, the unit n sets the value of the processed load data Wy [n] ′ at the time point t1 when the zero point measurement time Tz has elapsed as the zero point measurement value Wz [n]. That is, the zero point measurement value Wz [n] is obtained based on the following equation 1.

  This zero point measurement value Wz [n] is temporarily stored in the memory 194 in the unit n. The zero point measurement value Wz [n] includes the component of the initial load Wi [n] described above. The zero point measurement time Tz is the natural vibration of the load cell 180 at least at the time point t1, that is, when there is no load (Wx [n] = 0) state where no object is placed on the placement table 16. It is set longer than one cycle of the cycle Tx [n], for example, about 0.1 to 0.3 seconds. The natural vibration period Tx [n] will be described in detail later.

  After the zero point measurement, the empty container 12 is set in the unit n by the carry-in star wheel 52 as described above. Then, after the empty container 12 is set in this way, the unit n starts measuring the weight of the empty container 12, that is, the so-called tare weight, at the time t2 when it reaches the predetermined tare measurement start position Pr. This tare weight measurement is performed from the time t2 until a predetermined tare measurement time Tr elapses. Specifically, the unit n sets the value of the processed load data Wy [n] ′ at the time point t3 when the tare measurement time Tr has elapsed as the tare measurement value Wq [n]. That is, the tare measurement value Wq [n] is obtained based on the following equation 2.

The tare measurement value Wq [n] is also temporarily stored in the memory 194. Note that the tare measurement time Tr is at least one time period of the natural vibration period Tx [n] of the load cell 180 at the time point t3, that is, when the empty container 12 is mounted on the mounting table 16. Is set longer, for example, about 0.1 to 0.3 seconds.

  Then, the unit n subtracts the above-described zero point measurement value Wz [n] from the tare measurement value Wq [n], and sets the result (Wq [n] −Wz [n]) as the tare weight Wr [n]. That is, the tare weight Wr [n] is obtained based on the following Equation 3.

  Further, the unit n determines whether or not the container 12 set in the unit n conforms to a predetermined standard based on the tare weight Wr [n]. Specifically, it is determined whether or not the tare weight Wr [n] is within a predetermined range Wf ± β based on the standard value Wf defined in the standard. For example, when the tare weight Wr [n] is within the range Wf ± β (when Wf−β ≦ Wr [n] ≦ Wf + β), it is determined that the container 12 is set according to the standard. On the other hand, when the tare weight Wr [n] is outside the range Wf ± β (in the case of Wr [n] <Wf−β or Wr [n]> Wf + β), the container 12 is not set as specified. And waits for the opportunity to reach the zero measurement start position Pz.

  When it is determined that the tare measurement is completed and the container 12 conforming to the standard is set, the unit n immediately opens its valve 20. Thereby, the beverage starts to be supplied to the container 12. At this time, the valve 20 is opened with a relatively large diameter, so to speak, it is in a large input state. Then, when the weight of the beverage supplied to the container 12 reaches a predetermined switching weight Wa [n], specifically, at the time t4 when the following equation 4 is established, the unit n slightly reduces the diameter of the valve 20.

  Thus, by reducing the diameter of the valve 20 a little, the supply amount per unit time of the beverage from the valve 20 to the container 12 is reduced to about 1/2 to 1/3 of the large charging state, In other words, it becomes a small input state. The unit n closes the valve 20 when the weight of the beverage supplied to the container 12 reaches the predetermined supply stop weight Wb [n], specifically, at the time t5 when the following equation 5 is established.

  Even if the valve 20 is closed, the supply of the beverage to the container 12 is not immediately stopped, and the beverage is continuously supplied to the container 12 for a while. This is mainly due to the fact that there is a distance (drop) between the valve 20 and the container 12 and that there is a time lag between when Formula 5 is established and when the valve 20 is closed. Accordingly, until the supply of the beverage is completely stopped and the swing of the load cell 180 due to the stop of the supply of the beverage can be regarded as having converged to some extent, in other words, from the time t5 to the predetermined time Until the stable waiting time Tw elapses, the unit n is in a standby state. This stable waiting time Tw is, for example, about 0.3 to 0.5 seconds, although it depends on the magnitude of the total load Wy [n], the required measurement accuracy, and the like.

  At time t6 when the stabilization waiting time Tw elapses, the unit n starts final measurement in order to measure the final weight of the beverage supplied to the container 12. This final measurement is performed from the time t6 until a predetermined final measurement time Ts elapses. Specifically, the unit n substitutes the value of the processed load data Wy [n] ′ at the time point t7 when the final measurement time Ts has passed into the following Equation 6 to obtain the final weight of the filled beverage A final metric value Ws [n] representing is obtained. This final measured value Ws [n] is also temporarily stored in the memory 194.

  The final measured value Ws [n] is ideally equivalent to the target beverage filling weight, so-called target value Wt (Ws [n] = Wt). The supply stop weight Wb [n] described above is set so as to satisfy such a relationship. That is, the weight (Bt−ΔW [n]) obtained by subtracting the so-called drop amount ΔW [n] of the beverage supplied to the container 12 after the valve 20 is closed from the target value Wt is the supply stop weight Wb [ n]. The final measurement time Ts is the natural vibration period Tx [of the load cell 180 at least at the time point t7 (or t6), that is, when the container 12 is filled with a beverage having a weight substantially equal to the target value Wt. n] is set to be longer than one cycle, for example, about 0.1 to 0.3 seconds.

  After the end of the final measurement, at the time point t8 when the unit n reaches the predetermined end position Pe, the unit n sends the above-mentioned final measured value Ws [n] (specifically, data representing the final measured value Ws [n]) to the controller. 36. The controller 36 receives the final weighing value Ws [n] sent from the unit n.

  After a while after the unit n passes the end position Pe, the container 12 set in the unit n is removed by the unloading star wheel 54 as indicated by an arrow 106 in FIG. Thus, a series (one time) of the filling operation by the unit n is completed. The container 12 removed from the unit n is conveyed by the carry-out conveyor 56 as indicated by an arrow 108 and sent to the above-described sorter.

  On the other hand, when the controller 36 receives the final measured value Ws [n] sent from the unit n that has reached the end position Pe as described above, the controller 36 performs the control as specified by the unit n based on the final measured value Ws [n]. It is determined whether or not a constant amount of filling has been performed. Specifically, it is determined whether or not the final measured value Ws [n] is within a predetermined range Wt ± γ based on the target value Wt, and is within the range Wt ± γ ( In the case of Wt−γ ≦ Ws [n] ≦ Wt + γ), it is determined that the constant filling according to the standard is performed by the unit n. In this way, the container 12 determined to be a non-defective product is passed to the next process, for example, the capping process, through the sorting process by the sorter described above. On the other hand, when the final measurement value Ws [n] is out of the predetermined range Wt ± γ (when Ws [n] <Wt−γ or Ws [n]> Wt + γ), the controller 36 is defined by the unit n. Judging that the regular filling was not performed. Then, an instruction is given to the above-described sorter so that the container 12 determined to be a defective product is excluded from the production line.

  Note that the unit n cannot obtain the final measurement value Ws [n] until it reaches the end position Pe, for example, because the valve 20 is clogged, in other words, the time point t7 for the unit n is It may not arrive. In such a case, the unit n transmits data “Error” to the controller 36 as the final measured value Ws [n]. Also in this case, the controller 36 gives an instruction to the sorter so that the container 12 that is set to “Error” is excluded from the production line.

  By the way, the load cell 180 of the unit n always emits natural vibration. This natural vibration appears as noise with respect to the weighing signal Wy [n] output from the load cell 180. Of course, the natural vibration also appears in the raw load data Wy [n]. Moreover, the frequency, the so-called natural vibration frequency (natural frequency) fx [n] is not constant, and varies depending on the magnitude of the load Wy [n] applied to the load cell 180. Therefore, in order to perform accurate quantitative filling in the manner described above, it is necessary to remove such natural vibration components from the raw load data (measurement signal) Wy [n]. For this reason, in this embodiment, the above-mentioned adaptive filter 188 is provided. This will be described in detail below.

  That is, the natural vibration frequency fx [n] of the load cell 180 is expressed by the following formula 7.

  Here, h [n] is the spring constant of the load cell 180, and g is the gravitational acceleration. The load Wy [n] in Equation 7 includes the initial load Wi [n] in addition to the weight Wx [n] of the object placed on the placement table 16 as described above. The number 7 can be expressed as the following number 8.

  The adaptive filter 188 controls its own frequency characteristics so as to form an attenuation region in a band including the natural vibration frequency fx [n] expressed by this equation 8 (or equation 7). In order to realize this, as shown in FIG. 6, the adaptive filter 188 includes a moving average processing unit 200 as filter means to which the raw load data Wy [n] is input, and the moving average processing unit 200 includes A constant setting unit 202 as update means for setting the filter constant Cx [n]. The filter constant Cx [n] is an integer of 1 or more (Cx [n] ≧ 1).

  As shown in FIG. 7, the moving average processing unit 200 includes a plurality of, for example, Cmax registers 204, 204,. The raw load data Wy [n] input to the moving average processing unit 200 is first stored in the first register 204 (at the left end in FIG. 7). The raw load data Wy [n] stored in the first register 204 is input to the sampling timing of the A / D conversion circuit 184 every time new raw load data Wy [n] is input. At the same time (that is, at the same time interval as the sampling period Tc), the data is sequentially moved (shifted) to the register 204 on the right side. The raw load data Wy [n] moved to the last register (Cmax-th register at the right end in FIG. 7) is sent out at the next sampling timing by the A / D conversion circuit 184 and discarded. The moving average processing unit 200 uses the constant setting unit 202 to obtain new data with the sampling time among the plurality of raw load data Wy [n] stored in these registers 204, 204,. The number according to the set filter constant Cx [n] is extracted. That is, the raw load data Wy [n] stored in the first to Cx [n] -th registers 204, 204,. Then, an average value of the extracted raw load data Wy [n] is obtained, and this average value is output as processed load data Wy [n] '.

  The processed load data Wy [n] ′ output from the moving average processing unit 200 is also input to the constant setting unit 202. The constant setting unit 202 is based on the input processed load data Wy [n] ′, and is a filter constant Cx [n], that is, raw load data to be subjected to moving average processing (averaging) by the moving average processing unit 200. The number of Wy [n], that is, the so-called tap number is obtained. Specifically, first, the weight Wx [n] of the object currently mounted on the mounting table 16 is obtained by substituting the processed load data Wy [n] ′ into the following equation (9).

  Then, the constant setting unit 202 obtains the current natural vibration frequency fx [n] by substituting the weight Wx [n] obtained by the equation 9 into the above equation 8, and thus the natural vibration period which is the reciprocal thereof. Tx [n] (= 1 / fx [n]) is obtained. Further, the filter constant Cx [n] is obtained by substituting this natural vibration period Tx [n] into the following Equation 10.

  In Equation 10, ε is an integer equal to or greater than 1 (ε ≧ 1), and is, for example, “1” here. In addition, the decimal part of the solution Cr [n] of Formula 10 is rounded off, for example.

  The constant setting unit 202 newly sets the filter constant Cx [n] obtained by Equation 10 in the moving average processing unit 200. That is, the filter constant Cx [n] is updated. The constant setting unit 202 updates the filter constant Cx [n] every time new processed load data Wy [n] ′ is output from the moving average processing unit 200, in other words, the A / D conversion circuit 184. Is executed each time the weighing signal Wy [n] is sampled. It should be noted that the total number Cmax of the registers 204, 204,... Is equal to or slightly larger than the maximum value expected as the filter constant Cx [n].

  By updating the filter constant Cx [n] in this way, the moving average processing unit 200 can detect the natural vibration period Tx at the current sampling time, that is, the previous sampling time by the A / D conversion circuit 184. The average value of the raw load data Wy [n] for one cycle (or substantially one cycle) of [n] is output as processed load data Wy [n] ′. As a result, the frequency characteristic of the moving average processing unit 200 is a characteristic having a notch frequency at the same frequency (or the vicinity thereof) as the natural vibration frequency fx [n] and an integer multiple thereof, in other words, the natural vibration frequency. An attenuation band is formed in the frequency band including fx [n]. As a result, the natural vibration component included in the raw load data Wy [n] is effectively removed, and clean processed load data Wy [n] ′ that does not include the natural vibration component can be obtained.

  The natural vibration frequency fx [n] increases as the load Wy [n] applied to the load cell 180 increases, that is, as the weight Wx [n] of the object mounted on the mounting table 16 increases. In addition, the natural vibration frequency fx [n] varies depending on the size of the load cell 180, but is, for example, about 50 Hz to 200 Hz. In other words, the natural vibration period Tx [n] is about 5 ms to 20 ms.

  Here, as described above, the sampling period Tc of the A / D conversion circuit 184 is 1 ms, which is much shorter than the natural vibration period Tx [n] of about 5 ms to 20 ms. Then, the filter constant Cx [n] is updated at the same time interval as the sampling period Tc. That is, according to this embodiment, the filter constant Cx [n] is updated at a time interval Tc (<< Tx [n]) that is much shorter than the natural vibration period Tx [n]. Therefore, the conventional technique described above, specifically, the conventional technique in which it takes at least one period of the natural vibration period or an integral multiple thereof to update the integration time width corresponding to the filter constant Cx [n]. Compared with, extremely high output responsiveness can be obtained. Therefore, even when the weighing signal W [n] fluctuates drastically as in the weight-type filling device 10 described in this embodiment, the natural vibration component included in the weighing signal (raw load data) Wy [n] is reduced. It can be effectively removed and high measurement accuracy can be obtained.

  With reference to FIG. 8, the update process of the filter constant Cx [n] of the moving average processing unit 200 and the operation of the moving average processing unit 200 will be described in a little more detail.

  That is, at a certain sampling time k, for example, a value of “10” is set as the filter constant Cx [n, k] (k: index k representing the sampling time). That is, the natural vibration period Tx [n, k−1] at the previous sampling time k−1 is 10 ms, and the “10” is based on the natural vibration period Tx [n, k−1] of 10 ms. Assume that the filter constant Cx [n, k] is set. In this case, as shown in FIG. 8A, the ten raw load data Wy [n, k] to Wy [stored in the ten registers 204, 204,. The average value of n, k-9] is output as processed load data Wy [n, k] ′.

  Then, a new filter constant Cx [n, k + 1] obtained based on the processed load data Wy [n, k] ′, specifically, the processed load data Wy [n, k] is set to the above-mentioned number. The weight Wx [n, k] is calculated by substituting 9 and the natural vibration frequency fx [n, k] is calculated by substituting the weight Wx [n, k] into Equation 8, and this natural vibration is further calculated. A new filter constant Cx [n, k + 1] obtained by substituting the natural vibration period Tx [n, k], which is the reciprocal of the frequency fx [n, k], into the equation 10 is, for example, the previous filter constant Cx [ n, k]. In other words, the difference (| Tx [n, k] −Tx [n) between the natural vibration periods Tx [n, k] and Tx [n, k + 1] between the previous sampling time k and the current sampling time k + 1. , K + 1] |) is less than 0.5 ms. Also in this case, as in the previous case, as shown in FIG. 8 (b), ten raw load data Wy [n, stored in a total of ten registers 204, 204,. The average value of k + 1] to Wy [n, k-8] is output as processed load data Wy [n, k + 1] ′.

  Here, the previous sampling time k and the current sampling time k + 1 share much of the raw load data Wy [n] that is the target of the moving average process, specifically at any sampling time k and k + 1. Note that nine raw load data Wy [n, k] to Wy [n, k-8] that are common to each other are included in the object of the moving average process. This means that even if the filter constant Cx [n] is updated, the processed load data Wy [n] 'as the filter output does not vary greatly, that is, a smooth filter output is obtained.

  On the other hand, it is assumed that the updated filter constant Cx [n, k + 1] is larger by, for example, “1” than the previous filter constant Cx [n, k]. In other words, compared to the natural vibration period Tx [n, k] at the previous sampling time k, the natural vibration period Tx [n, k + 1] at the current sampling time k + 1 is increased by 0.5 ms or more and less than 1.5 ms. Suppose that In this case, as shown in FIG. 8C, 11 raw load data Wy [n, k + 1] to Wy [stored in 11 registers 204, 204,. The average value of n, k-9] is output as processed load data Wy [n, k + 1] ′. Of these 11 raw load data Wy [n, k + 1] to Wy [n, k-9], 10 raw load data Wy [n, k] to Wy [n, k-9] This is common with the target of the moving average process at the previous sampling time k shown in FIG. Therefore, also in this case, the processed load data Wy [n] ′ does not vary greatly before and after the filter constant Cx [n] is updated, and a smooth filter output is obtained.

  On the contrary, it is assumed that the updated filter constant Cx [n, k + 1] is smaller by “1” than the previous filter constant Cx [n, k]. In other words, compared with the natural vibration period Tx [n, k] at the previous sampling time k, the natural vibration period Tx [n, k + 1] at the current sampling time k + 1 is decreased by 0.5 ms or more and less than 1.5 ms. Suppose that In this case, as shown in FIG. 8D, nine raw load data Wy [n, k + 1] to Wy [stored in nine registers 204, 204,. The average value of n, k-7] is output as processed load data Wy [n, k + 1] ′. Of these nine raw load data Wy [n, k + 1] to Wy [n, k-7], eight raw load data [n, k] to Wy [n, k-7] This is common with the target of the moving average process at the previous sampling time k. Therefore, even when the filter constant Cx [n] is decreased by “1” in this way, a smooth filter output can be obtained.

  If the filter constants Cx [n, k] and Cx [n, k + 1] are different from each other by “2” or more at the previous sampling time k and the current sampling time k + 1, the difference (| Cx [n, As k] −Cx [n, k + 1] |) is larger, the objects of moving average processing (number and value of raw load data Wy [n]) at the respective sampling times k and k + 1 are different. As the moving average processing target deviates in this way, the processed load data Wy [n] ′ may fluctuate relatively large, and a smooth filter output may not be obtained. However, increasing or decreasing the filter constant Cx [n] by “2” or more means that the natural vibration period Tx [n] is 1 within the short update period (one period of the sampling period Tc) of the filter constant Cx [n] of 1 ms. .5 ms or more, and this is extremely rare (none). Therefore, in this embodiment, the filter constant Cx [n] does not change extremely, and a smooth filter output is always obtained.

  On the other hand, for example, when the above-described prior art is applied to the weight filling apparatus 10, the following problem occurs. That is, in the prior art, as described above, it takes time for one cycle of the natural vibration period of the measuring unit or an integral multiple thereof to update the integration time width. For this reason, a big difference arises between the integration time width actually used for the integration process and the integration time width required for removing the natural vibration component. As a result, the natural vibration component included in the measurement signal cannot be accurately removed, and the weight data after the integration process varies greatly. That is, the prior art cannot be applied to a so-called dynamic balance that requires high responsiveness like the weight-type filling device 10, in other words, high responsiveness is not required like that of a platform scale (substantially). It can only be applied to so-called static balances, which measure the weight of an object to be weighed from a constant and constant weighing signal.

  In order to obtain the natural vibration frequency fx [n] based on the above equation 8, it is necessary to accurately obtain the spring constant h [n] and the initial load Wi [n] included in the equation 8 in advance. . That is, the spring constant h [n] is not only the spring constant of the load cell 180 itself, but also the spring constant of the mounting table 16 coupled to the load cell 180, and the spring constant of a coupling member (not shown) that couples both. Also involved. On the other hand, the initial load Wi [n] includes not only the weight of the mounting table 16 but also the weight of the coupling member, and is also influenced by the weight (self-weight) of the load cell 180 itself. Therefore, it is very difficult to directly measure the spring constant h [n] and the initial load Wi [n], and it is substantially impossible to obtain the value by a calculation (theoretical value) on a desk. Therefore, in this embodiment, the following adjustment mode is provided in order to accurately obtain the spring constant h [n] and the initial load Wi [n]. It is assumed that zero point adjustment, span adjustment, and temperature drift adjustment have already been performed by a known method before entering this adjustment mode.

  In the adjustment mode, first, the natural vibration frequency fz [n] at no load is obtained for each unit 14. Specifically, for each unit 14, a predetermined impact is applied, for example, by an impulse hammer (not shown) in a state where no object is placed on the mounting table 16. At this time, the raw load data Wy [n] of the unit 14 to which the impact is applied is input to the digital recorder 44 via the controller 36 and is recorded by the digital recorder 44 over a certain period of time. The natural vibration frequency fz [n] is read by the operator from the raw load data Wy recorded in the digital recorder 44, and the read natural vibration frequency fz [n] is written down in, for example, a note.

  Next, for each unit 14, the natural vibration frequency fm [n] when the existing test load Wm is placed on the placement table 16 is obtained. Specifically, a predetermined impact is applied to each unit 14 by the above-described impulse hammer while the test load Wm is placed. Then, the raw load data Wy [n] of the unit 14 to which this impact is applied is recorded in the digital recorder 44 over a certain period of time as described above. The natural vibration frequency fm [n] is read from the raw load data Wy [n] recorded in the digital recorder 44, and the read natural vibration frequency fm [n] is written down in a note. The natural vibration frequency fm [n] is a value smaller than the natural vibration frequency fz [n] at no load.

  In this way, after the natural vibration frequencies fz [n] and fm [n] are obtained for all the units 14, 14,..., These natural vibration frequencies fz [n] and fm [n] Input from 36 operation keys 370 (by the operator). The input natural vibration frequencies fz [n] and fm [n] (strictly, these data) are stored in the memory 374 in the controller 36 and then transmitted to the corresponding unit 14 (the weighing machine 18). The

  Subsequently, by operating the operation key 370, the above-described parameters, specifically, the zero point measurement time Tz, the tare measurement time Tr, the stable waiting time Tw, the final measurement time Ts, the tare standard value Wf, and the tare standard value Wf. , The reference weight Wb [n], the supply stop weight Wb [n], the target value Wt, and the allowable value γ based on the target value Wt are input. These parameters are also stored in the memory 374 in the controller 36, and then transmitted to the units 14, 14,... Except for the target value Wt and the allowable value γ. Strictly speaking, among the parameters, the switching weight Wa [n] and the supply stop weight Wb [n] are set to individual values for each unit 14, and this is the purpose of the present invention. In this case, it is assumed that common values Wa [n] and Wb [n] are set to the units 14, 14,.

  On the other hand, in each unit 14, the natural vibration frequencies fz [n] and fm [n] sent from the controller 36 and the parameters (Tz, Tr, Tw, Ts, Wf, β, Wa [n] ] And Wb [n]) are stored in the memory 194. Furthermore, based on the received natural vibration frequencies fz [n] and fm [n], the spring constant h [n] and the initial load Wi [n] are obtained as follows.

  That is, the natural vibration frequencies fz [n] and fm [n] are expressed by Equation 11 and Equation 12, respectively.

  Then, by the simultaneous equations consisting of these formulas 11 and 12, arithmetic expressions for obtaining the spring constant h [n] and the initial load Wi [n] are derived as the following formulas 13 and 14, respectively.

  The spring constant h [n] and the initial load Wi [n] are obtained by these equations 13 and 14, and the obtained values h [n] and Wi [n] are stored in the memory 194.

  Thus, after the spring constant h [n] and the initial load Wi [n] are obtained for each unit 14 in the adjustment mode, the operation mode is switched. Then, the filling operation is performed in the manner described above by this operation mode. In the filling operation in this operation mode, the time points t0 to t1 are based on the zero point measurement time Tz, the tare measurement time Td, the stabilization wait time Tw, and the final measurement time Ts as parameters stored in the memory 194 in the adjustment mode. Zero measurement over time, tare measurement over time points t2 to t3, stabilization wait over time points t5 to t6, and final measurement over time points t6 to t7. Then, based on the tare standard value Wf and the allowable value β stored as the parameters, it is determined whether or not the standard container 12 has been set. Furthermore, based on the switching weight Wa [n] and the supply stop weight Wb [n], a time point t4 that is a beverage supply amount switching timing and a time point t5 that is a supply stop timing are determined.

  In order to realize a series of operations in each of the adjustment mode and the operation mode, the CPU 360 of the controller 36 and the CPU 186 of each unit 14 (the weighing machine 18) perform multitask processing as follows. Note that, during this process, for example, when an interrupt operation for forcible termination is performed by the operation key 370, the process being executed is forcibly terminated immediately. In the adjustment mode, it is assumed that the digital recorder 44 is connected to the data output terminal 42 of the controller 36 in advance. .., And zero adjustment, span adjustment, and temperature drift adjustment are performed in advance (before entering the adjustment mode) by a known method as described above.

  First, when the adjustment mode is selected by operating the operation key 370, the CPU 360 of the controller 36 executes the adjustment mode task shown in FIGS. That is, in step S101 in FIG. 9, all the units 14, 14,... Are notified that the adjustment mode has been selected. In step S103, in order to validate the unit 14 whose identification number n is “1” among the units 14, 14,..., A value “1” is set in the index n for identifying the identification number n. Set. In step S105, the raw load data Wy [n] are sequentially acquired from the activated unit 14 (hereinafter, this activated unit 14 is also expressed as unit n), and the obtained raw data is obtained. The load data Wy [n] is sequentially transferred to the data output circuit 362. Thus, the raw load data Wy [n] is input to the digital recorder 44 connected to the data output terminal 44 via the data output circuit 362. In step S107, the CPU 360 displays a message on the display 372 requesting that the unit n be measured for the natural vibration frequency fz [n] at no load.

  After displaying this message, the CPU 360 proceeds to step S109 and waits for the measurement of the natural vibration frequency fz [n] for the unit n to end. That is, during this period, an impact is applied to the unit n by the impulse hammer as described above in a state where no object is placed on the mounting table 16 of the unit n. The raw load data Wy [n] of the unit n at this time is recorded in the digital recorder 44, and the natural vibration frequency fz [n] of the unit n is read by the operator from the recorded raw load data Wy [n]. It is done. The read natural vibration frequency fz [n] is written down in a note by the operator.

  After the natural vibration frequency fz [n] of the unit n is obtained in this way (written in a note), an operation to the effect that the measurement of the natural vibration frequency fz [n] of the unit n is completed by the operation key 370. When the process is completed, the CPU 360 proceeds from step S109 to step S111. In step S111, it is determined whether or not the value of the index n described above is equal to the maximum value “N”. Here, when the value of the index n is not equal to the maximum value “N”, that is, when there is a unit 14 for which the measurement of the natural vibration frequency fz [n] has not been completed, the natural vibration frequency of the unit 14 is determined. Proceed to step S113 to measure fz [n]. Then, after the value of the index n is incremented by “1” in this step S113, the process returns to step S105. On the other hand, when the value of the index n is equal to the maximum value “N”, that is, when the measurement of the natural vibration frequency fz [n] is completed for all the units 14, 14,..., The process proceeds from step S 111 to step S 115.

  In step S115, the CPU 360 sets a value “1” to the index n in order to validate the unit 14 with the identification number n “1”. In step S117, the raw load data Wy [n] is sequentially acquired from the validated unit n, and the acquired raw load data Wy [n] is sequentially transferred to the data output circuit 362. In step S119, a message requesting that measurement of the natural vibration frequency fm [n] with the test load Wm applied to the unit n is requested is displayed on the display 372. Then, the process proceeds to step S121 and waits for the measurement of the natural vibration frequency fm [n] for the unit n to end.

  During this time, an impact is applied to the unit n by the impulse hammer while the test load Wm is placed on the placing table 16 of the unit n. The raw load data Wy [n] of the unit n at this time is recorded in the digital recorder 44, and the natural vibration frequency fm [n] of the unit n is read from the recorded raw load data Wy [n]. Then, the read natural vibration frequency fz [n] is written down in a note.

  After the natural vibration frequency fz [n] of the unit n is obtained in this way, when the operation indicating that the measurement of the natural vibration frequency fm [n] of the unit n is finished is performed by the operation key 370, the CPU 360 is performed. Advances from step S121 to step S123. In step S123, it is determined whether or not the value of the index n is equal to the maximum value “N”. Here, when the value of the index n is not equal to the maximum value “N”, that is, when there is a unit 14 for which the measurement of the natural vibration frequency fm [n] has not yet been completed, the natural vibration frequency of the unit 14 is determined. The process proceeds to step S125 to measure fm [n]. And then. In step S125, after the value of index n is incremented by “1”, the process returns to step S117. On the other hand, when the value of the index n is equal to the maximum value “N”, that is, when the measurement of the natural vibration frequency fm [n] is completed for all the units 14, 14,..., The process proceeds from step S 123 to step S 127 in FIG. move on.

  In step S127, the CPU 360 sets a value “1” to the index n in order to re-enable the unit 14 having the identification number n “1”. In step S129, a message that requests input of the natural vibration frequency fz [n] when the activated unit n is unloaded is displayed on the display 372. Then, the process proceeds to step S131 and waits for the natural vibration frequency fz [n] to be input.

  When the natural vibration frequency fz [n] of the unit n is input from the operation key 370 in step S131, the CPU 360 proceeds to step S133. In step S133, the input natural vibration frequency fz [n] (strictly, the data representing the natural vibration frequency fz [n]) is stored in the memory 374, and in step S135, the natural vibration frequency fz [n] is stored. The frequency fz [n] is transmitted to the unit n. Then, the process proceeds to step S137.

  In step S137, the CPU 360 displays on the display 372 a message requesting input of the natural vibration frequency fm [n] when the test load Wm is applied to the unit n. Then, the process proceeds to step S139 and waits for the input of the natural vibration frequency fm [n].

  Here, when the natural vibration frequency fm [n] of the unit n is input by operating the operation key 370, the CPU 360 proceeds to step S141. In step S141, the input natural vibration frequency fm [n] is stored in the memory 374. In step S143, the natural vibration frequency fm [n] is transmitted to the unit n, and then in step S145. move on.

  In step S145, the CPU 360 determines whether or not the value of the current index n is equal to the maximum value “N”. Here, when the value of the index n is not equal to the maximum value “N”, that is, when there is a unit 14 for which the input of the natural vibration frequencies fz [n] and fm [n] by the operation key 370 has not been finished yet. Advances to step S147 to request the unit 14 to input the natural vibration frequencies fz [n] and fm [n]. In step S147, the index n is incremented by “1”, and the process returns to step S129. On the other hand, when the value of the index n is equal to the maximum value “N”, that is, when the input of the natural vibration frequencies fz [n] and fm [n] is completed for all the units 14, 14,. No. 11 step S149.

  In step S149, the CPU 360 sends a message to the display 372 requesting input of the parameters (Tz, Tr, Tw, Ts, Wf, β, Wa [n], Wb [n], Wt, and γ) described above. indicate. Then, the process proceeds to step S151 and waits for the input of the parameter. Here, when any parameter is input from the operation key 370, the CPU 360 proceeds to step S 153 and stores the input parameter in the memory 374. In step S155, it is determined whether all parameters have been input. If all parameters have not yet been input, the process returns to step S151. On the other hand, if all parameters have been input, the process proceeds from step S155 to step S157.

  In step S157, the CPU 360 transmits the parameters stored in the memory 374 to all the units 14, 14,. However, as described above, among the parameters, the target value Wt and the allowable value γ are not transmitted. In step S159, the adjustment mode is now in the final setting stage, and more specifically, the above-described spring constant h [n] and initial load Wi [n] are being calculated in each unit 14. A message indicating that is displayed on the display 372 for a certain period of time. Then, the adjustment mode task is completed with the elapse of the predetermined time.

  For the operation of the CPU 360 on the controller 36 side, the CPU 186 on each unit 14 (unit n) side executes the pair adjustment mode task shown in FIG. 12 in the adjustment mode.

  That is, upon receiving a notification that the adjustment mode has been selected from the controller 36, the CPU 186 proceeds to step S201, and the raw load data Wy [n] input from the A / D conversion circuit 184 is left as it is. Transfer to the communication circuit 192. Thus, the raw load data Wy [n] is transmitted to the controller 36 via the communication circuit 192.

  Then, in step S203, the CPU 186 waits for the no-load natural vibration frequency fz [n] (specifically, data representing the natural vibration frequency fz [n]) to be sent from the controller 36. When the natural vibration frequency fz [n] is received, the process proceeds to step S205, and the received natural vibration frequency fz [n] is stored in the memory 194. Further, the process proceeds to step S207, and this time, it waits for the natural vibration frequency fm [n] when the test load Wm is applied to be sent from the controller 30.

  In step S207, upon receiving the natural vibration frequency fm [n] from the controller 30, the CPU 186 proceeds to step S209, and stores the received natural vibration frequency fm [n] in the memory 194. In step S211, the spring constant h [n] and the initial load Wi [n] are calculated based on these natural vibration frequencies fz [n] and fm [n], that is, based on the above-described equations 13 and 14. The calculation results h [n] and Wi [n] are stored in the memory 194.

  Further, the CPU 186 proceeds to step S213 to obtain an initial value Ci [n] of the filter constant Cx [n]. Specifically, the spring constant h and the initial load Wi [n] obtained in step S211 are substituted into the above equation 8, and zero (0) is substituted as the weight Wx [n] in the equation 8, The natural vibration frequency fx [n] at the time of loading is obtained. Then, by substituting the natural vibration period Tx [n], which is the reciprocal number of the natural vibration frequency fx [n] at the time of no load, into the above-described formula 10, the filter constant Cx [n] for the no load is obtained. The no-load filter constant Cx [n] is set as an initial value Ci [n]. The reciprocal of the natural vibration frequency fz [n] at the time of no load (received from the controller 36 in step S203) stored in the memory 194 in step S205 described above may be used as the initial value Ci [n]. The initial value Ci [n] is stored in the memory 194.

  After obtaining the initial value Ci [n] in this way, the CPU 186 proceeds to step S215. In step S215, the controller 30 waits for the above parameters (Tz, Tr, Tw, Ts, Wf, β, Wa [n] and Wb [n]) to be sent. When these parameters are received, the process proceeds to step S217, and the received parameters are stored in the memory 194. With the storage of this parameter, the series of pair adjustment mode tasks are completed.

  On the other hand, in the operation mode, the CPU 360 on the controller 36 side executes the operation mode task shown in FIG. In this operation mode, it is assumed that the digital recorder 44 is removed from the data output terminal 42 of the controller 36.

  That is, when the operation mode is selected by operating the operation key 370, the CPU 360 proceeds to step S301 and notifies all units 14, 14,... That the operation mode has been selected. In step S303, it is determined whether or not an operation start operation has been performed with the operation key 370, that is, whether or not an operation start command has been input from the operation key 370.

  When an operation start command is input in step S303, the CPU 360 proceeds to step S305, starts the motor 34, and instructs the external device such as the carry-in conveyor 50 to start. In step S307, all units 14, 14,... Are instructed to start operation. In step S309, the execution of the timing control task is started. This timing control task will be described in detail later.

  In step S311, the CPU 360 continues until an empty container 12 is set in each unit 14, 14,..., For example, until a certain time elapses when it can be sufficiently considered that the container 12 is set. It will be in a standby state. And after progress of this fixed time, it progresses to step S313 and starts performing a data acquisition task. Thereby, the filling operation by each unit 14 is repeatedly performed in the manner described above. The data acquisition task in step S313 will be described in detail later.

  While the filling operation is performed in this way, in step S315, the CPU 360 determines whether or not an operation stop operation has been performed by the operation key 370, that is, whether or not an operation stop command has been input from the operation key 370. Judging. Here, when an operation stop command is input, the process proceeds to step S317 to instruct all units 14, 14,. Further, after finishing the above-described timing control task in step S319, the data acquisition task is finished in step S321. In step S323, the motor 34 is stopped and the external device is instructed to stop driving. After execution of step S323, the series of operation mode tasks are terminated.

  Details of the above-described timing control task will be described with reference to FIG. In this timing control task, the CPU 360 first acquires the output pulses of the encoders 38 and 40 from the pulse input circuit 364 in step S401. In step S403, “1” is set to the index n described above, and then the process proceeds to step S405.

  In step S405, the CPU 360 determines whether the unit n has reached the zero point measurement start position Pz based on the output pulses of the encoders 38 and 40 acquired in step S401. Here, when the unit n has reached the zero point measurement start position Pz, in step S407, the fact is notified to the unit n by the position data described above, and the process proceeds to step S409. On the other hand, if the unit n has not reached the zero point measurement start position Pz, the process skips step S407 and proceeds directly to step S409.

  In step S409, the CPU 360 determines whether the unit n has reached the tare measurement start position Pr based on the output pulses of the encoders 38 and 40 described above. Here, when the unit n reaches the tare measurement start position Pr, the process proceeds to step S411. After that is notified to the unit n by the position data, the process proceeds to step S413. On the other hand, if the unit n has not reached the tare measurement start position Pr, the process skips step S411 and proceeds directly to step S413.

  In step S413, the CPU 360 determines whether the unit n has reached the end measurement start position Pe based on the output pulses of the encoders 38 and 40. Here, if the unit n has reached the end measurement start position Pe, in step S415, the fact is notified to the unit n by the position data, and the process proceeds to step S417. On the other hand, if the unit n has not reached the end measurement start position Pe, the process skips step S415 and proceeds directly to step S417.

  In step S417, the CPU 360 determines whether or not the value of the index n is equal to the maximum value “N”. Here, when the value of the index n is not equal to the maximum value “N”, that is, when all the units 14, 14,... Proceed to step S419 to execute step S405 to step S415. In step S419, the value of the index n is incremented by “1”, and the process returns to step S405. On the other hand, when the value of the index n is equal to the maximum value “N”, that is, when the steps S405 to S415 have been executed for all the units 14, 14,. In order to execute step S405 to step S415 sequentially from the unit 14, the process returns to step S401.

  Next, details of the data acquisition task will be described with reference to FIG. In this data acquisition task, the CPU 360 first sets “1” to the index n described above in step S501. In step S503, it is determined whether the unit n has reached the end position Pe. If the unit n has reached the end position Pe, the process proceeds to step S505.

  In step S505, the CPU 360 requests the unit n to obtain the final measurement value Ws [n]. In step S507, it is determined whether or not the final measurement value Ws [n] is sent from the unit n. If it is sent, the process proceeds to step S509.

  In step S509, the CPU 360 determines whether or not the final measurement value Ws [n] received from the unit n is “Error”. Here, if it is not “Error”, the process further proceeds to step S511, and it is determined whether or not the final measured value Ws [n] is within the above-described range Wt ± γ. If it is within the range Wt ± γ (Wt−γ ≦ Ws [n] ≦ Wt + γ), the process proceeds to step S513.

  On the other hand, if the final measurement value Ws [n] is “Error” in step S509, the CPU 360 proceeds to step S515. In step S515, an instruction is given to the above sorter so that the unit n is excluded from the production line, and then the process proceeds to step S513. Similarly, when the final measurement value Ws [n] is out of the above-described range Wt ± γ in step S511 (Ws [n] <Wt−γ or Ws [n]> Wt + γ), the process goes through step S515 in the same manner. The process proceeds to S513.

  In step S513, the CPU 360 determines whether the value of the index n is equal to the maximum value “N”. Here, when the value of the index n is not equal to the maximum value “N”, that is, when the final measured values Ws [n] of all the units 14, 14,. Then, after the value of the index n is incremented by “1” in step S517, the process returns to step S503. On the other hand, when the value of the index n is equal to the maximum value “N”, that is, when the final measured values Ws [n] of all the units 14, 14,. In order to obtain the final measured value Ws [n] from the “numbered unit 14”, the process returns to step S501.

  On the other hand, the CPU 186 of each unit 14 is in the operation mode. It operates as follows.

  In other words, upon receiving notification from the controller 36 that the operation mode has been selected, the CPU 186 executes the anti-operation mode task shown in FIG. In this anti-operation mode, first, in step S601, the controller 36 waits for an operation start instruction sent from the controller 36. When this operation start instruction is received, the process proceeds to step S603, where the above-described initial value Ci [n] (see step S213 in FIG. 12) is set as the filter constant Cx [n]. Furthermore, after starting execution of a filter constant update task described later in step S605, execution of a weighing task described later is started in step S607.

  After executing step S607, the CPU 186 proceeds to step S609, and determines whether or not an acquisition request for the final measurement value Ws [n] has been received from the controller 36. Here, when accepting the acquisition request, the CPU 186 proceeds to step S 611 and transmits the final measurement value Ws [n] to the controller 36. After this transmission, the process returns to step S609. On the other hand, when the acquisition request for the final measurement value Ws [n] is not accepted in step S609, the process proceeds to step S613, and it is determined whether an operation stop instruction is accepted from the controller 36. If the operation stop instruction has not been received, the process returns to step S609.

  Upon receiving an operation stop instruction in step S613, the CPU 186 proceeds to step S615 to end the execution of the weighing task started in step S607 described above, and in step S617, the filter constant update task started in step S605 described above. The execution of is terminated. Then, upon completion of this filter constant update task, a series of anti-operation mode tasks are terminated.

  The filter constant update task will be described in detail with reference to FIG. In this filter constant update task, the CPU 186 first determines whether or not new raw load data Wy [n] is output from the A / D conversion circuit 184 in step S701, in other words, the adaptive filter 188 (moving average processing unit). 200), whether or not the new raw load data Wy [n] is input is determined.

  When new raw load data Wy [n] is input in step S701, the CPU 186 proceeds to step S703, executes the moving average process by the application filter 188 (moving average processing unit 200), and sets a new processed load. Data Wy [n] ′ is obtained. Then, the process proceeds to step S705, and the processed load data Wy [n] 'is substituted into the above-described equation 9 to calculate the weight Wx [n] currently applied to the mounting table 16. Note that, at the first (first) execution of step S705, the measurement value obtained by the above-described zero adjustment is substituted as the zero measurement value Wz [n] in Equation 9. Thereafter, the zero point measurement value Wz [n] obtained by the zero point measurement at time points t0 to t1 is substituted for each filling operation.

  Then, the CPU 186 proceeds to step S707, and calculates a new filter constant Cx [n] based on the processed load data Wy [n] ′ calculated in step S705 described above. That is, the natural vibration frequency fx [n] is calculated by substituting the processed load data Wy [n] ′ into Equation 8, and the natural vibration period Tx [n] that is the reciprocal of the natural signal frequency fx [n]. ] Is substituted into Equation 10 to calculate a new filter constant Cx [n]. In step S709, the new filter constant Cx [n] is set in the applied filter 188 (moving average processing unit 200), that is, the filter constant Cx [n] is updated. After updating the filter constant Cx [n], the process returns to step S701, and steps S701 to S709 are repeatedly executed.

  Next, the weighing task will be described in detail with reference to FIGS. 18 and 19. In this weighing task, the CPU 186 first determines whether or not itself (strictly, the unit 14 having the CPU 186) has reached the zero point measurement start position Pz in step S801 of FIG. This determination is made based on the above-described position data sent from the controller 36. If it is determined that the zero point measurement start position Pz has been reached, the process proceeds to step S803, where a timer for measuring the zero point measurement time Tz is reset and immediately started. In step S805, it is determined whether the zero point measurement time Tz has elapsed.

  When the zero point measurement time Tz elapses, the CPU 186 proceeds from step S805 to step S807, and obtains the zero point measurement value Wz [n] based on the above-described equation 1. That is, the value of the processed load data Wy [n] ′ at the time point t1 when the zero point measurement time Tz has elapsed is set as the zero point measurement value Wz [n]. Then, this zero point measurement value Tz is temporarily stored in the memory 194. The zero point measurement value Wz [n] is also used to determine the weight Wx [n] in step S705 of FIG. 17 (filter constant update task) as described above.

  Then, the CPU 186 proceeds to step S809 to determine whether or not the CPU 186 has reached the tare measurement start position Pr. This determination is also made based on the position data sent from the controller 36. If it is determined that the tare measurement start position Pr has been reached, the process proceeds to step S811, where a timer for measuring the tare measurement time Tr is reset and immediately started. In step S813, it is determined whether or not the tare measurement time Tr has elapsed.

  When the tare measurement time Tr elapses, the CPU 186 proceeds from step S813 to step S815, and obtains the tare measurement value Wq [n] based on the above-described formula 2. That is, the value of the processed load data Wy [n] ′ at the time point t3 when the tare measurement time Tz has elapsed is set as the tare measurement value Wq [n]. The tare measurement value Wq [n] is also temporarily stored in the memory 194.

  Then, the CPU 18 proceeds to step S816 to calculate the tare weight Wr [n] based on the above-described equation 3. That is, the tare weight Wr [n] is calculated by substituting the measured tare value Wq [n] obtained in step S815 and the zero point measured value Wz [n] obtained in step S807 described above into Equation 3.

  Further, the CPU 18 proceeds to step S817 and determines whether or not the container 12 conforming to the standard has been set in itself (unit 14) based on the tare weight Wr [n] calculated in step S816 described above. Specifically, it is determined whether or not the tare weight Wr [n] is within a predetermined range Wf ± β, and if it is present (Wf−β ≦ Wr [n] ≦ Wf + β), it conforms to the standard. Assuming that the container 12 is set, the process proceeds to the next step S819. On the other hand, when the tare weight Wr [n] is out of the predetermined range Wf ± β (Wr [n] <Wf−β or Wr [n]> Wf + β), the container 12 according to the standard is not set. It is considered that the process is returned to step S801.

  In step S819, the CPU 186 controls the valve 20 so that the supply of the beverage to the container 12 is started. As a result, the valve 20 enters a large input state. Then, the CPU 186 proceeds to step S821 in FIG.

  In step S821, the CPU 186 determines whether or not the weight of the beverage supplied to the container 12 has reached the switching weight Wa [n]. That is, it is determined whether or not the above equation 4 is established. Here, when the weight of the supplied beverage has not yet reached the switching weight Wa [n], the process proceeds to step S823 to determine whether or not the beverage has reached the end position Pe. This determination is also made based on the position data described above. If the end position Pe has not yet been reached, the process returns from step S823 to step S821.

  On the other hand, when the weight of the beverage already supplied reaches the switching weight Wa [n] in step S821, in other words, when the time t4 arrives at itself, the CPU 186 proceeds to step S825 and slightly reduces the aperture of the valve 20. As a result, the valve 20 transitions from the large input state to the small input state.

  Then, the CPU 186 proceeds to step S827 and determines whether or not the weight of the supplied beverage has reached the supply stop weight Wb [n]. That is, it is determined whether or not the above formula 5 is established. If the weight of the supplied beverage has not yet reached the supply stop weight Wb [n], the process proceeds to step S829 to determine whether or not the beverage has reached the end position Pe. If the end position Pe has not yet been reached, the process returns from step S829 to step S827.

  On the other hand, when the weight of the beverage already supplied reaches the supply stop weight Wb in Step S827, in other words, when the time t5 arrives at itself, the CPU 186 proceeds to Step S831, and closes the valve 20. Thereby, after the beverage of the above-described drop amount ΔW [n] is supplied to the container 12, the supply of the beverage is completely stopped. In step S833, the CPU 186 resets a timer for measuring the stable waiting time Tw and starts it immediately, and then proceeds to step S835.

  In step S835, the CPU 186 determines whether or not the stabilization waiting time Tw has elapsed. Here, if the stable waiting time Tw has not yet elapsed, the process proceeds to step S837, and it is determined whether or not it has reached the end position Pe. If the end position Pe has not been reached, the process returns to step S835.

  In step S835, when the stable waiting time Tw elapses, in other words, when the time point t6 itself arrives, the CPU 186 proceeds to step S839, resets the timer for measuring the final measurement time Ts, and starts immediately. In step S841, it is determined whether or not the final measurement time Ts has elapsed. If not, the process proceeds to step S843 to determine whether or not the terminal has reached the end position Pe. If the end position Pe has not been reached, the process returns to step S841.

  When the final measurement time Ts has elapsed in step S841, the CPU 186 proceeds to step S845, and calculates the final measurement value Ws [n] based on the above-described equation 6. Specifically, the processed load data Wy [n] ′ at the time point t7 when the final measurement time Ts has elapsed and the tare measurement value Wq [n] stored in the memory 194 in step S815 described above are substituted into Equation 6. Thus, the final measured value Ws [n] is obtained. Then, after storing the final measured value Ws [n] in the memory 194, the process returns to step S801 in FIG. 18 to repeat the filling operation.

  When the end position Pe is reached in step S843, that is, when the final waiting time Tw has elapsed but the final weighing value Ws [n] has not been obtained, the CPU 186 proceeds to step S847, and the final weighing value is reached. The above-described data “Error” is set as Ws [n]. And after execution of this step S847, it returns to step S801 of FIG. Similarly, when the end position Pe is reached in step S837, that is, when the final weighing value Ws [n] is not obtained although the filling of the beverage into the container 12 is completed, the process proceeds to step S847 in FIG. The process returns to step S801.

  Further, when the end position Pe is reached in step S829 described above, that is, when the end position Pe is reached in the small injection state, the CPU 186 closes the valve 20 in step S849. Then, after step S847, the process returns to step S801 in FIG. In addition, when the end position Pe is reached in step S823, that is, when the end position Pe is reached in the large input state, the process returns to step S801 in FIG. 18 through step S849 and step S847.

  As described above, according to this embodiment, the filter constant Cx [n] is updated at a time interval Tc that is much shorter than the natural vibration period Tx [n] of the load cell 180. Accordingly, an extremely high output compared to the above-described conventional technique in which it takes at least one period of the natural vibration period or an integral multiple thereof to update the integration time width corresponding to the filter constant Cx [n]. Responsiveness is obtained. Therefore, even when the weighing signal (raw load data) Wy [n] fluctuates violently with the passage of time, the natural vibration component can be effectively removed from the weighing signal Wy [n]. Obtainable. That is, accurate quantitative filling can be realized in the weight-type filling device 10 that could not be handled by the prior art.

  Further, the spring constant h [n] and the initial load Wi [n] necessary for updating the filter constant Cx [n] are extremely difficult to directly measure as described above or to be obtained by calculation on the desk. However, in this embodiment, the natural vibration frequencies fz [n] and fm [n] are measured under two different conditions, and the spring constant h [n] and the initial load Wi [n] are measured based on the actual measurement results. ] Is demanded. Therefore, for example, the natural vibration frequency fx [n] when an arbitrary weight Wx [n] is applied as compared with the case where the estimated values are applied as the spring constant h [n] and the initial load Wi [n]. ] Can be accurately obtained, and as a result, an appropriate filter constant Cx [n] can be obtained to remove the natural vibration component of the frequency fx.

  In this embodiment, the unit 14, 14,... Has been described by taking as an example a so-called rotary weight-type filling device 10 that performs a filling operation while rotating, but the present invention is not limited to this. For example, the present invention is applied to a so-called parallel (line) type intensive filling device in which a plurality of units are arranged in a row and filling operations are performed in parallel (so as to badge processing) by these units. You may apply.

  In addition, although the number N of the units 14, 14,... Is “36”, other numbers may be used. In addition, the present invention is applied not only to the so-called multiple-type heavy-weight filling device 10 having a plurality of units 14, 14,..., But also to a so-called single-type filling device having only one unit. May be.

  Furthermore, the objects to be weighed are not limited to beverages but may be other liquids such as alcohol and oil. Moreover, this invention is applicable not only to a liquid but to the apparatus which fills solids, such as a granular material, as a to-be-measured object.

  The A / D conversion circuit 184 is a Δ-Σ type, but is not limited thereto. For example, a circuit of another type such as a successive approximation type may be used. However, in order to satisfy the sampling theorem, it is important to use a circuit that can sample at least a period Tc shorter than the natural vibration period Tx [n]. Although the sampling period Tc is 1 ms, other values may be used as long as the sampling theorem is satisfied.

  Further, each of the flowcharts shown in FIGS. 9 to 19 is only an example until it is tired, and is not limited to these as long as the same operation and effect as described in this embodiment can be obtained. . For example, in this embodiment, the spring constant h [n] and the initial load Wi [n] are obtained in each unit 14 (the weighing machine 18), but the spring constant h of all these units 14, 14,. [N] and the initial load Wi [n] may be obtained collectively by the controller 36. Further, in order to simplify the operation in the adjustment mode, the spring constant h [n] and the initial load Wi [n] are obtained only for one arbitrary (or specific) unit 14 and the obtained values are set to other values. The spring constant h [n] and the initial load Wi [n] of the units 14, 14,.

  As the adaptive filter 188, a filter that performs the following double moving average process may be used. That is, as shown in FIG. 20, a double moving average processing unit 210 that receives raw load data Wy [n] and performs a double moving average process on the raw load data Wy [n], The heavy moving average processing unit 210 is provided with a constant setting unit 212 that sets two filter constants, a first filter constant Cx [n] and a second filter constant Dx [n]. Note that the first filter constant Cx [n] and the second filter constant Dx [n] are both integers and different values. For example, the first filter constant Cx [n] is larger by “1” than the second filter constant Dx [n] (Cx [n] ≠ Dx [n] = Cx [n] −1).

  As shown in FIG. 21, the double moving average processing unit 210 includes a plurality of, for example, Cmax first registers 214, 214,... Arranged in a row, and a plurality of, for example, Dmax arranged in a row. The second registers 216, 216,... The total number Cmax of the first registers 214, 214,... Is the same as the total number Cmax of the registers 204, 204,. The total number of registers 214, 214,... Is slightly smaller than Cmax.

  The raw load data Wy [n] input to the double moving average processing unit 210 is first stored in the first register 214 (at the left end in FIG. 21) of the first registers 214, 214,. The The raw load data Wy [n] stored in the first first register 214 is sampled by the A / D conversion circuit 184 every time new raw load data Wy [n] is input. In accordance with the timing (that is, at the same time interval as the sampling period Tc), the first register 214 is sequentially moved (shifted) to the right. Then, the raw load data Wy [n] moved to the last (Cmax-th) register 214 at the right end in FIG. 21 is sent out at the next sampling timing by the A / D conversion circuit 184 and discarded.

  The raw load data Wy [n] stored in the first registers 214, 214,... Are data having consecutive sampling times, and the first filter constant Cx [n] while shifting the sampling times one by one. Is averaged by the number according to The plurality of average values Wy [n] ″ having different sampling times (average times) are stored in the individual second registers 216, 216,..., For example. In other words, the average value Wy [n] of the raw load data Wy [n] stored in the first and subsequent consecutive Cx [n] first registers 214, 214,. ] "(Wy [n, k]") is obtained, and the average value Wy [n] "is stored in the second register 216 (at the left end in FIG. 21). Further, the average value Wy [n] ”(Wy [n, k + 1]” of the raw load data Wy [n] stored in the second and subsequent consecutive Cx [n] first registers 214, 214,. The average value Wy [n] ″ is stored in the second second register 216. Hereinafter, the raw load data Wy stored in the third and subsequent first registers 214, 214,. Similarly, [n] is averaged by Cx [n], and each average value Wy [n] ″ is stored in the individual second registers 216, 216,.

  Then, among the average values Wy [n] ”stored in the second registers 216, 216,..., New ones with a new sampling time are extracted according to the second filter constant Dx [n]. The average value Wy [n] ″ stored in the registers 216, 216,. Then, further average values of the extracted Dx [n] average values Wy [n] ″ are obtained, and this further average value is output as processed load data Wy [n] ′. Wy [n] ′ is also input to the constant setting unit 212.

  The constant setting unit 212 uses the same procedure as the constant setting unit 202 in FIG. 6 described above to obtain the filter constant Cx [n] based on the input processed load data Wy [n] ′. Cx [n] is obtained. Further, a value smaller than the first filter constant Cx [n] by “1” is set as a second filter constant Dx [n], and the first filter constant Cx [n] and the second filter constant Dx [n] are set as the second filter constant Cx [n]. Set in the double moving average processing unit 210. That is, the filter constants Cx [n] and Dx [n] are updated. These filter coefficients Cx [n] and Dx [n] are updated every time new processed load data Wy [n] ′ is output from the double moving average processing unit 210, in other words, A / D conversion. It is executed whenever the weighing signal Wy [n] is sampled by the circuit 184.

  By setting the two filter constants Cx [n] and Dx [n] having values close to each other (differing by “1”) in this way, the double moving average processing unit 200 has the natural vibration frequency fx [n]. The notch frequency is formed not only in the same frequency as that in FIG. 5 but also in the surrounding frequency, that is, an attenuation region is formed in a wider frequency band including the natural vibration frequency fx [n]. As a result, the natural vibration component included in the raw load data Wy [n] is more reliably removed, and higher measurement accuracy can be obtained.

  Note that the configuration of the adaptive filter 188 is not limited to that described in this embodiment. That is, instead of the moving average processing unit 200 (or the double moving average processing unit 210) serving as a filter unit, a filter having another configuration such as an FIR (Finite-duration Impulse Response) type or an IIR (Infinite Impulse Response) type may be used. You may employ | adopt what has.

  The digital recorder 44 is used to obtain the natural vibration frequency fz [n] when no load is applied to each unit 14 and the natural vibration frequency fm [n] when the test load Wm is applied. Not exclusively. For example, by analyzing the raw load data Wy [n] of each unit 14 in the frequency domain using a spectrum analyzer (FFT (Fast Fourier Transform) processing), the natural vibration frequencies fz [n] and fm [n ] May be obtained. Further, a function for automatically obtaining these natural vibration frequencies fz [n] and fm [n] may be provided in the controller 36 or each of the units 14, 14,. Further, in place of the natural vibration frequency fz [n] at no load, a natural vibration frequency fm [n] ′ when a test load Wm ′ (≠ Wm) different from the test load Wm is applied is obtained. Even if the spring constant h [n] and the initial load Wi [n] are obtained based on the natural vibration frequency fm [n] ′ and the natural vibration frequency fm [n] ′ when the test load Wm is applied. Good.

  Further, a plurality of supply amount switching timings such as the time point t4 described above may be provided. That is, the present invention can be applied even when the beverage supply amount is switched in a plurality of stages. Further, the present invention can be applied to a so-called different kind of filling device in which a plurality of kinds of objects to be weighed are filled with a predetermined weight.

  And this invention is applicable not only to the weight-type filling apparatus 10 but to other dynamic balances, such as a weighing conveyor, a combination scale, and a weight-type sorter. Needless to say, the present invention can be applied not only to a dynamic balance but also to a static balance.

It is a schematic block diagram of one Embodiment of this invention. It is an illustration figure for demonstrating the operation | movement of the embodiment. It is a block diagram which shows the structure of the controller in the same embodiment. It is a block diagram which shows the structure of each unit in the embodiment. It is an illustration figure for demonstrating operation | movement of the unit. It is an illustration figure which shows notionally the structure of the adaptive filter implement | achieved by CPU of the unit. It is an illustration figure which shows notionally the structure of the moving average process part in FIG. It is an illustration figure which shows notionally the operation | movement of the same moving average process part. It is a flowchart which shows the content of the adjustment mode task which CPU of the controller in the embodiment performs. It is a flowchart following FIG. It is a flowchart following FIG. It is a flowchart which shows the content of the pair adjustment mode task which CPU of each unit in the embodiment performs. It is a flowchart which shows the content of the operation mode task which CPU of the controller in the embodiment performs. It is a flowchart which shows the content of the timing control task which CPU of the controller performs. It is a flowchart which shows the content of the data acquisition task which CPU of the controller performs. It is a flowchart which shows the content of the operation mode task which CPU of each unit in the embodiment performs. It is a flowchart which shows the content of the filter constant update task which CPU of the unit performs. It is a flowchart which shows the content of the measurement task which CPU of the unit performs. It is a flowchart following FIG. It is a block diagram of the adaptive filter which shows an example different from FIG. It is an illustration figure which shows notionally the structure of the moving average process part in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Weight type filling device 14 Unit 16 Mounting stand 18 Weighing machine 180 Load cell 184 A / D conversion circuit 186 CPU
188 Adaptive filter 200 Moving average processing unit 202 Constant setting unit

Claims (5)

Sampling means for sampling a weighing signal obtained from the weighing means at a period shorter than the period of the natural vibration of the weighing means;
Filter means for subjecting the discrete metric signal sampled by the sampling means to filter processing according to a certain filter constant;
Based on the processed metric signal after processing by the filter means, the filter constant is such that the frequency characteristic of the filter means becomes a characteristic of attenuating the frequency band component including the natural vibration frequency of the discrete metric signal. Updating means for updating
The weight measuring apparatus, wherein the updating means updates the filter constant every time sampling is performed by the sampling means.   The weight measuring apparatus according to claim 1, wherein a sampling period by the sampling unit is shorter than a half of a period of the natural vibration. The filtering process is a moving average process,
The weight measuring device according to claim 1 or 2, wherein the filter constant is the number of the discrete weighing signals to be subjected to the moving average process.   The updating means is applied with a known second weight different from the frequency of the natural vibration when the known first weight is applied to the object placing portion of the weighing means and the first weight. The weight measuring device according to claim 1, wherein the filter constant is updated based on an arithmetic expression including the frequency of the natural vibration at the time. A sampling process for sampling a measurement signal obtained from the measurement means at a period shorter than the period of the natural vibration of the measurement means;
A filtering process for applying a filtering process according to a certain filter constant to the discrete metric signal after sampling by the sampling process;
Based on the processed metric signal after processing by the filtering process, the frequency characteristic in the filtering process is such that the frequency band component including the natural vibration frequency of the discrete metric signal is attenuated. An update process for updating the filter constant,
A weight measurement method in which, in the update process, the filter constant is updated each time sampling is performed in the sampling process.

Images