JP5669551B2 - Load cell failure diagnosis device - Google Patents

Description

  The present invention relates to a load cell failure diagnosis device that detects a span abnormality occurring in a strain gauge type load cell used in a measuring instrument.

  As an apparatus for detecting a span abnormality occurring in one load cell, for example, there is one disclosed in Patent Document 1.

JP-A-5-172661

  In the failure diagnosis device disclosed in Patent Document 1, the zero point adjustment device is manually operated by determining that the weighing instrument is in a specific state, for example, that the operator has no object to be weighed on the weighing table. Or, in the case of automatic, the zero sensor is detected by the zero sensor automatic detection means by detecting the presence or absence of the object to be weighed on the weighing platform by the article sensor, and the two bridges constituting the full bridge circuit of the load sensor are configured. Each output of the half bridge circuit is compared with a predetermined allowable value based on the output of the zero point, and if the result of the comparison is that the output of any half bridge circuit exceeds the allowable value, a failure notification is made. ing.

  However, the above-described conventional failure diagnosis apparatus has the following problems regarding determination of span abnormality.

(1) Since the outputs of the half-bridge circuits are compared in a state where a specific force or load is applied, it is not possible to detect a span abnormality during normal measurement work.
(2) When the zero point becomes abnormal, a span abnormality often occurs at the same time. However, in the transition period of temperature change, the zero points of the two half-bridge circuits are different even if the load cell is in the normal range. If a tolerance value is set to warn of a zero-point abnormality, and a small tolerance value is set, a zero-point abnormality is frequently output as a warning and is difficult to use.
(3) Even if the fluctuation of the zero is somewhat large, the operator can adjust it with the zero adjustment switch during normal weighing work. Therefore, it is not appropriate to give an abnormal alarm to a very small fluctuation of the zero. On the other hand, an abnormality in the span cannot be adjusted easily by the operator like the fluctuation of the zero point, and even if the fluctuation amount is small, the ratio of the fluctuation amount to the rated load is the measurement accuracy of the specification. Therefore, it is necessary to accurately determine the abnormality and give an alarm. However, when zero fluctuation and span fluctuation occur, there is no description about extracting only the fluctuation of the span and giving an alarm in a normal weighing operation.
(4) Each of the two half-bridge circuits has different output magnitudes for the same load, and if the spans are different, the difference cannot be accurately determined even if both outputs are compared. Cannot be detected.
(5) Some of the two half-bridge circuits have different hysteresis error characteristics. Even if the span is adjusted with a specific load, a large error may occur even if the span is not abnormal under any load. Yes, an accurate span abnormality cannot be detected for any load.

  The present invention has been made in view of the above-described problems, and eliminates zero fluctuations individually at the outputs of the two half-bridge circuits of the load cell used in the weighing instrument, and further eliminates hysteresis errors individually. It is an object of the present invention to provide a load cell failure diagnosis apparatus that can accurately detect only the span variation by correcting to.

In order to achieve the above object, a load cell failure diagnosis apparatus according to the first invention comprises:
In a load cell failure diagnosis device using output signals from two terminals in a Wheatstone bridge circuit composed of a plurality of strain gauges affixed to a strain generating portion as load signals,
Two load signal output means for performing calculation processing so that the load signals output from the two terminals become load signals of the same magnitude with respect to the load applied to the load cell, and the two terminals In the load signal output from, zero point adjustment means for independently removing zero point components that fluctuate during normal weighing work is provided.

  The second invention is the load signal display according to the first invention, wherein the two load signals output from the two load signal output means are individually displayed or the absolute value of the difference between the two load signals is displayed. Means are provided.

According to a third invention, in the first invention, the absolute value of the difference between the two load signals output from the two load signal output means is zero or a value near zero, or a value near zero or near zero. It is characterized by comprising zero point adjustment state display means for displaying a state that is not.

The fourth invention is Oite the third invention, a hysteresis error correcting means for individually correcting the hysteresis error caused in accordance with the increase or decrease of two load signals output from the two load signal output means It is characterized by comprising.

Further, a fifth invention comprises the monotonic increase determining means in the fourth invention, wherein the hysteresis error correcting means determines that each of the two load signals is a monotone increase starting from the vicinity of the zero point, The hysteresis error correction means is configured to correct the hysteresis error when it is determined by the monotone increase determination means that the two load signals are monotonically increasing starting from the vicinity of the zero point.

The load cell failure diagnosis apparatus according to the sixth invention is:
In a load cell failure diagnosis device using output signals from two terminals in a Wheatstone bridge circuit composed of a plurality of strain gauges affixed to a strain generating portion as load signals,
Two load signal output means for performing arithmetic processing so that the load signals output from the two terminals become load signals of the same magnitude with respect to the load applied to the load cell;
A stability determining means for determining the stability of the two load signals output from the two load signal output means, a latest stable weight value generated based on the stability determination by the stability determining means,
A load change amount detecting means for detecting a difference from the stable weight value generated at the timing immediately before the latest stable weight value;
Span abnormality determining means for determining a load cell span abnormality based on the difference between the latest stable weight value detected by the load change amount detecting means and the previous stable weight value. Is.

According to a seventh invention, in the sixth invention, the span abnormality determining means is configured such that the latest stable load value generated from one of the two load signal output means and the previous stable load value. The difference between the weight value and the difference between the latest stable load value generated from the other load signal output means and the previous stable weight value is compared. The load cell span is determined to be abnormal.

According to an eighth aspect of the present invention, in the sixth or seventh aspect of the present invention, hysteresis error correction that individually corrects a hysteresis error that occurs in response to an increase or decrease in the two load signals output from the two load signal output means. Means are provided.

According to a ninth invention, in the eighth invention, the hysteresis error correction means further comprises a monotone increase determination means for determining that each of the two load signals is a monotone increase starting from the vicinity of the zero point. The hysteresis error correcting means is configured to correct the hysteresis error when it is determined by the monotone increase determining means that the two load signals are monotonically increasing starting from the vicinity of the zero point.

According to the present invention, there is a different zero point drift for each individual half bridge by allowing the operator to individually adjust the zero point for each half bridge or when the operator forgets to operate the zero adjustment switch. However, it is possible to accurately detect only the span fluctuation by removing the zero fluctuation from the output of each half bridge, and the span abnormality can be accurately determined. Moreover, this span abnormality determination can be carried out while continuing normal measurement work without requiring any special work.
In the present invention, if the hysteresis error generated according to the increase / decrease of the two load signals is corrected individually and accurately, the magnitude of the output appearing according to the increase / decrease of the load load in the two half bridges may be different. A span abnormality can be detected more accurately.

FIG. 2 is a use state diagram of a load cell to which a load cell failure diagnosis apparatus according to an embodiment of the present invention is applied, a state diagram used for a compression load cell (a) and a state diagram used for a dual beam load cell (b). Schematic system configuration diagram of a load cell failure diagnosis apparatus according to an embodiment of the present invention Functional block diagram of the central processing unit Load cell full bridge circuit output voltage diagram Hysteresis characteristics diagram of load cell half-bridge output (a) and diagram showing half-bridge output error (b) Flow chart showing the procedure of span abnormality determination calculation by each half bridge output Schematic system configuration diagram of a load cell failure diagnosis apparatus according to another embodiment of the present invention

  Next, a specific embodiment of a load cell failure diagnosis apparatus according to the present invention will be described with reference to the drawings.

  FIG. 1 is a use state diagram of a load cell to which a load cell failure diagnosis apparatus according to an embodiment of the present invention is applied, a state diagram (a) used for a compression type load cell and a state used for a dual beam type load cell. Each figure (b) is shown. FIG. 2 is a schematic system configuration diagram of the load cell failure diagnosis apparatus according to the present embodiment.

  A compression type load cell 1 shown in FIG. 1A includes a support portion 2, a strain generating portion 3 provided on the support portion 2, and a force receiving portion 4 provided on an upper portion of the strain generating portion 3. Thus, the strain generating portion 3 is configured to be compressed when the impact receiving portion 4 receives a force. Strain gauges 11 a and 13 a are attached to the strain generating portion 3 in parallel with the axial direction of the strain generating portion 3, and strain gauges 12 a and 14 a are attached to the strain generating portion 3 at right angles to the axial direction of the strain generating portion 3. The strain gauges 11a and 13a detect compressive force, and the strain gauges 12a and 14a detect tensile force.

  In the double beam type load cell 5 shown in FIG. 1B, a strain generating portion 8 is formed so as to constitute two beams (beams) 6 and 7. Strain gauges 11b and 12b are attached to the beam 6, and strain gauges 13b and 14b are attached to the beam 7 along the longitudinal direction of the beam. When the object to be weighed 10 is placed on the weighing table 9 coupled to the strain generating part 8, a load corresponding to the weight of the object to be weighed 10 acts on the strain generating part 8, and the strain gauges 12b and 14b are bonded to the gauge. The strain gauges 11b and 13b receive a bending stress in a direction in which the gauge bonding surface contracts.

  In the following description, the strain gauges 11a and 11b are collectively referred to as the strain gauge 11, the strain gauges 12a and 12b are collectively referred to as the strain gauge 12, and the strain gauges 13a and 13b are collectively referred to as the strain gauge 13. The strain gauges 14a and 14b are collectively referred to as a strain gauge 14.

  As shown in FIG. 2, the strain gauges 11, 12, 13, and 14 are connected to each other so as to constitute a full bridge circuit 15. Here, in the full bridge circuit 15, if both the strain gauges 12 and 14 constituting the opposite sides are tensile forces, the tensile force is obtained, and both the strain gauges 11 and 13 constituting the opposite sides are both compressive forces. Then, it is wired to receive the same type of force as compression force.

  In the full bridge circuit 15, a DC voltage for excitation is applied to two opposing connection points 16 and 17, and force or load is detected from the connection points 18 and 19 positioned at right angles to these connection points 16 and 17. The voltage is taken out.

  A fault diagnosis device 20 is provided for the above-described full bridge circuit 15. The fault diagnosis apparatus 20 includes two voltage reference fixed resistors 21 and 22, an analog adder circuit 23, and two analog-digital converters (hereinafter referred to as “A / D converters”) 24, 25 and an arithmetic circuit 26. Here, the fixed resistors 21 and 22 are connected in series to each other and are connected to connection points 16 and 17 of the full bridge circuit 15. The fixed resistors 21 and 22 and the strain gauges 12 and 13 form a half bridge circuit 15a, and the fixed resistors 21 and 22 and the strain gauges 11 and 14 form a half bridge circuit 15b.

The analog adder circuit 23 includes a first operational amplifier 31, a second operational amplifier 32, a third operational amplifier 33, and a fourth operational amplifier 34.
In the first operational amplifier 31, the input positive terminal 31 a is connected to the connection point 18 of the full bridge circuit 15, the input negative terminal 31 b is connected to the output terminal 31 c, and the output terminal 31 c is connected to the resistor 40.
In the second operational amplifier 32, the input positive terminal 32a is connected to the connection point 41 of the two fixed resistors 21 and 22, the input negative terminal 32b is connected to the output terminal 32c, and the output terminal 32c is connected to the resistors 42 and 43. It is connected.
In the third operational amplifier 33, the input positive terminal 33a is connected to the circuit ground 44, the input negative terminal 33b is connected to the resistors 40 and 42, and is connected to the output terminal 33c via the resistor 45, The output terminal 33c is connected to the A / D converter 24.
In the fourth operational amplifier 34, the input positive terminal 34 a is connected to the connection point 19 of the full bridge circuit 15, and the input negative terminal 34 b is connected to the resistor 43 and connected to the output terminal 34 c via the resistor 46. The output terminal 34 c is connected to the A / D converter 25.

  The A / D converters 24 and 25 convert the analog load signal from the analog adder circuit 23 into a digital load signal. Digital load signals from the A / D converters 24 and 25 are output to the arithmetic circuit 26.

The arithmetic circuit 26 includes an input / output circuit (I / O) 51, a central processing unit (CPU) 52, and a memory block (MEM) 53.
In the arithmetic circuit 26, output signals of the A / D converters 24 and 25 are read from the input / output circuit 51 into the memory block 53 via the central processing unit 52.
The memory block 53 is composed of a storage element (semiconductor element) such as a RAM that primarily stores data for input, output, and computation, an EEPROM that continuously stores setting data, and a PROM that continuously stores predetermined programs.
Calculation formulas such as load signals Wan, Wbn and hysteresis errors Ea1, Eb1, which will be described later, are stored in the PROM, and an allowable value Rh for determining a span abnormality, which will be described later, and an allowable value Wz of zero fluctuation are set in the EEPROM. .
A display device (DIS) 54, a key switch (KEY) 55, an alarm device (ALARM) 56, and the like are connected to the arithmetic circuit 26, and values of Wan and Wbn, which will be described later, are displayed on the display device 54, and data Operations such as setting and zero adjustment are performed by the key switch 55, and an alarm for notifying a failure is issued from the alarm device 56.

  In the central processing unit 52, by executing a predetermined program stored in the memory block 53, a filter processing unit 52a, a load signal calculation unit 52b, a stability determination unit 52c, a load, as shown in FIG. Functions of the change amount calculation unit 52d, the hysteresis error correction unit 52e, the span abnormality determination unit 52f, and the like are realized.

The filter processing unit 52a performs a predetermined filtering process on the digital load signals from the A / D converters 24 and 25.
The load signal calculation unit 52b performs arithmetic processing such that each digital load signal from the A / D converter 24 and the A / D converter 25 becomes an output signal having the same magnitude with respect to the load having the same magnitude. Execute.
The stability determination unit 52c receives the digital load signal from the A / D converter 24 when a force or load of an arbitrary magnitude is applied to the compression load cell 1 or the double beam load cell 5. The digital load signal from the A / D converter 25 determines whether or not the load is stable depending on whether or not it exceeds an allowable value determined based on a predetermined weighing accuracy or a predetermined rated load.
The load change amount calculation unit 52d detects a difference between the latest stable weight value obtained by the stability determination unit 52c and the stable weight value generated at the timing immediately before the latest stable weight value.
Based on the digital load signal from the A / D converter 24 and the digital load signal from the A / D converter 25 obtained through the arithmetic processing by the load signal calculation unit 52b, the hysteresis error correction unit 52e The hysteresis error that occurs according to the increase / decrease is individually corrected.
Based on the digital load signal from the A / D converter 24 and the digital load signal from the A / D converter 25 obtained through the arithmetic processing by the load signal calculation unit 52b, the span abnormality determination unit 52f It is determined whether or not is abnormal.

  In the failure diagnosis apparatus 20 having the configuration described above, the analog load signals eoa and eob output from the two sets of half-bridge circuits 15a and 15b are converted into analog signals via an analog adder circuit 23 as shown in FIG. The load signals are eoa ′ and eob ′. These analog load signals eoa ′ and eob ′ are converted into digital load signals Wa and Wb by A / D converters 24 and 25, respectively. These digital load signals Wa and Wb are taken into the arithmetic circuit 26, and as shown in FIG. 3, a predetermined filtering process is performed in the filter processing unit 52a to obtain digital load signals Wax and Wbx.

  Next, with respect to the contents of the arithmetic processing based on the digital load signals Wax and Wbx in the arithmetic circuit 26, an example of a measuring instrument (see FIG. 1 (b)) in which the measuring platform 9 is mounted on the double beam type load cell 5 is used. It will be described below.

  At the time of adjustment of the weighing device, when the power is applied to the double beam load cell 5 without placing the object 10 on the weighing table 9 (which may be a weighing container such as a weighing hopper), two sets of half bridge circuits 15a, From 15b, analog load signals eoa and eob to which an unbalance component of the bridge resistance and a tare load are added are output. These analog load signals eoa and eob are converted into digital load signals Wax and Wbx via the analog adder circuit 23, A / D converters 24 and 25, and the filter processing unit 52a. Assume that Wax = Wai and Wbx = Wbi are obtained as the digital load signals Wax and Wbx at this time. The values of Wai and Wbi are stored in the initial load memory of the memory block 53 as an initial load by operating the initial load storage key switch in the key switch 55.

Now, the following formulas (1) and (2) are determined as calculation formulas for the output load signal Wan on the half bridge circuit 15a side and the output load signal Wbn on the half bridge circuit 15b side, respectively.
Wan = ka. (Wax-Wai) -Wza (1)
Wbn = kb · (Wbx−Wbi) −Wzb (2)
Here, ka and kb are span coefficients, and Wza and Wzb are zero point load memories.

The above formulas (1) and (2) are obtained by independently generating span coefficients ka and kb, initial load memories Wai and Wbi, and zero load memories Wza and Wzb for the output load signals Wan and Wbn of both half bridge circuits 15a and 15b. And Wan and Wbn are set to the same value for the same load, and the zero adjustment function can be implemented independently. By defining these equations, for example, in the transition period of temperature change, the zero point fluctuation that is not common to both half-bridge circuits occurred because the heat transfer to the strain-generating metal to which the strain gauge is attached is not uniform. Even in this case, it is possible to avoid the inconvenience that the zero point fluctuation is included in the output load signals Wan and Wbn as they are and appears as an output difference.
Note that when the initial load values Wai and Wbi are stored in the initial load memory, Wax = Wai and Wbx = Wbi, so that Wan = Wbn = 0 for any numerical value of the span coefficients ka and kb which is not zero.

The span coefficients ka and kb are adjusted so that the value of Wan and the value of Wbn each represent the value of Ws when the rated load Ws (the value of Ws is known) is set, that is, the following equation is satisfied. .
Wan = Wbn = Ws (3)
Thus, the weight measurement value of the object to be weighed is calculated and displayed by the following equation.
Wn = 1/2 · (Wan + Wbn) (4)
Note that Wan, Wbn, and Wn are set so as to have at least four times the resolution of the display value as the weight measurement value of the object to be weighed in the weighing machine.

Next, the response to the zero point fluctuation of each half bridge output will be described.
(1) When zero variation is performed by manual adjustment means For the span abnormality during the weighing operation, the allowable value Rh for the span variation rate is set in advance corresponding to the measurement accuracy of the specification, and the span is Is determined to be abnormal.
| Wan−Wbn | / {(Wan + Wbn) / 2}> Rh (5)

If the zero adjustment switch provided on the measuring instrument is manually operated, the two output load signals Wan and Wbn are individually adjusted to zero at the same time. That is, the value of Wan is added to Wza, and as a result, Wan = 0 is adjusted. The value of Wab is added to Wzb, and the result is adjusted to Wbn = 0.
However, if, for example, Wan and Wbn vary from the zero point in the positive and negative directions by the same amount or substantially the same amount, if the load signal is displayed by the above equation (4) which is a conventional display method, both zero variation Since the amount is added to a large value, it is erroneously determined that the span is abnormal. Therefore, in order to accurately determine the span abnormality by relying on the zero adjustment operation by the operator's zero adjustment switch, the value of | Wan−Wbn | which is the absolute value of the difference is displayed as the zero evaluation. Alternatively, the output load signals Wan and Wbn of the two half bridge circuits 15a and 15b may be displayed, respectively.

  If there is no object to be weighed on the weighing platform and this value | Wan-Wbn | is not zero, when the object to be weighed is placed, either of Wan and Wbn will not have any span fluctuation. Since this is evaluated as a span variation of the half-bridge circuits 15a and 15b, this value | Wan−Wbn | is set as an index value that prompts the operator to press the zero adjustment switch.

Alternatively, wz is determined as an allowable value of zero point fluctuation,
| Wan-Wbn |> wz
If is established, a sign indicating that the zero point has not been correctly adjusted (zero point fluctuation sign) is displayed by a display device such as a lamp, and the operator is prompted to press the zero point adjustment switch.
On the contrary, it is possible to display on the display unit that the above inequality is not satisfied, and to make the operator judge that it is not necessary to press the zero adjustment switch.

(2) When zero-point fluctuation is performed by automatic adjustment means In order to detect a span abnormality automatically and accurately, the fluctuation amount due to the load load with respect to the output load signals Wan and Wbn of the half bridge circuits 15a and 15b is as follows. A load change amount detecting means for calculating only the load is provided.
That is, Wan (Wbn) is generated at a time interval Ta (= several 100 msec) by a load signal read from the A / D converter 24 (25), and M shift registers are prepared. Wan (Wbn) is discarded in order of generation, the oldest value is discarded, and the latest M Wan (Wbn) are always stored. Further, an allowable value of the stability limit is set in advance, and it is determined that the value is stable if the value of M Wan (Wbn) is within the allowable value. This stability determination is sequentially performed at time intervals Ta. Thus, by calculating the average value of the values stored in the shift register and setting it as Wan (Wbn) for the weight measurement value, the stable weight value of the object to be weighed can be obtained at every time interval Ta at the shortest.

  In this case, since the object to be weighed is placed on the weighing platform, the value of the shift register deviates from the allowable value due to the value of Wan (Wbn) read at a certain timing, and the stability condition is not satisfied. Since the latest value that entered the shift register at the timing just before deviating may be the limit value of the allowable range, the most recent value was entered to calculate the value of Wan (Wbn) as the stable weight value. It is appropriate to exclude the value and obtain the average value with M−1 values.

  For this purpose, one storage memory for storing the stable weight value immediately before the value of Wan (Wbn) becomes unstable is prepared and continuously generated when there is no object to be weighed on the weighing table. The stable weight value is stored in the storage memory and the content is updated. However, at the time of updating, the storage data (= the previous stable weight value) is called from the storage memory to another memory and then updated to a new stable weight value.

With this configuration, when an object to be weighed is placed on the weighing platform, the stability determination condition is not satisfied and the storage memory is not updated while the load signal vibrates due to the placement. Is in a state in which the last data that satisfies the stability condition immediately before placing the object to be weighed is stored. When the vibration of the load signal converges, the stability determination condition is satisfied, and the latest stable weight value is obtained. At this time, the data called from the storage memory is the stable weight value at the previous timing, and the latest stable weight value is the stable weight value at the current timing.
During this time, the time difference required for the object to be weighed to be placed and convergence of the vibration of the load signal is short, so the load signal does not drift during this time, or even if the zero signal drifts, it is extremely small. Therefore, the difference between these stable weight values does not include the zero-point drift and accurately represents the amount of change corresponding to the weight of the object to be weighed. Therefore, the zero-point fluctuation component is excluded and used for accurate evaluation of only the span fluctuation. Suitable for

  In the above description, the case where the object to be weighed is placed from the state where there is no object to be weighed on the weighing table has been described. The same can be done for the placement. In this case, the difference in the stable weight value corresponds to the weight of the newly placed object. Therefore, even a weight measuring device where the object to be weighed stays on the weighing table for a long time and cumulatively weighs on the weighing table can accurately diagnose the span abnormality without being affected by the zero point fluctuation. it can.

In such a case, assuming that the output load signal Wan in the stable state of the previous timing and the current timing is Wanu and Wanv, respectively, the change amount Wamp due to the load load of the half bridge output is expressed by the following equation. calculate.
Wamp = | Wanv-Wan | (6)
Then, the amount of change Wamp is compared with a predetermined value Wh, and the following expression Wamp> Wh (7)
If is established, a flag is set during the calculation process assuming that the load change amount is sufficient to determine the span abnormality. If this change amount Wamp is small, the S / N ratio with noise becomes small, and accurate determination cannot be made. When the change amount Wamp is small and the above equation (7) is not satisfied, the flag is reset.

Similarly, when the flags based on the calculation for the output load signal Wbn are gathered, the abnormality of the span is determined based on the change amounts Wamp and Wbnp due to the load load of the half bridge output at the time when the flags are set together. The span abnormality determination is carried out by the following equation as in the above equation (5).
| Wamp−Wbnp | / {(Wamp + Wbnp) / 2}> Rh (8)

  In this way, even if the operator forgets to operate the zero adjustment switch or there is a different zero drift for each individual half bridge, the zero fluctuation value in the weight measurement values Wan and Wbn is offset, and an accurate span error Can be determined.

  Next, a method for correcting a hysteresis error between two half-bridge outputs and more accurately and easily determining a span abnormality will be described.

  The output load signals Wan and Wbn of the two sets of half-bridge circuits 15a and 15b have different hysteresis characteristics even if the span coefficient is adjusted so that the output is the same with respect to the load of the rated load. As shown in FIG. 5A, the output when the load load is gradually increased (monotonically increased) and the output when the load load is gradually decreased (monotonically decreased) are different values with respect to the intermediate load load. In FIG. 5A, a1 indicates an output locus when the output load signal Wan is gradually increased, a2 indicates an output locus when the output load signal Wan is gradually decreased, and b1 indicates a load load of the output load signal Wbn. An output locus at the time of gradual increase, b2 indicates an output locus at the time of gradual decrease of the load of the output load signal Wbn.

  Even if there is no span variation in the half bridge, the hysteresis error characteristic causes a difference in the value of Wan-Wbn in accordance with the change in load. Therefore, the accuracy of the span abnormality determination is impaired by this difference. . The amount of hysteresis varies depending on the state of application to the strain cell load cell strained portion of the strain gauge (for example, the thickness of the adhesive), and therefore the error characteristics differ for each half bridge. For this reason, in order to accurately compare the outputs of the half bridges, it is necessary to individually correct the hysteresis error.

  This hysteresis error changes in a complex manner when the load increases or decreases in various magnitudes between the zero point and the rated load, so it cannot be corrected easily and accurately, and accuracy may be impaired. is there. Here, instead of correcting the measured value of the object to be weighed, it is only necessary to accurately detect an abnormality in the span, so the output of the two half bridges is individually corrected for the hysteresis error in the increase or decrease of the load load. However, it is preferable to employ a simplified hysteresis correction method that is convenient for abnormality determination.

  In the process where the load increases monotonously toward the rated load or decreases monotonously from the rated load toward zero, the hysteresis error that occurs changes almost uniformly and continuously, so the error corresponding to the load Is relatively easy to estimate. In addition, due to the nature of weighing work, the process of monotonically increasing the load load up to the rated load, that is, when placing one object to be weighed on the weighing table, and when stacking another object to be weighed on it It is preferable to carry out an operation for determining a span abnormality each time an object to be weighed is placed.

In consideration of these points, in this embodiment, hysteresis correction is performed as follows. That is, in the adjustment mode, several types of known load loads having different sizes from 0 to the weighing platform are applied to the rated load Ws while gradually increasing (monotonically increasing), and then the sizes are different after being applied to Ws. Multiply several known load loads to obtain hysteresis error data for the two half-bridge outputs for a monotonically increasing change in load. Based on this error data, the relationship between the load changes Wan and Wbn and the hysteresis errors Ea1 and Eb1 when the load is monotonously increased by a method such as the least square method is expressed by a function of a quadratic expression (error correction function). ) And expressed as Then, this equation is stored in the memory block 53 of the arithmetic circuit 26.
Ea1 = fa (Wan)
Eb1 = fb (Wbn) (9)

  The errors Ea1 and Eb1 shown in FIG. 5 (b) appear only when the load load increases monotonously from 0 to the rated load Ws. For example, the errors Ea1 and Eb1 once decrease after the load load increases to Ws / 2. Even if it increases toward the rated load Ws again, the error does not appear as shown in FIG.

  Thus, the error functions fa (Wan) and fb (Wbn) determined in the adjustment mode are used, and the span abnormality is detected only in the process in which the load is monotonously increased from near zero or less than zero.

In short, it is assumed that the following conditions are satisfied when performing span abnormality determination that compensates for the hysteresis error in the monotonically increasing process.
(1) Both sets of output load signals Wan and Wbn satisfy the stability condition.
(2) Both sets of output load signals Wan and Wbn are in a monotonically increasing process starting from the zero point. For this reason, if there is a decrease of a predetermined value or more once in the process of increasing the output load signal, both Wan and Wbn are not applied to the abnormality determination unless they return to near zero. The reason for this is that if there is a load decrease of a predetermined value abnormal during the increase process, an error that is completely different from the function set based on the monotonic increase occurs, and the load that increases monotonically from the zero point is corrected. Shall.
(3) Since the determination is based on the difference between the change amount of the previous stored load output and the current load output, the difference between the current measured value and the previous stored value is larger than a predetermined value for both Wan and Wbn (for example, the rated load Ws An abnormality of the span is determined with a change amount equal to or more than 1/4.
Here, assuming that the corrected outputs are Wan ′ and Wbn ′, error correction is expressed by the following equation using equation (9).
Wan '= Wan-fa (Wan)
Wbn ′ = Wbn−fb (Wbn) (10)

  Next, the procedure of span abnormality determination calculation by each half bridge output when the load load is monotonously increased from 0 toward the rated load Ws of the measuring instrument will be described with reference to the flowchart shown in FIG. In the flowchart, S1 to S16 indicate steps.

S1 to S2: Read the latest M Wan and Wbn, and store the respective values in the M shift registers. Then, an allowable value of the stability limit is set in advance, and the stability determination is performed depending on whether or not the values of M Wan and Wbn are within the allowable value.
S3 to S7: When the stability condition is satisfied, the stable weight values for Wan and Wbn are calculated. If the stability condition is not satisfied, the process returns.
The monotonically increasing process of the load change starts when both outputs Wan and Wbn are near zero or below zero. Therefore, the near zero outputs Wan and Wbn of this time are stored in the memory. However, if Wan and Wbn are negative values, they are replaced with 0, and 0 is stored. When the values of Wan and Wbn are less than or equal to zero, a flag Fz indicating that the process is in the monotonic increase process is set to 1, and the monotonous increase start condition is satisfied.

S8: At the next read timing of Wan and Wbn, if both outputs Wan and Wbn are less than or equal to zero again, the above operation is repeated, but in the determination of step S6, one or both of Wan and Wbn are If the output exceeds near zero or less, whether or not the amount of change of the current read load from the previous read load exceeds a predetermined value Wq (for example, Wq = Ws / 4), in other words, It is determined whether or not the formula is satisfied.
Wan-Wan> Wq
Wbn-Wbnu> Wq (11)
However, Wan and Wbnu are the stable weight values immediately before Wan and Wbn, respectively.

S9 to S13: If the equation (11) is established, it is determined in step S9 whether or not both outputs are in the process of monotonic increase (whether Fz = 1). If it is in the process (Fz = 1), the previous stored values Wanu and Wbnu are called from the memory in the processing after step S10, and the error is corrected based on the equation (10) to be Wanv and Wbnv. Then, in step S13, the amount of change in the two half-bridge outputs due to the current and previous correction outputs.
Da = Wanv-Wanu
Db = Wbnv−Wbnu (12)
Therefore, the following formula: | Da−Db | / {(Da + Db) / 2}> Rab (13)
However, Rab determines whether or not a preset allowable value of the span variation rate is satisfied. And if Formula (13) is materialized, it will warn that it is span abnormality. Here, the value of Rab is set to Rab = E, for example, if the weighing accuracy of the weighing instrument is E on the specification.

  If there is a course in which the load load decreases in the middle, then there is a process in which the load increases again, and even if equation (11) is satisfied and step S9 is reached, the condition of the monotonous increase process is broken and the flag Fz is 0. Since it has been reset to, abnormality determination is not performed. That is, the abnormality determination is not performed until both the outputs once return to near zero or less and Fz = 1 is set.

S14 to S16: This is a case where the expression (11) is not satisfied in the determination of step S8. This includes the case where only one output of the current read values Wan and Wbn is near zero or less. There are cases where the amount of change from the previous time is small even if it is a large output exceeding a level below zero. In these cases, the process proceeds to step S14.
Here, in Expression (11), if the load change is not monotonically increasing or the current state is not maintained, that is, the allowable value of the output change for determining that the condition of monotonic increase is broken, Wr, If the current read value is smaller than the previous read value and the absolute value of the difference is larger than Wr, that is, if the following equation is satisfied, the flag Fz is set to 0 because the monotonically increasing process has collapsed. (Step S15).
Wan-Wanu <0 and | Wan-Wanu> Wr
Wbn-Wbnu <0 and | Wbn-Wbnu |> Wr
However, when the output takes a negative value, it is replaced with 0, so that Wan, Wan ≧ 0 and Wbn, Wbnu ≧ 0.

  The value of Wr satisfies Wr <Wq and is set to a value slightly larger than a value that changes due to slight disturbance vibration when a certain load is placed. In addition, Wr is a small allowable load that does not change the state of error generation due to a monotonous increase even if the load may decrease from the previous time. However, if Wr = 0 is set, the condition of the monotonous increase process is lost even if the load fluctuates slightly due to disturbance, and the span abnormality cannot be determined again unless the load load is once lowered from the weighing platform. That is, it becomes impossible to detect the span abnormality due to the increase.

  On the other hand, if it is determined in step S14 that the monotonically increasing process is maintained, the current Wan and Wbn are error-corrected based on Expression (10) and stored in the memory in preparation for the next determination. (Step S16).

  As described above, according to the present embodiment, two half-bridge outputs can be easily and accurately based on the monotonic measurement addition determination unit that determines that the load change of the measuring instrument is a monotonic increase starting from the zero point. It is possible to correct the hysteresis error and accurately determine the span abnormality. Moreover, the method of the present embodiment is a method that can be carried out while the weighing operation is normally continued, and does not require a special operation for determining an abnormality.

  In the above description, the range from the zero point to the rated load is expressed as a single function. However, the error from the zero point to Ws / 2 and the error from Ws / 2 to the rated load at the time of adjustment are different. You may make it express with the function of.

<About a fault diagnosis apparatus according to another embodiment>
Instead of the system configuration of the failure diagnosis apparatus 20 in the previous embodiment, a system configuration as shown in FIG. 7 may be employed. In the failure diagnosis apparatus 20A, the same or similar parts as those in the failure diagnosis apparatus 20 of the previous embodiment are designated by the same reference numerals in the drawing, and detailed description thereof will be omitted. The description will focus on differences from the failure diagnosis apparatus 20 of the embodiment.

The failure diagnosis apparatus 20A of the present embodiment includes an analog adder circuit 57, an A / D converter 58, and an arithmetic circuit 26 that are provided for the full bridge circuit 15.
Here, the full bridge circuit 15 and the arithmetic circuit 26 are the same as those used in the failure diagnosis apparatus 20 of the previous embodiment.
In the failure diagnosis apparatus 20 of the previous embodiment, two A / D converters 24 and 25 are used, but in the failure diagnosis apparatus 20A of this embodiment, one A / D converter 58 is used.
In the fault diagnosis apparatus 20 of the previous embodiment, a DC voltage is applied in which the potential at the connection point 16 in the full bridge circuit 15 is + V and the potential at the connection point 17 is zero. A DC voltage is applied such that the potential at the connection point 16 is + V and the potential at the connection point 17 is −V. In this case, the half-bridge circuits 15a and 15b using the voltage reference fixed resistors 21 and 22 that are required in the failure diagnosis apparatus 20 of the previous embodiment are not necessary.

The analog adder circuit 57 includes a first operational amplifier 61, a second operational amplifier 62, a third operational amplifier 63, a fourth operational amplifier 64, and a fifth operational amplifier 65.
In the first operational amplifier 61, the input positive terminal 61a is connected to the connection point 18 of the full bridge circuit 15, the input negative terminal 61b is connected to the output terminal 61c, and the output terminal 61c is connected to the resistors 66 and 67. .
In the second operational amplifier 62, the input positive terminal 62 a is connected to the connection point 19 of the full bridge circuit 15, the input negative terminal 62 b is connected to the output terminal 62 c, and the output terminal 62 c is the resistor 68 and the fourth operational amplifier 64. Each is connected to the input positive terminal 64a.
In the third operational amplifier 63, the input positive terminal 63a is connected to the resistor 68, and is connected to the circuit ground 70 via the resistor 69, and the input negative terminal 63b is connected to the resistor 66. The output terminal 63 c is connected to the A / D converter 58 via the analog switch 72.
In the fourth operational amplifier 64, the input positive terminal 64 a is connected to the output terminal 62 c of the second operational amplifier 62, and the input negative terminal 64 b is connected to the circuit ground 70 via the resistor 73 and the resistor 74. The output terminal 64 c is connected to the A / D converter 58 via the analog switch 75.
In the fifth operational amplifier 65, the input positive terminal 65 a is connected to the circuit ground 70 via a resistor 76, and the input negative terminal 65 b is connected to a resistor 67, and is also connected to a output terminal 65 c via a resistor 77. The output terminal 65 c is connected to the A / D converter 58 via the analog switch 78.

  In the failure diagnosis apparatus 20A of the present embodiment, the analog load signal for the weighing instrument is synthesized by the third operational amplifier 63. The output load signal of the half-bridge circuit 15a on the connection point 18 side is eob, the load signal of the half-bridge circuit 15b on the connection point 19 side is eoa, and only one A / D converter 58 is used for all signals. The input is switched by analog switches 72, 75, 78. Thus, the number of A / D converters used may be selected as appropriate.

In this fault diagnosis device 20A, the full bridge output is detected separately from the half bridge output, the span coefficient is obtained independently, and adjusted so that Wn = Wan = Wbn for the same load load, and the initial weight storage Memory, zero weight storage memory,
Wn = kab. (Wabx-Wabi) -Wab
To be represented as

  The load signal calculation unit 52b in the present embodiment corresponds to the “load signal output unit” in the present invention, and the display device 54 in the present embodiment corresponds to the “load signal display unit” and the “zero point adjustment state display unit” in the present invention. The stability determination unit 52c in the present embodiment corresponds to the “stability determination unit” in the present invention, and the load change amount calculation unit 52d in the present embodiment corresponds to the “load change amount detection unit” in the present invention. The hysteresis error correction unit 52e in the present embodiment corresponds to the “hysteresis error correction unit” in the present invention, and the span abnormality determination unit 52f in the present embodiment corresponds to the “span abnormality determination unit” in the present invention.

  The load cell failure diagnosis device of the present invention can be suitably used for detecting an abnormality of a span occurring in a strain gauge type load cell used in a measuring instrument such as a conveyor scale, a hopper scale, a truck scale, a platform scale, a fee scale, etc. it can.

DESCRIPTION OF SYMBOLS 1 Compression type load cell 5 Double beam type load cell 9 Weighing table 15a, 15b Half bridge circuit 20, 20A Fault diagnosis device 26 Calculation circuit 52 Central processing unit 52b Load signal calculation part (load signal output means)
52c Stability determining unit (stability determining means)
52d Load change amount calculation unit (load change amount detection means)
52e Hysteresis error correction unit (hysteresis error correction means)
52f Span abnormality determination unit (span abnormality determination means)
53 Memory block 54 Display device (Load signal display means, zero point adjustment status display means)

Claims (9)

In a load cell failure diagnosis device using output signals from two terminals in a Wheatstone bridge circuit composed of a plurality of strain gauges affixed to a strain generating portion as load signals,
Two load signal output means for performing calculation processing so that the load signals output from the two terminals become load signals of the same magnitude with respect to the load applied to the load cell, and the two terminals A load cell fault diagnosis device comprising: zero point adjusting means for independently removing zero point components that fluctuate during normal weighing work in the load signal output from the load cell.   The load signal display means for individually displaying the two load signals output from the two load signal output means or displaying the absolute value of the difference between the two load signals is provided. The load cell failure diagnosis device according to claim 1.   Zero adjustment state display means for displaying a state where the absolute value of the difference between the two load signals output from the two load signal output means is zero or a value near zero, or a state where the absolute value is not zero or a value near zero. The load cell failure diagnosis apparatus according to claim 1, further comprising: 4. The failure of the load cell according to claim 3 , further comprising hysteresis error correcting means for individually correcting a hysteresis error that occurs in response to an increase or decrease in the two load signals output from the two load signal output means. Diagnostic device. The hysteresis error correction means includes monotonic increase determination means for determining that each of the two load signals is a monotonic increase starting from the vicinity of the zero point, and the two load signals are generated by the monotone increase determination means. 5. The load cell failure diagnosis apparatus according to claim 4 , wherein the hysteresis error correction means corrects the hysteresis error when it is determined that the increase is a monotonic increase starting from the vicinity of the zero point. In a load cell failure diagnosis device using output signals from two terminals in a Wheatstone bridge circuit composed of a plurality of strain gauges affixed to a strain generating portion as load signals,
Two load signal output means for performing arithmetic processing so that the load signals output from the two terminals become load signals of the same magnitude with respect to the load applied to the load cell;
A stability determining means for determining the stability of the two load signals output from the two load signal output means, a latest stable weight value generated based on the stability determination by the stability determining means,
A load change amount detecting means for detecting a difference from the stable weight value generated at the timing immediately before the latest stable weight value;
Span abnormality determining means for determining a load cell span abnormality based on the difference between the latest stable weight value detected by the load change amount detecting means and the previous stable weight value. Load cell failure diagnosis device. The span abnormality determination means includes a difference between the latest stable load value generated from one of the two load signal output means and the previous stable weight value, and the other load signal. Comparing the difference between the latest stable load value generated from the output means and the previous stable weight value and determining that the span of the load cell is abnormal when one difference is greater than the other difference The load cell failure diagnosis apparatus according to claim 6 . Load cell according to claim 6 or 7, characterized in that it comprises a hysteresis error correcting means for individually correcting the hysteresis error caused in accordance with the increase or decrease of two load signals output from the two load signal output means Fault diagnosis device. The hysteresis error correction means includes monotonic increase determination means for determining that each of the two load signals is a monotonic increase starting from the vicinity of the zero point, and the two load signals are generated by the monotone increase determination means. 9. The load cell failure diagnosis apparatus according to claim 8 , wherein the hysteresis error correction means corrects the hysteresis error when it is determined that the increase is a monotonic increase starting from the vicinity of the zero point.

Images