JPH112573A - Method and device for compensating age-based error of force detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable high-accuracy metering even in the case of continuously metering plural times by compensating a creep error of a metering signal and a creep recovery error. SOLUTION: A metering unit 1 converts displacement δ(t) corresponding to the load produced in a metering table by deflection of a deflective body 50 to an analog metering signal f(t) by a displacement signal converting unit 3. This is converted to a digital metering signal f(nt) by an A/D converter 4. When it is discriminated by a signal discriminator 11 in a compensating device 10 that the metering signal f(nt) includes a creep error, a compensating operation device 12 operates an inverse transfer function H1 (Z) of a transfer function G1 (Z) for generating the metering signal f(nT), and an accurate metering signal y(nT) where a creep error is compensated is output. When it is discriminated that a creep recovery error is included, similarly that creep recovery error is compensated to output an accurate metering signal y(nT) showing no load. Thus, the creep recovery error is removed so that the next material to be metered can be accurately metered after the creep error is compensated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a load based on force or mass, for example, a creep error accompanying a change with time generated when a load is applied to a force detector for measuring a load, and a load. A method and apparatus for compensating for creep recovery errors with aging that are created when removed.

[0002]

2. Description of the Related Art A creep error and a creep recovery error of the above-mentioned force detector will be described with reference to FIGS. FIG. 8 shows an example of a force detector, and shows a structure of a so-called strain gauge load cell type weighing unit. In this weighing unit, an elastic body having a rectangular parallelepiped shape having a uniform thickness and an opening 51 is used as a strain-generating body 50, and four corners thereof are particularly cut to form a strain-generating part 52. This strain body 5
0 is fixed to the base 53, and the weighing table 54 is provided on the load-side block 59. An object 55 to be weighed is placed on the weighing table 54.
Is loaded, the flexure element 50 bends at the flexure section 52, deforms into a parallelogram like a Roberval mechanism, and the weighing table 54 generates a displacement δ (t) corresponding to the load (see FIG. 5). In order to measure this displacement, four sets of strain gauges 56 are adhered to the surface of each strain generating portion 52 via an insulating material 57 as shown in an enlarged sectional view of FIG. Is covered with an insulating cover 58. The four sets of strain gauges 56 are bridge-connected, and the bridge output voltage is amplified and the analog weighing signal f
(T) (see FIG. 6), which is further subjected to A / D conversion to become a digital weighing signal f (nT). n is a sampling number, T is a sampling period, and nT represents time t.

[0003] As shown in FIG.
When load is applied at time t _0, the weighing signal is immediately generated the initial output f ₁ (t _0), the initial output f ₁ (t
₀ ) and does not stabilize, but slightly increases with time, as shown by curve 1. The weighing signal f (t) reaches a certain stable value after a long time, that is, a final output f ₁ (∞). Also, when the applied load is removed at time t ₁ , the weighing signal does not immediately return to zero, but slightly decreases with time as shown by curve 2 in FIG. A certain stable value is reached, i.e. zero of the final output.

The reason why the weighing signal increases with time as shown by the curve 1 is that the weighing unit has a creep characteristic, and as shown by the curve 2, The reason for this is that the measurement unit has creep recovery characteristics. Even when the applied load is small, the weighing signal gradually increases as shown by the curves 3 and 4 due to the creep characteristics and the creep recovery characteristics.
Decrease gradually. However, FIG. 6 shows an example in which the creep characteristic curve gradually increases and the creep recovery characteristic curve gradually decreases. On the contrary, the creep characteristic curve gradually decreases due to the mechanical characteristics of the strain element 50 and the strain gauge 56 and the like. In some cases, the creep recovery characteristic curve gradually increases. Assuming that such a creep characteristic is a specific dynamics model G _R (s) and the constant load is a stepwise input w (s), the weighing signal f (s) output from the weighing unit is

F (s) = G _R (s) w (s) (1)

[0006] However, the responsiveness of the output weighing signal to the applied load of the weighing unit is much higher than the response of the creep characteristic, and thus the responsiveness of the output weighing signal to the applied load of the weighing unit is ignored.

Then, the pulse transfer function G _R (z) of the discrete-time system of the model G _R (s) is obtained by setting the sampling time of the output weighing signal as T, and as shown in equation (2), this G _{R is obtained.} By inputting the output weighing signal f (nT) of the weighing unit to the inverse transfer function H _R (z) of (z), it is possible to obtain the weighing signal y _R (nT) on which creep error compensation has been performed.

Y _R (nT) = H _R (z) f (nT) (2)

[0009] That is, as shown in FIG. 7, n _a T~n _b
Output weighing signal f (nT) of the weighing unit at time T
Is compensated by the inverse transfer function H _R (z) to obtain a weighing signal y _R (nT) for which creep error compensation has been performed.
As a result, the weight of the object 55 can be accurately measured (see Japanese Patent Application Laid-Open No. 4-1221).

[0010]

However, according to the conventional creep error compensation method described above, as shown in FIG. 7, when the object 55 to be weighed is loaded on the weighing table 54, n _a T to n _b
Although the compensated weighing signal y _R (nT) can be obtained by compensating for the creep error at the time T, the creep recovery error generated when the object 55 is removed from the weighing platform 54 cannot be compensated. In the next weighing, the next weighing signal y _R (nT) includes the weighing error g (n _c T) based on the creep recovery error, so that the weight of the object 55 can be accurately weighed. There is a problem that can not be.

That is, as shown in FIG. 7, the weighing signal y _R
(NT) is, n _b in T~n _c T time gradually decreases along the creep recovery characteristic curve g (nT), the time n _c
Zeros T In the metering apparatus will be reduced by g (n _c T), the g (n _c T) is the next metering errors. Here, if the n _b T~n _c T time is considered as not done compensating creep error by the inverse transfer function H _R (z), and that to compensate for creep error, also large An error will occur.

An object of the present invention is to provide a compensation method and apparatus capable of performing high-precision weighing even when weighing is performed a plurality of times continuously by compensating for a creep error and a creep recovery error of a force detector. Aim.

[0013]

According to a first aspect of the present invention, there is provided a method for compensating for a temporal error of a force detector, comprising the steps of: generating a weighing signal including a creep error accompanied by a temporal change when a load is applied; A compensating method for compensating for the creep error and the creep recovery error of the force detector that generates a weighing signal including a creep recovery error that changes with time when the weighing signal includes a creep error;
Or a step of determining whether the signal includes a creep recovery error, and for a weighing signal including the creep error based on the determination result, a creep in which an inverse transfer function of a transfer function for generating a creep error signal compensates for the creep error. An error-compensated weighing signal is generated, and for a weighing signal including the creep-recovery error, a creep-recovery-error-compensated weighing signal in which the inverse transfer function of the transfer function for generating the creep-recovery error signal compensates for the creep-recovery error And (c) performing the following steps.

A time error compensating device for a force detector according to a second aspect of the present invention generates a weighing signal including a creep error accompanied by a time change when a load is applied, and generates a time signal when the load is removed. A compensating device for compensating the creep error and the creep recovery error when used with a force detector that generates a weighing signal including a changing creep recovery error;
Determining means for determining whether the weighing signal contains a creep error or containing a creep recovery error; and a transfer function for generating a creep error signal for the weighing signal containing the creep error based on the determination result. Generates a creep error compensated weighing signal in which the creep error is compensated, and for the weighing signal including the creep recovery error, the inverse transfer function of the creep recovery error signal generation transfer function is the creep recovery error. And a compensating means for generating a weighing signal with a compensated creep recovery error.

According to the present invention, when a weighing signal generated by the force detector is input, it is determined whether the weighing signal contains a creep error or a creep recovery error, and the creep error is determined based on the result of the determination. For a weighing signal containing, the inverse transfer function of the transfer function for generating a creep error signal generates a creep error compensated weighing signal with the creep error compensated, and for a weighing signal containing a creep recovery error, The inverse transfer function of the transfer function for generating the error signal can generate a creep recovery error compensated weighing signal in which the creep recovery error is compensated.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a weighing device to which a compensating device using a method of compensating for a temporal error of a force detector according to the present invention will be described with reference to the drawings. This weighing device comprises a weighing unit 1 and a compensating device 10 as shown in FIG.
It has. The weighing unit 1 is equivalent to the conventional one shown in FIGS. 8 and 9, and includes a strain body 50 and a displacement signal conversion unit 3. One end of the flexure element 50 is fixed to the base 53, and the weighing table 54 is
Is provided. When the object 55 to be weighed is loaded on the weighing table 54, the flexure element 50 bends at the flexure portion 52, and the weighing table 54 generates a displacement δ (t) corresponding to the load (see FIG. 5).

The displacement signal conversion unit 3 converts the displacement of the weighing table 54 into an analog weighing signal f (t) and outputs the analog weighing signal f (t). Four sets of strain gauges adhered to the surface of each strain generating section 52 56 and these four sets of strain gauges 5
Bridge circuit (not shown) formed by bridging 6
And The bridge output voltage of the bridge circuit is amplified by an amplifier provided in the displacement signal conversion unit 3 and output as an analog measurement signal f (t).

This analog weighing signal f (t) is A / D converted by an A / D converter (analog-digital converter) 4 shown in FIG. 1 to become a digital weighing signal f (nT) (see FIG. 6). ). n is a sampling number, T is a sampling period, and nT represents time t.

The compensator 10 includes a signal discriminator 11 and a compensation calculator 12. The signal discriminator 11 outputs the weighing signal f
Whether (nT) contains a creep error, a creep recovery error, or a weighing signal f (nT)
Is determined to be constant, and the result of this determination is used as the compensation calculator 1
2 is output. However, in this embodiment,
When the weighing signal f (nT) includes a creep error, the signal gradually increases as shown by a curve 1 or 3 in FIG. 6, and when the weighing signal f (nT) includes a creep recovery error, the signal gradually decreases as shown by a curve 2 or 4 in FIG. It shall be. And
Whether the weighing signal f (nT) is gradually increasing, gradually decreasing, or constant is determined by f (nT)> f ((n-1)
T), f (nT) <f ((n-1) T), and f (n
T) = f ((n-1) T) by sequentially determining which one. And f (nT)> f
((N−1) T), the weighing signal f
If (nT) includes a creep error and is determined to increase gradually, and if it is determined that f (nT) <f ((n−1) T), the weighing signal f (nT) is determined to be a creep recovery error. And it is determined that the value gradually decreases, and f (nT) = f
((N−1) T), the weighing signal f
It is determined that (nT) is constant.

The compensation calculator 12 performs a compensation calculation of the weighing signal f (nT) based on the determination result of the signal determiner 11,
It outputs the compensated weighing signal y (nT) shown in FIG. The output weighing signal y (nT) is displayed on a display unit provided in the weighing device, not shown in the figure. That is, when the signal discriminator 11 determines that the weighing signal f (nT) includes a creep error and f (nT)> f ((n−1) T), the weighing unit 1 including the creep error of the weighing unit 1 Transfer function G that generates signal f (nT)
₁ (z) inverse transfer function H ₁ of (z), to generate the entered the measuring signal f (nT) creep error compensated weight signal y to compensation calculation to to to compensate for creep error in the (nT). When the signal discriminator 11 determines that the weighing signal f (nT) includes the creep recovery error and f (nT) <f ((n-1) T), the creep recovery error of the weighing unit 1 is determined. The inverse transfer function H ₂ (z) of the transfer function G ₂ (z) for generating the weighing signal f (nT) includes a compensation operation for the input weighing signal f (nT) to compensate for the creep recovery error. Recovery error compensated weighing signal y (n
T). When the signal discriminator 11 determines that the weighing signal f (nT) is constant and f (nT) = f ((n−1) T), the weighing signal f (nT) is converted to the weighing signal y ( nT). Note that the transfer function G
₁ (z), inverse transfer function H ₁ (z), transfer function G
₂ (z) and the inverse transfer function H ₂ (z) will be described later.

According to the weighing device having the above structure, when the object 55 is loaded on the weighing table 54 of the weighing unit 1 at the time n _a T as shown in FIG. A weighing signal f (nT) including an error is generated, and in response to this input, the compensation calculator 12 performs inverse transfer of a transfer function G ₁ (z) that generates a weighing signal f (nT) including a creep error signal. By calculating the function H ₁ (z) and compensating for the creep error, an accurate weighing signal y (nT) representing the weight of the object 55 can be output. The weighing signal y (nT) is output as a signal of substantially constant magnitude until the time n _b T when the object 55 is removed from the weighing table 54.

Then, as shown in FIG. 4, at time n _b T,
When the weighing object 55 is removed from the weighing platform 54, the weighing signal f including the creep recovery error with the gradually decreasing time-dependent change.
(NT) is generated, and in response to this input, the compensation calculator 12 outputs the weighing signal f (n) including the creep recovery error signal.
T) the inverse transfer function H of the transfer function G ₂ (z)
₂ (z) is calculated to compensate for the creep recovery error, and to output an accurate weighing signal y (nT) (= 0) indicating that there is no load on the weighing platform 54 with no object 55 loaded. it can. The weighing signal y (nT) (= 0) is output as a signal having a substantially constant magnitude until the time n _c T at which the next object 55 is loaded on the weighing table 54. As a result, FIG.
In the conventional compensating device shown in FIG.
When 5 except the creep recovery error g (n _c T) but is generated, in this embodiment, creep recovery error g (n _c
T) can be eliminated. This allows time n _c
At T, when the next weighing object 55 is loaded on the weighing table 54, an accurate weighing signal y ( _ncT ) representing the weight of the weighing object 55 with the creep error compensated can be output. In this way, the weight of the new object 55 can be sequentially and accurately measured.

Next, how to calculate the inverse transfer functions H ₁ (z) and H ₂ (z), which are the contents of the compensation operation of the compensation operation unit 12, will be described. First, assuming that the creep characteristic model and the creep recovery characteristic model are different from each other, the creep error and the creep recovery error are approximated using different dynamics models. That is, the stair-like input w (s) of the constant load is applied to the weighing table 54.
The creep characteristic model corresponding to the weighing signal f ₁ (s) including the creep error generated by the weighing unit 1 when applied to the weighing unit 1 is defined as G ₁ (s), and the stair-like input w (s) is input from the weighing table 54. The creep recovery characteristic model corresponding to the weighing signal f ₂ (s) including the creep recovery error generated by the weighing unit 1 after the removal is defined as G ₂ (s). However,
w (s), f ₁ (s) and f ₂ (s) are w
(T), f ₁ (t) and f ₂ (t) are Laplace transformed. Here, the models G ₁ (s) and G ₂ (s)

G _i (s) = 1 + β _i / (1 + τ _i s) (3)

A first-order delay element model is used. Where i
= 1 represents a creep characteristic model, and i = 2 represents a creep recovery characteristic model. Β _i and τ _i are the gain and the time constant of the creep characteristic model and the creep recovery characteristic model. However, although the first-order lag element model G _i (s), or as a secondary or higher order lag element model.

Next, how to determine the gain β _i and the time constant τ _i will be described with reference to FIGS. First, gain β ₁
Is determined using equation (4).

Β ₁ = (δ ₁ (∞) -δ ₁ (t ₀ )) / δ ₁ (t ₀ ) = (f ₁ (∞) -f ₁ (t ₀ )) / f ₁ (t ₀ )・ ・ (4)

However, as shown in FIGS. 5 and 6, δ ₁
(T ₀ ) is the initial deflection when a unit load is applied to the weighing table 54, and f ₁ (t ₀ ) is a weighing signal corresponding to δ ₁ (t ₀ ). δ ₁ (∞) is the final deflection at the time when about 2 hours or more at which the creep is substantially stabilized after the unit load is applied, and f ₁ (∞) is a weighing signal corresponding to δ ₁ (∞). Therefore, the gain β ₁ can be obtained by measuring f ₁ (t ₀ ) and f ₁ (∞).
The time constant τ ₁ is obtained using the equation (5).

(Df / dt) _{t = t0} = β ₁ / τ ₁ (5)

[0030] However, (df / dt) t = t0 , as shown in FIGS. 5 and 6, the initial differential coefficient at time t ₀ in applying the unit load on the weighing table 54. Thus, continuous data of the measuring signal f (t) is measured to determine the time constant tau ₁ the data obtained by calculating the a (df / dt) t = _t0 and beta ₁ are substituted into Equation (5) be able to. Next, the gain β ₂ is obtained using Expression (6).

Β ₂ = (δ ₂ (∞) −δ ₂ (t ₁ )) / δ ₂ (t ₁ ) = (f ₂ (∞) −f ₂ (t ₁ )) / f ₂ (t ₁ ) ·・ ・ (6)

However, as shown in FIGS. 5 and 6, δ ₂
(T ₁ ) is the initial return amount from δ ₁ (∞) when the unit load applied to the weighing table 54 is removed, and f ₂ (t
₁ ) is a weighing signal corresponding to δ ₂ (t ₁ ). δ
₂ (∞) is the δ at the time when about 2 hours or more after the unit load is removed and the recovery of creep is almost stabilized.
₁ (∞) is the final return amount, f ₂ (∞) is δ
_{2 This} is a weighing signal corresponding to (∞). Therefore, f ₂ (t
_The gain β ₂ can be obtained by measuring ₁ ) and f ₂ (∞). The time constant τ ₂ is obtained using equation (7).

(Df / dt) _{t = t1} = −β ₂ / τ ₂ (7)

[0034] However, (df / dt) t = t1 , as shown in FIGS. 5 and 6, the initial differential coefficient at time t ₁ when removing the unit load that is applied to the weighbridge 54. Thus, continuous data of the measuring signal f (t) is measured to determine the time constant tau ₂ the data obtained by calculating the a (df / dt) t = _t1 and beta ₂ into Equation (7) be able to. The pulse transfer function G _i (z) of the discrete time system of the model G _i (s) obtained as described above is

[0035] _{G i (z) = (1} + B i z -1) / (1 + A i z -1) a (8). _{However, A i = -exp (-T /} τ i) B i = - a _{[exp (-T / τ i) -β} i {1-exp (-T / τ i)} ].

The inverse transfer function H _i of the pulse transfer function G _i (z) (z) is

[0037] _{H i (z) = (1} + A i z -1) / (1 + B i z -1) ··· (9)

## EQU4 ## The weighing signal f (nT) passes through the inverse transfer function H _i (z) to compensate for the creep error or the creep recovery error by the weighing signal y (n).
T).

Y (nT) = H _i (z) f (nT) (10)

That is, as shown in FIG.
Means that when the signal discriminator 11 determines that the weighing signal f (nT) is a gradually increasing signal including a creep error, the weighing signal f (nT) is H ₁ (z) = (1 + A ₁ z ⁻¹ ) / ( 1
+ B ₁ z ⁻¹ ) to output a weighing signal y (nT) in which the creep error has been compensated. Then, the weighing signal f (n
When the signal discriminator 11 determines that T) is a gradually decreasing signal including a creep recovery error, the weighing signal f (nT)
Is passed through H ₂ (z) = (1 + A ₂ z ⁻¹ ) / (1 + B ₂ z ⁻¹ ) to compensate for the creep recovery error.
(NT) is output.

In the above embodiment, however, as shown in FIG. 1, the displacement δ (t) of the flexure element 50 is converted by the displacement signal conversion unit 3 into an analog measurement signal f (t), for example, an analog electric signal of voltage or current. e (t), and the analog weighing signal f (t) is converted into a digital weighing signal f (nT) by the A / D converter 4 and output.
Instead, as shown in FIG.
(T) is converted by the displacement frequency converter 5 into a digital weighing signal f
(NT) and output. The displacement frequency converter 5 generates a pulse having a frequency corresponding to the deflection δ (t) of the flexure element 50 and counts the number of pulses within a predetermined time repeatedly at predetermined time intervals. The counted values form a discrete metric signal f (nT). As a result, the output of the displacement frequency converter 5 becomes A
The signal can be directly input to the compensator 10 without passing through the / D converter 4.

In the above embodiment, the signal discriminator 1
1 determines whether the weighing signal f (nT) is gradually increasing, gradually decreasing, or constant, by determining f (nT)> f ((n−
1) T), f (nT) <f ((n-1) T), and f
(NT) = f ((n-1) T) is performed by successively determining which of the following, but instead of this, f (nT) -f ((n-1) T ) ≧ C, f (nT)
−f ((n−1) T) <− C and f (nT) = f
((N-1) T) is sequentially determined,
When it is determined that f (nT) −f ((n−1) T) ≧ C, it is determined that the weighing signal f (nT) includes a creep error and gradually increases, and f (nT) ) -F
When it is determined that ((n−1) T) <− C, the weighing signal f (nT) includes a creep recovery error and is determined to be gradually decreasing, and f (nT) = f ((n -1) When it is determined that T), the measurement may be performed by determining that the weighing signal f (nT) is constant. Here, C is a constant. The reason why such a configuration is adopted is that, in the above-described embodiment, if error oscillation is included in f (nT), the determination of the gradual increase and decrease of f (nT) may be erroneous. This is because if the amplitude of the error oscillation is equal to or smaller than C, it is not erroneous to judge whether f (nT) is gradually increased or decreased.

When the compensation calculator 12 of the above embodiment is capable of performing high-precision compensation, the compensation may be performed on the assumption that the gain β _{i of the} creep coefficient is a function of the temperature θ. Here, the gain β _i is represented by a quadratic polynomial. For example, the value at 20 ° C. is β (2
0), the following equation holds.

Β _i (θ) = β _i (20) [1 + C ₁ (θ−20) + C ₂ (θ−20) ² (11)

The coefficient β _i (20), C ₁ , C
The method of determining ( ₂ ) is disclosed in Japanese Patent Application Laid-Open No. 4-12221, which is conventionally known, and a detailed description thereof will be omitted.
Then, the gain β _i determined in this way can be used for the inverse transfer function H _i (z) of the compensation calculator 12. That is, the gain β ₁ is used for the inverse transfer function H ₁ (z), and the gain β ₂ is used for the inverse transfer function H ₂ (z).

[0046]

According to the first aspect of the present invention, the transmission of the creep error signal generation is performed in response to the input of the weighing signal including the creep error accompanied by the time change generated when a load is applied to the force detector. Since the inverse transfer function of the function is configured to compensate for the creep error, the creep error can be compensated while the load is applied, and an accurate weighing signal indicating the magnitude of the load can be output. Thus, for example, when the weight of the object is measured, a constant weight value can be obtained, and the weight of the object can be accurately measured.

For an input of a weighing signal including a creep recovery error accompanied by a change with time generated when the load is removed from the force detector, an inverse transfer function of the transfer function for generating the creep recovery error signal is obtained. Is configured to compensate for the creep recovery error, so that the creep recovery error can be compensated with the load removed, and an accurate weighing signal indicating no load can be output. As a result, conventionally, when the load is removed after the current measurement, a creep recovery error g (n _c T) is generated. In the present invention, the creep recovery error g (n _{c
Since cT} ) can be eliminated, it is possible to output an accurate weighing signal in which the creep error is compensated in the next weighing. Therefore, the weight of each weighing object can be accurately measured even when the weighing of the weighing objects is performed plural times continuously.

Further, according to the present invention, even when, for example, half of the load is removed from the force detector, an accurate weighing signal corresponding to the other half of the load can be obtained. Therefore, if the present invention is applied to, for example, performing loss-in weighing of the object to be weighed discharged from the weighing hopper, the weight of the object to be weighed discharged from the weighing hopper can be accurately measured.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an electric circuit of a weighing device using a device for compensating for a temporal error of a force detector according to an embodiment of the present invention.

FIG. 2 is a block diagram showing an electric circuit of a weighing device according to another embodiment of the present invention.

FIG. 3 is a block diagram showing a compensating device of the embodiment.

FIG. 4 is a diagram showing a creep error and a creep recovery error compensated weighing signal output from the compensator of the embodiment.

FIG. 5 is a diagram showing a relationship between displacement of the flexure element of the embodiment and time.

FIG. 6 is a diagram showing a relationship between an electric signal corresponding to the displacement of the flexure element of the embodiment and time.

FIG. 7 is a diagram showing a weighing signal after creep error compensation for explaining a conventional compensating device.

FIG. 8 is a front view showing a strain body provided in the weighing device of the embodiment.

FIG. 9 is an enlarged cross-sectional view of the strain gauge attached to the flexure element of the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Measuring unit (force detector) 50 Flexure element 10 Compensator 11 Signal discriminator 12 Compensation calculator

Claims (2)

[Claims] 1. A force for generating a weighing signal including a creep error with a change with time when a load is applied, and generating a metric signal including a creep recovery error with a change with time when the load is removed. A compensating method for compensating the creep error and the creep recovery error of the detector, comprising the steps of: determining whether the weighing signal includes a creep error or including a creep recovery error; and For a weighing signal including an error, an inverse transfer function of a transfer function for generating a creep error signal generates a creep error compensated weighing signal in which the creep error is compensated, and for a weighing signal including the creep recovery error, , A creep recovery error compensated weighing signal in which the inverse transfer function of the creep recovery error signal generation transfer function compensates for the creep recovery error Time error compensation method of the force detectors, which comprises generating the steps of, a. 2. A force for generating a metric signal including a creep error with a change with time when a load is applied, and generating a metric signal including a creep recovery error with a change with time when the load is removed. A compensating device that is used together with a detector to compensate for the creep error and the creep recovery error. A discriminating means for discriminating whether the weighing signal includes a creep error or a creep recovery error. For the weighing signal including the creep error based on the result, the inverse transfer function of the transfer function for generating the creep error signal generates a creep error compensated weighing signal in which the creep error is compensated, and the weighing signal including the creep recovery error. For signals, the inverse transfer function of the creep recovery error signal generation transfer function compensated for the creep recovery error. Time error compensation apparatus of the force detector characterized by comprising a compensation calculating means for generating a compensated weighing signal.