JPH0212018A - Electromagnetic flow meter - Google Patents

Abstract

PURPOSE:To compensate and eliminate a DC offset voltage without producing tailing caused by spike noises by integrating the output voltage of an amplifier by means of an integration circuit and feeding back the output voltage of the integration circuit to the input side of the amplifier so that the output voltage can be inverted in polarity. CONSTITUTION:An excitation circuit 5 supplies an exciting current I to exciting coils 3 and 4 synchronously to a control signal P4 from a control circuit 6. The electromotive force produced across electrodes 2a and 2d is inputted to a differential amplifier 7 and the output e1 of the amplifier 7 is connected to one input side of an adder circuit 8. The input side of an amplifier 9 having an amplification degree A is connected to the output side of the circuit 8. A signal V3 corresponding to the difference between values V1 and V2 respectively stored in sample Eidophor circuits 14 and 15 is produced on the output side of a subtractor circuit 16. Moreover, a compensation circuit 12 which eliminates a DC offset voltage is constituted of the circuit 8, the amplifier 9, a switch S1, and an integrator 11. Occurrence of tailing is prevented by opening the switch S1 only during the period in which spike noises are produced.

Description

[Detailed Description of the Invention] [Industrial Application Field] Does the present invention compensate for the DC offset voltage generated in the electrodes? This invention relates to an improvement of an electromagnetic flowmeter having a 1i1 circuit.

[Prior art and its problems]

In addition to a signal voltage proportional to the fluid flow rate, a DC offset voltage is generated at the electrodes of an electromagnetic flowmeter due to electrochemical action. This DC offset voltage ranges from several meters to several hundred mV.
This will cause an obstruction to flow rate measurement.

Therefore, a compensation circuit for removing the DC offset voltage has been proposed in Japanese Patent Laid-Open No. 146113/1983.

This prior art is shown in FIG.

FIG. 4 shows an internally insulated tube 1, in which two excitation coils 3.4 produce a magnetic field H perpendicular to the axis of the tube 1. Inside the tube 1 are electrodes 2a and 2b from which an electromotive force proportional to the flow rate of the fluid across the magnetic field is extracted. The excitation circuit 5 generates a DC excitation current I in synchronization with the control signal P4 from the wi control circuit 6.
is supplied to the excitation coil 3.4. The electromotive force generated at the electrodes 2a and 2b is input to a differential amplifier 7. The output C1 of the differential amplifier 7 is connected to one input side of an adder circuit 8, and the input side of an amplifier 9 having an amplification factor A is connected to the output side of the adder circuit 8. Two sample idle and hold circuits 14 and 15 are connected in parallel to the output side of the amplifier 9. The sample idle and hold circuit 14 includes a switch S2 operated by a control signal P2 from the control circuit 6.
and a capacitor C2. The sample idle and hold circuit 15 includes a switch S3 operated by a control signal P3 from the control circuit 6 and a capacitor C3. The output sides of the two sample eye-hold circuits 14.15 are connected to the two input sides of the subtraction circuit 16, and the output side of the subtraction circuit 16 is connected to the sample eye-hold circuit 14.1.
A signal v3 corresponding to the difference between the values V+ and V2 stored in 5
occurs. The output signal v3 is a measurement signal proportional to the flow velocity in the tube 1.

A switch S4 operated by a control signal P7 from the control circuit 6 is connected to the output side of the amplifier 9, and a resistor R is connected to the output side of the amplifier 9.
4, an integrator 21 consisting of a capacitor C4 and an operational amplifier 20 is connected. The output side of the integrator 21 is connected to the second input side of the adder circuit 8. addition circuit 8,
The amplifier 9, the switch S4, and the integrating circuit 21 constitute a compensation circuit 22 that removes DC offset.

The timing chart of the prior art shown in FIG. 4 is shown in FIG. 5. During a certain period Tb or Td immediately before the excitation current is switched, the output e5 is compensated to zero by feedback, and the feedback le6 for this compensation is Next half period Tc
or Ta', and the adder 8 on the input side of the amplifier 9
The DC offset voltage is removed by continuously adding
It amplifies the signal voltage. Specifically, period Tb
When the O switch is opened, the integrator 21 outputs the output e of the amplifier 9 with the time constant 04·R4 of the capacitor C4 and the resistor R4.
5 is integrated and its output e6 is added to the other input of the adder 8 as a compensation voltage. Since the output e6 has a value opposite in polarity to the input e1, and the time constant 04·R4 has a smaller value than the period Tb, e5 becomes zero during the period Tb.
When the switch is opened at time tc, the output voltage e6 of the integrator 21 is held at the output voltage e6(tc) at time tc.

If the input e1 includes a DC offset voltage, the integrator output e et al (tc) in the period Tb has a value that compensates for zero including the DC offset. During period Tc, the sum of input signal e1 and compensation voltage e et al (tc) increases! Special equipment 9
It is amplified by , and becomes the output e5. Since the voltage 6 (tc) also includes the compensation voltage for the DC offset,
If the DC offset voltage is constant between the period Tb and the period Tc, this can be removed.

According to this method, the DC electromotive force generated in the electrode, that is, the DC offset voltage, can be removed, but if noise is superimposed on the input signal e1 at the end of the period Tb, that is, just before the switch S4 is opened, the noise In order to compensate the value including 0 to zero, and to maintain the compensation voltage e6 necessary for the compensation over the next half-cycle period Tc and continue to apply it to the adder 8,
An error occurs in the output voltage during the period Tc due to noise.
In addition to the DC offset voltage mentioned above, random noise called high-speed fluid noise that increases in proportion to the flow velocity of the fluid is generated at the electrodes of the electromagnetic flowmeter. Further, white noise generated by the semiconductor forming the differential amplifier 7 is also superimposed on the input signal el and input to the compensation circuit 22.

The high-speed fluid noise is 1 when the flow velocity is 7 to 10 m/s.
On the other hand, in the case of low power consumption type electromagnetic flowmeters such as 2-wire electromagnetic flowmeters and battery operated electromagnetic flowmeters, the flow velocity signal is a minute voltage of about 10μV when the flow velocity is Lm/s, so this The influence of random noise on the compensation circuit 22 cannot be ignored.

If random noise of a large value is superimposed just before the switch s4 is opened in the compensation circuit 22, a large difference in M will occur in the output during the next half cycle. Conversely, if the random noise has a small value, the error between the next half cycles will be small.

For the above reasons, when the input signal has variations due to high-speed fluid noise, the compensation circuit 22 further amplifies the variations in the output, resulting in a drawback.

Therefore, there is a circuit shown in FIG. 6 as a conventional technique for removing the DC offset voltage generated in the electrodes without increasing the variation in output due to random noise such as high-speed fluid noise. The circuit in the sixth part is the same as the circuit in FIG. 4 except for the switch S4, which is connected 7, and the time constant C and R1 are increased.

Feedback for constant compensation from the output of the amplifier 9 is performed through an integrating circuit 11 consisting of a resistor R1, a capacitor C1, and an operational amplifier 10, and the time constant 01·R+ of the integrating circuit II is set for a half cycle period 7+10↑2. Since it is a relatively large value, even if there is a large peak value of high-speed fluid noise, the variation will not be magnified.

The circuit shown in FIG. 6 amplifies an input signal of about 6 Hz to 20 Hz, that is, a flow velocity signal for an electromagnetic flowmeter, in the same way as a general amplifier, while removing slow changes in the input signal and DC components.

By the way, with the electromagnetic flow rate of rectangular wave excitation, large spike-like noise occurs at the moment the excitation current is switched. When this noise is input to the circuit of FIG. 6, it is differentiated by the auxiliary IM circuit 12, producing an error voltage that remains at the output until after the spike-like noise has disappeared. This point will be explained with reference to FIGS. 6 and 7.

■ in Fig. 7 indicates the excitation current, and el indicates the electrodes 2a and 2b.
The waveform obtained by amplifying the electromotive force generated in the differential amplifier 7 becomes the input signal of the compensation circuit 12. els is the input signal e1
Of the waveforms, only the effective signal proportional to the flow velocity and eLN represent only the spike-like noise due to switching of excitation, respectively, as square waves. e4s is the waveform of the output signal of the amplifier 9 for els, and 84N is the waveform of the output signal of the amplifier 9 for elN. The output signal e4 of the amplifier 9 is e4s and e
4N is added and synthesized, and is an output waveform for a general input signal 1 containing spike noise. e2s is the output waveform of the integrator 11 for els, and 62N is the output waveform of the integrator 11 for elN.

First, an ideal case where only the flow velocity signal els is used will be described assuming that the operation starts from time t1.

At time t, the input signal els takes a positive value and maintains a constant value during the first half period of excitation, that is, during T++Tz.

This value is passed through an adder 8 to the input of an amplifier 9;
It is multiplied by , and becomes 84s. The integrator 11 has a time constant C1・R
1, the above e4s is integrated to generate a compensation voltage e2s, which is applied to one input of the adder 8. The adder 8, the amplifier 9, and the integrator 11 constitute a compensation circuit 12.

Since the time constant C+-R1 is a sufficiently long value with respect to the time width T1+T2, this compensation loop operates in the direction of slowly reducing the output e4s to zero. At time t3, input signal el
When s switches from positive to negative polarity, the amplifier 9 immediately amplifies the change and outputs it, but the integrator 110 output 22s does not change instantaneously. Therefore, the change in e4s is e
The value is the change in ls multiplied by A.

In the second half cycles T3 and T4, the input signal els has a negative polarity, and the compensation voltage 82s is the same square as in the first half cycle.
The output e4S has a waveform that slowly approaches zero.

If el includes a DC component, the integrator 11 outputs e4
It has a function of integrating the DC component included in the output e4 to generate a compensation voltage e2 having a polarity opposite to that of the DC component, and adding it to the adder 8 to make the DC component included in the output e4 zero. This is the basic function of the circuit of FIG.

According to the following explanation, if the input signal e1 is only the flow velocity signal Qis, the rectangular wave input signal will become a waveform like 24s in FIG. 7 due to the compensation operation of the integrating circuit 11, but as a signal amplification of the electromagnetic flow rate Does not cause any problems.

Next, consider the case where only spike noise elN is input.

When spike noise elN is input only during period T1, 1. It is multiplied by the amplifier 9 to become e4N. Integral 511
integrates 24N with a time constant of 61·R+ to generate a compensation voltage e2)f(tz) having the opposite polarity to exN at time tz.

Correspondingly, e4N decreases slowly throughout period T1. Although elN becomes zero at time t2, the output voltage of the integrator does not change instantaneously and remains at 62N (tz). This value is multiplied by A via the adder 8 to become e=r(12), which has the opposite polarity to e4N in period T1, and in period T2, due to the compensation operation of the integrator 11, the time constant C1·R
At 1, the waveform slowly decreases. This 84N(tz
) is the initial value and the waveform that decreases is spike noise elN
This is the error voltage due to the operation of Second half cycle T3・T'i
However, by a similar compensation operation, the output waveform 84N, which is symmetrical to the first half cycle by one day, is reduced by 4F. The general operating state of the electromagnetic current meter is an output voltage waveform e4 that is a combination of the output signals ells and e4N, and includes the above operation in periods T2 and T4. When spike noise CIN occurs in period T1 or T3, an output voltage is generated in period T2 or T4 even though the human power elN is zero. This is because feedback is applied from the output to the input side via the integrator 11, causing a differential result.
The operation occurred during noise-free periods Tz and T4. The error voltage due to this operation is unrelated to the flow velocity,
This is of a nature that cannot be removed by signal processing by the sample eye-hold circuits 14 and 15 and the subtraction circuit 16 at the subsequent stage.

The peak value of the spike noise generated by switching the excitation current changes depending on the temperature and conductivity of the fluid, which causes the magnitude of operation to change, causing zero point fluctuations in the electromagnetic flowmeter.

The signal level of a low power consumption type electromagnetic flowmeter is when the flow velocity is 1 m.
/Sec, is as small as about 10 μV. Flow rate Q, 1
m/Sec, to guarantee an accuracy of 10.5%, the fault voltage at a flow velocity Q of 1 m/Sec is set to 5 μV.
Must be below. However, the spike noise generated at the electrode can reach 10 μV, and the temperature
It also changes depending on the conductivity of the fluid. Therefore, it is practically impossible to reduce the fault voltage due to operational phenomena to 5 μV or less using the circuit shown in FIG.

In view of the above, an object of the present invention is to propose an electromagnetic flowmeter that compensates for and eliminates DC offset voltage without causing operation due to spike noise and without increasing output variation due to random noise. .

[Means to solve the problem]

In order to achieve the above object, the electromagnetic flowmeter of the present invention is an electromagnetic flowmeter that measures the flow rate of fluid using a DC magnetic field whose steady value changes periodically, and integrates the output voltage of an amplifier using an integrating circuit. In this compensation circuit that compensates for the DC offset voltage superimposed on the signal voltage by feeding back the output voltage of the integrating circuit to the input side of the amplifier so as to have a reverse polarity, it is possible to switch the steady value of the DC magnetic field. The present invention is characterized in that the input of the integrating circuit is opened during a period including the time during which spike noise occurs.

The input of the integrating circuit may be opened only during the period when spike noise occurs.

Further, it is preferable that the time constant of the integrating circuit be set longer than the period in which the steady-state value of the DC magnetic field changes.

In addition, the signal voltage generated at the electrode is at a flow rate of 1 m/See.

It is particularly effective in electromagnetic flowmeters with a voltage of about 10 μV or less, and even more effective in low-power electromagnetic flowmeters that use batteries as a power source.

〔Example〕

FIG. 1 shows an internally insulated tube 1 in which two excitation coils 3, 4 generate a magnetic field H perpendicular to the axis of the tube. Inside the tube 1 are electrodes 2ai2b from which an electromotive force proportional to the flow rate of the fluid across the magnetic field is extracted. The excitation circuit 5 supplies an excitation current I to the excitation coils 3 and 4 in synchronization with a control signal P4 from the control circuit 6.

The electromotive force generated at the electrodes 2a and 2b is input to a differential amplifier 7. The output e1 of the differential amplifier 7 is the output e1 of the adder circuit 8.
The amplification degree A is connected to the output side of the adder circuit 8.
The input side of an amplifier 9 having . Two sample eye-hold circuits 14 on the output side of the amplifier path
．． 15 are connected in parallel. The sample idle and hold circuit 14 includes a switch S2 operated by a control signal P2 from the control circuit 6 and a capacitor C2.
The sample idle circuit 15 includes a switch S3 and a capacitor C operated by a control signal P3 from the control circuit 6.
Consists of 3.

Capacitor C7 after switch 52 and S3 are opened
To prevent discharge of L and C, an impedance converter is generally connected after the capacitors C2 and C3, but 2 is omitted here for simplicity.

The output sides of the two sample eye-hold circuits 14 and 15 are connected to the two input sides of the subtraction circuit 16, and the output side of the σ divisor circuit 16 is connected to the sample eye-hold circuit 1.
A signal v3 is generated which corresponds to the value v1 stored at 4.15 and the VzO difference. The output signal V3 is a measurement signal proportional to the flow velocity in the tube 1.

A switch S1 operated by a control signal P1 from the control circuit 6 is connected to the output example of the amplifier 9, and a resistor R
1, an integrating circuit 11 consisting of a capacitor C1 and an operational amplifier 10 is connected. The output side of the integrating circuit 11 is connected to the second input side of the adding circuit 8. The adder circuit 8, the amplifier 9, the switch S1, and the integrator 11 constitute a compensation circuit 12 that removes the DC offset voltage. Further, in the above configuration, a switch S1 is inserted between the input resistor R1 of the amplifier 9 and the integrator 11 in the conventional circuit shown in FIG.

By opening the switch S1 only during the period T1 during which spike noise occurs, the occurrence of operations due to integration of spike noise is prevented.

FIG. 2 shows waveforms at various parts in FIG. 1. el is an input signal in which spike noise is superimposed on the flow velocity signal.

C4 shows the output waveform.

First, an ideal case with only the flow velocity signal els will be described assuming that the operation starts from time L1.

During period T1, switch S1 is open and integrator 1
Since the output e2S of the amplifier 9 also maintains a constant value (initial value), the output e4s of the amplifier 9 also maintains a constant value. When the switch S1 is closed at time t2, the integrator 11 outputs a positive output e.
4s with a time constant of (:1-11+) to obtain the compensation voltage e2
Add s to adder 8. C2s increases in the negative direction over the period TZ, and accordingly, the output e4s of the amplifier 9 also has a waveform that slowly approaches zero. els at time t5
changes to negative polarity and switch S1 is opened, so ?
! The value e2s(tj) of the lightning compensation voltage time t3 is held.

Therefore, the variation width of the output e4s of the amplifier 9 at time t3 is equal to the variation width of els multiplied by A.

During period T3, switch S1 is open, so the output of integrator 1j maintains a constant value e2s(t3), and C4
s also becomes a constant value. When the switch S1 is closed at time ℃4, the integrator 11 converts the negative polarity output voltage e4s to CI·Ih.
The compensating voltage 2s is added to the adder 8 by integrating with a time constant of .

As a result of integrating C4s during the period T4, the compensation voltage e2s changes from the negative polarity C2s (tx) to e, +5(t5>), which has the opposite polarity to els.
4s has a waveform that slowly approaches zero throughout the period T2.

The above is an explanation of only the flow velocity signal, but if el includes a DC component, the integrator 11 integrates the DC component included as the output e4 of the amplifier 9 and sends the adder 8 a compensation voltage for the DC component. By adding e2, it has a function of removing the DC component included in the output e4.

This is exactly the same as the circuit shown in FIG.

Next, a case where only the spike noise subtraction N is input will be explained. During the period T1, when the spike noise eIN is input, the switch S1 is open, so the integrator 11
does not perform an integral operation and the output 82N is held at zero value. Therefore, the output voltage of the amplifier 9 has a value equal to 64N=A-elN. At the moment when the input CJN becomes a zero value at time t2, the output voltage e2N of the integrator 11 also holds a zero value, so the two inputs of the adder 8 both have a zero value. Therefore, the output voltage 64N of the amplifier 9 has a zero value. When the switch s1 is closed in this state, the integrator 11 integrates a zero value, and the output e2N remains zero. In other words, since the switch S+ is open during the period T1, the compensation circuit does not operate and differentiates the spike noise elN, and no operation occurs during the period TZ.The second half of the cycle, T3, is exactly the same as the first half. In operation, negative polarity noise CI
N is simply multiplied by A and becomes 64N. During the period T4, the output e4N becomes a zero value and no operation occurs.

As mentioned above, the period T1 or T3 in which the noise elN exists
Since the switch S1 is open, the integrator 11 does not operate, and the output is always held at zero.

Next, consider a general case where spike noise elN is superimposed on the flow velocity signal els.

The output e4 at this time is shown as 04 in FIG. 2 by simply adding and combining the e4s and C4N. This synthesized output e4 has a period T2. It is clear from the explanation of C4N that there is no operation in T4.

By opening the switch S1 only during the period when spike noise occurs as described above, no operational phenomenon occurs and accurate measurement without zero point error is possible.

In the explanation of FIG. 2, the period T1 during which the switch S1 is open is
Although T3 and T3 are made to coincide with the period in which spike noise occurs, this is for ease of explanation and is not limited thereto. In Figure 2, the spike noise is expressed as a rectangular wave, but generally the waveform is a triangular wave with a sharp peak accompanied by ringing, etc., and since the point at which the noise disappears is not clear, the open period T1 and T2 of the switch S1 is It is desirable to make it a little larger.

Up to now, we have explained the timing of the excitation current using two values, positive and negative.
As shown in the figure.

3 in FIG. 3 is a waveform of an excitation current with a pause period, and et is an input signal waveform in which spike noise occurs every time the excitation is switched. A timing signal for controlling the switch S1 in FIG. 1 corresponding to this input signal e1 is indicated by Pl.
Although the number of times spike noise occurs increases as shown in FIG. 3, the purpose can be achieved by opening the switch S1 only during the period when noise occurs.

〔Effect of the invention〕

According to this invention, even when random noise such as high-speed fluid noise is superimposed, variations in the output signal are not amplified, and
Moreover, it is not adversely affected by the spike noise that occurs when switching the excitation current (and it is possible to compensate and eliminate the DC offset voltage. Therefore, it can be used in low power consumption electromagnetic flowmeters over a wide measurement range. This is effective because an electromagnetic flowmeter with a stable zero point can be realized.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a timing chart of the embodiment of Fig. 1, Fig. 3 is a timing chart of another embodiment, and Figs. 4 and 6 are conventional techniques. The block diagrams of FIGS. 5 and 7 show timing charts for the devices of FIGS. 4 and 6, respectively. 3.4... Excitation coil, 9... Amplifier, 11...
Integrating circuit, 12... Compensation circuit, SI... Switch that opens the input of the integrating circuit, C1... Capacitor, R1
···resistance

Claims (1)

[Claims] 1. An electromagnetic flowmeter that measures the flow rate of fluid using a DC magnetic field whose steady-state value changes periodically, which integrates the output voltage of an amplifier using an integrating circuit, and calculates the output voltage of the integrating circuit. In a compensation circuit that compensates for the DC offset voltage superimposed on the signal voltage by feeding back the opposite polarity to the input side of the amplifier, the time period during which spike noise occurs due to the switching of the steady-state value of the DC magnetic field. period including,
An electromagnetic flowmeter characterized in that the input of the integrating circuit is opened. 2. The electromagnetic flowmeter according to claim 1, in which the input of the integrating circuit is opened only during the period when spike noise occurs. 3. Claim 1 in which the time constants C_1 and R_1 of the integrating circuit are set to be longer than the period in which the steady value of the DC magnetic field changes.
The electromagnetic flowmeter according to item 1 or 2. 4. The signal voltage generated at the electrode is set at a flow rate of 1 m/Sec. The electromagnetic flowmeter according to claim 1, 2, or 3, wherein the electromagnetic flowmeter has a voltage of about 10 μV or less. 5. An electromagnetic flowmeter according to claim 1, 2, 3, or 4, which uses a battery as a power source.