Abstrict An electromagnetic flow meter which calculates the electric conductivity
of a fluid based on the ratio of an output signal generated when
electrodes are grounded through short-circuit resistors Rs.sub.1
Rs.sub.2 to an output signal generated when the electrodes are not
grounded, and more accurately corrects a flow measurement using
the electric conductivity, characterized in that, for reducing measurement
errors due to fluctuation in DC offset voltage accompanied by an
on/off operation of the short-circuit switch S1 outputs are measured
when the short-circuit switch S1 is on and off in a magnetization
state, the output are next measured when the short-circuit switch
S1 is on and off in a non-magnetization stage, and the outputs in
the non-magnetization state are subtracted from the outputs in the
magnetization state, thereby removing the influence of "DC
offset voltage fluctuation" occurring accompanied by the on/off
operation of the short-circuit switch S1.
Claims I claim:
1. A electromagnetic flow meter comprising:
a pair of electrodes positioned opposite to each other inside a
pipe through which a fluid pass;
an earth electrode for grounding said pipe;
a short-circuit switch disposed between said pair of electrodes
and said earth electrode for performing an on/off operation;
an amplifier for amplifying an electric signal from said pair of
electrodes;
means for generating a first output as a flow signal when said
short-circuit switch is turned off to operate said amplifier in
a high input impedance state;
means for generating a second output when said short-circuit switch
is turned on to operate said amplifier in a low input impedance
state; and
processing means for calculating the electric conductivity of said
fluid based on the ratio of said first output to said second output,
characterized by generating the outputs when said short-circuit
switch is on and off in a magnetization state, generating the outputs
when said short-circuit switch is on and off in a non-magnetization
state, and subtracting the outputs in the non-magnetization state
from the outputs in the magnetization state to remove the influence
of fluctuation in DC offset voltage occurring to the accompaniment
of an on/off operation of said short-circuit switch.
2. An electromagnetic flow meter according to claim 1 characterized
in that a signal in one cycle after said short-circuit switch is
turned on or off is not used, and a signal in the second cycle is
sampled as an output signal.
Description BACKGROUND OF THE INVENTION
The present invention relates generally to an electromagnetic flow
meter, and more particularly to an electromagnetic flow meter which
removes the influence of fluctuation in a direct-current (DC) offset
voltage.
The applicant has disclosed in JP-A-5-312610 improvements on an
electromagnetic flow meter proposed in International Publication
No. WO93-05367 wherein an electromagnetic flow meter having a short-circuit
switch, for measuring electric conductivity, which intermittently
connects electrodes for inducing a flow signal to a ground through
short-circuit resistors, and an offset compensation circuit for
removing a DC component in the flow signal, characterized in that
the output of the offset compensation circuit immediately after
the short-circuit switch for measuring electric conductivity is
turned on and off is prohibited from being sampled as the flow signal.
The above-mentioned electromagnetic flow meter was intended to
improve the accuracy of flow measurements by preventing an error
due to a change in DC offset voltage component occurring when the
short-circuit switch for measuring electric conductivity is turned
on and off from introducing into a signal indicative of the result
of a flow measurement.
FIG. 1 is a circuit diagram illustrating the electromagnetic flow
meter of the prior art. The electromagnetic flow meter comprises
a cylindrical flow pipe through which a fluid passes; electrodes
2a, 2b mounted on the inner wall of the flow pipe at opposite positions
to each other for inducing a flow signal; and an earth electrode
2c, wherein the flow pipe 1 the electrodes 2a, 2b, the earth electrode
2c, and a magnetizing coil, not shown, constitute a detector of
the electromagnetic flow meter which is well known in the art.
A short circuit used for measuring electric conductivity is composed
of a series circuit comprising a short-circuit resistor R.sub.S1
connected between the electrode 2a and the earth electrode 2c and
the short-circuit switch S1 for measuring electric conductivity
and a series circuit comprising a short-circuit resistor R.sub.S2
connected between the electrode 2b and the earth electrode 2c and
the short-circuit switch S1 for measuring electric conductivity.
A pre-amplifier A1 amplifies a voltage induced across the electrodes
2a, 2b. An offset compensation circuit 5 is composed of an inverting
amplifier 51 with a gain of -G comprising resistors Ra, Rb and an
operational amplifier OP1 connected as illustrated; an integrator
52 having a resistor R1 a capacitor C1 and an operational amplifier
OP2 connected as illustrated; and a switch S2.
The switch S2 and the short-circuit switch S1 are turned on and
off in synchronism with a magnetizing current having a square waveform
which is generated by a magnetizing coil, not shown.
A signal induced across the electrodes 2a, 2b, after amplified
by the pre-amplifier A1 is supplied to the offset compensation
circuit 5 for removing "fluctuation in DC voltage". The
offset compensation circuit 5 is essential because the electronic
circuit is more likely to be saturated when the DC voltage is larger.
When the short-circuit switch S1 in FIG. 1 is turned on and off
at the timing shown in a waveform chart of FIG. 2-(B), a change
in the offset voltage occurs between the electrodes 2a, 2b as illustrated
in the waveform of FIG. 2-(A). The influence of level differences
indicated by labels .delta.2 and .delta.1 in FIG. 2-(A) cannot be
removed by the offset compensation circuit 5 in FIG. 1.
To solve this problem, the prior art technique disclosed in JP-A-5-312610
does not sample (in other words, used) the output of the offset
compensation circuit as a flow signal immediately after the short-circuit
switch S1 is turned off and turned on in order to avoid the adverse
influence of the level differences.
This prior art technique can remove the influence of the level
differences .delta.2 and .delta.1 as illustrated in FIG. 2-(A).
In addition, the prior art technique can also remove the influence
of "DC offset voltage fluctuation", if it is regarded
as linear, as illustrated in FIG. 2-(A).
In the prior art technique mentioned above, the DC offset voltage,
after the short-circuit switch S1 is turned off, fluctuates in such
a manner that a level difference .delta.2 occurs, and thereafter
the offset voltage continues to increase exponentially, not linearly,
as illustrated in FIG. 3-(A). It takes a predetermined time, for
example, 30 seconds or more after the short-circuit switch S1 has
been turned off that the increase in the offset voltage can be practically
regarded as linear.
Also, the magnitude of the "DC offset voltage fluctuation"
changes over time. Specifically, in addition to the level difference
.delta.2 which becomes larger each time the short-circuit switch
S1 is turned off, the exponential increase in the offset voltage
also becomes larger during a later off period of the short-circuit
switch S1. In other words, the level difference .delta.2 becomes
larger immediately after the short-circuit switch S1 is turned off,
and the gradient of the exponential increase of the offset voltage
also tends to increase from a curve .alpha. to a curve .beta., as
illustrated in FIG. 4 due to the influences of aging changes and
foreign substances possibly attached on the electrodes 2a, 2b, and
so on.
As illustrated in FIG. 3 after the short-circuit switch S1 is
turned on, a level difference .delta.1 occurs in the offset voltage,
and the offset voltage tends to substantially converge to a fixed
value after it presents a slight decrease for a short time period.
In the prior art technique as described above, the offset compensation
circuit 5 can remove the adverse influence of a linearly increasing
DC offset voltage, but not the adverse influence of the DC offset
voltage which continuously presents an exponentially change.
The reason for the inability of the offset compensation circuit
5 to remove the adverse influence of the exponentially changing
DC offset voltage will be described in detail with reference to
FIG. 1 and waveform charts of FIG. 5.
FIG. 5-(A) illustrates an ideal square-wave input signal supplied
to the offset compensation circuit 5; FIG. 5-(B) illustrates the
timing of turning on and off the switch S2; and FIG. 5-(C) illustrates
an output signal of the offset compensation circuit 5 when the input
signal of FIG. 5-(A) is supplied thereto.
Since the signal of FIG. 5-(A) is assumed to be free from noise
and DC offset voltage and the switch S2 is off during a period T1
a peak value E1 of the square-wave input is amplified by a predetermined
amplification ratio in the inverting amplifier 51 and outputted
from the offset compensation circuit 5 as having a level -V1 as
illustrated in FIG. 5-(C).
During a period T2 in which the switch S2 is on, the output of
the inverting. amplifier 51 is integrated by the integrator 52.
Since an output of the integrator 52 is connected to a non-inverting
input of the operational amplifier OP1 constituting the inverting
amplifier 51 a feedback operation is performed so as to lead the
output of the inverting amplifier 51 to zero. This feedback operation
is a basic compensation operation performed by the offset compensation
circuit 5. With the compensation operation, the output of the offset
compensation circuit 5 reaches V2 at time t2. The ratio of V2 to
the foregoing V1 (V2/V1) is given by:
where G=Rb/Ra.
Simultaneously with the switch S2 being turned off at time t2
the polarity of the input signal to the offset compensation circuit
5 is inverted. Similarly to the operation during the period T1
the input signal is amplified to generate an output value V3 which
is held for a period T3.
During the period T3 the switch S2 is off and the output of the
integrator 52 maintains a fixed value, so that a changing portion
of the input to the offset compensation circuit 5 at time t2 is
multiplied by -G in the inverting amplifier 51.
When the switch S2 is turned on at time t3 the compensation operation
acts to reduce the output V3 of the inverting amplifier 51 to zero,
and the output of the inverting amplifier 51 is reduced to V4 at
the end time t4 of a period T4. In this case, the ratio of V4 to
V3 (V4/V3) is given by:
During one cycle of the offset compensation circuit 5 the above
described operations are performed.
Next described is the operation of the offset compensation circuit
5 when a DC offset voltage linearly changes.
FIG. 5-(D) illustrates a DC offset voltage linearly changing in
a decreasing direction, and FIG. 5-(E) illustrates the output of
the offset compensation circuit 5 generated when the DC offset voltage
changes as illustrated in FIG. 5-(D). During the period T1 the
switch S2 is off, and a changing portion of the input waveform of
FIG. 5-(D) is multiplied by -G so that the output of the offset
compensation circuit 5 reaches Va1 at time t1. When the switch S2
is turned on at time t1 the compensation operation acts to reduce
the output to zero. At the end time t2 of the period T2 the output
is reduced to Vb1.
A decrease in the input during the next period T3 is equal to a
changing portion of the input in the minus direction during the
period T1. This changing portion is amplified by the same amplification
ratio as that in the period T1 in the inverting amplifier 51 whereby
the output of the offset compensation circuit 5 reaches Va2. In
this event, Va2=Va1 is satisfied.
During the period T4 since the switch S2 is on, the compensation
operation acts to reduce the output to zero, whereby the output
of the offset compensation circuit 5 is reduced to Vb2 at time t4.
Likewise, in this event, Vb2=Vb1 is satisfied.
A sampling circuit 61 at a subsequent stage samples an area ml
during the period T2 and an area m2 during the period T4 to generate
m.sub.1 -m.sub.2 as an output signal Vo.
In this event, since Va1=Va2 and Vb1=Vb2 are satisfied and the
output waveform of the offset compensation circuit 5 has the same
gradient during the periods T2 and T4 as illustrated in FIG. 5-(E),
m.sub.1 =m.sub.2 is satisfied, and the output signal Vo is given
by:
In this way, when the DC offset voltage linearly changes in the
decreasing direction as illustrated in FIG. 5-(D), the offset compensation
circuit 5 can remove its influence.
Next, FIG. 5-(G) illustrates an output waveform of the offset compensation
circuit 5 when a DC offset voltage exponentially changes in the
decreasing direction as illustrated in FIG. 5-(F).
During the period T1 a changing portion of the input waveform
illustrated in FIG. 5-(F) during the period T1 is multiplied by
-G in the inverting amplifier 51 and the output of the offset compensation
circuit is increased to Vc1 at time t1. When the switch S2 is on
during the next period T2 the compensation operation acts to reduce
the output of the offset compensation circuit 5 to zero, whereby
the output is reduced to Vd1 at time t2. During the next period
T3 since the switch S2 is off again, a changing portion of the
input during the period T3 is multiplied by -G, and the output of
the offset compensation circuit 5 is increased to Vc2 at time t3.
Since the change in the input is exponentially decreasing as illustrated
in FIG. 5-(F), a changing amount of the input during the period
T3 is smaller than the changing amount of the same during the period
T1 so that Vc2 is smaller than Vc1.
Also, the Vd2 compensated to approximately zero by the compensation
operation during the period T4 is of course smaller than Vd1.
The gradually decreasing areas in FIG. 5-(G) corresponding to the
periods T2 T4 are sampled by the sampling circuit 61.
Assuming that the areas are labelled Mg.sub.1 Mg.sub.2 respectively,
it is apparent that Mg.sub.1 is smaller than Mg.sub.2. The output
Vo is given by:
The output Vo is not reduced to zero.
In summary, when the input includes an exponential fluctuation
as illustrated in FIG. 5-(F), its influence cannot be removed by
the offset compensation circuit 5.
The foregoing description clearly indicates that the prior art
technique disclosed in JP-A-5-312610 cannot remove the influence
of the "DC offset voltage fluctuation".
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide an
electromagnetic flow meter which is capable of removing the influence
of an exponentially changing DC offset voltage to accomplish accurate
measurements of a flow rate and electric conductivity of a fluid.
To achieve the above object, an electromagnetic flow meter according
to a first aspect of the present invention comprises a short-circuit
switch (S1) disposed between a pair of electrodes (2a, 2b) and an
earth electrode (2c) for performing an on/off operation, wherein
an output generated when the short-circuit switch (S1) is turned
off to operate a pre-amplifier (A1) in a high input impedance state
is used as a flow signal, and the electric conductivity of a fluid
is calculated based on the ratio of an output generated when the
short-circuit switch (S1) is turned on to operate the pre-amplifier
(A1) in a low input impedance state to the output generated in the
high input impedance state,
characterized by, generating the outputs when the short-circuit
switch (S1) is on and off in a magnetization state, subsequently
generating the outputs when the short-circuit switch (S1) is on
and off in a non-magnetization state, subtracting the outputs in
the non-magnetization state from the outputs in the magnetization
state to remove the influence of fluctuation in "DC offset
voltage fluctuation" occurring to the accompaniment of an on/off
operation of the short-circuit switch (S1).
A second aspect of the present invention is characterized in that
a signal in the first cycle after the short-circuit switch S1 is
turned on or off is removed, and a signal in the second cycle is
sampled as a target signal.
A signal generated across the electrodes during a magnetization
period includes both of a flow signal and the influence of "DC
offset voltage fluctuation". On the other hand, an output generated
during a non-magnetization period only includes the influence of
"DC offset voltage fluctuation" without the flow signal.
By subtracting the output during a non-magnetization period from
the output during a magnetization period, the influence of "DC
offset voltage fluctuation" is canceled, thus making it possible
to only extract an effective signal proportional to a flow rate.
Also, even if the "DC offset voltage fluctuation" increases
over time, the outputs generated in adjacent periods with a slight
time difference therebetween are subjected to the subtraction, the
influence of the increase in the "DC offset voltage fluctuation"
can be removed.
The effective signal derived by the subtraction includes a flow
signal as a normal electromagnetic flow meter in a high input impedance
state with the short-circuit switch (S1) being off, and information
for calculating the electric conductivity of a fluid in a low input
impedance state with the short-circuit switch (S1) being on. Even
if the magnitude of the "DC offset voltage fluctuation"
increases due to the influence of foreign substances attached to
the electrodes or the like, it is possible to correctly make measurements
of not only a flow rate but also the electric conductivity of a
fluid.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram illustrating a conventional electromagnetic
flow meter;
FIG. 2 is a waveform chart showing the relationship between an
offset voltage (A) and on/off states of a short-circuit switch of
the electromagnetic flow meter illustrated in FIG. 1;
FIG. 3 is a waveform chart showing the relationship between an
offset voltage (A) and on/off states of the short-circuit switch
of the electromagnetic flow meter illustrated in FIG. 1;
FIG. 4 is a waveform chart illustrating DC offset voltage fluctuation;
FIG. 5 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 1;
FIG. 6 is a block diagram illustrating an electromagnetic flow
meter according to a first embodiment of the present invention;
FIG. 7 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 6;
FIG. 8 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 6;
FIG. 9 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 6;
FIG. 10 is a block diagram illustrating an electromagnetic flow
meter according to another embodiment of the present invention;
FIG. 11 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 10;
FIG. 12 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 10; and
FIG. 13 is a waveform chart showing the operation timing of the
electromagnetic flow meter illustrated in FIG. 10.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 6 illustrates an electromagnetic flow meter according to the
preferred embodiment of the present invention. The electromagnetic
flow meter comprises a cylindrical flow pipe 1 through which a fluid
passes; a pair of electrodes 2a, 2b mounted on the inner wall of
the flow pipe 1 at positions opposite to each other for extracting
an induced flow signal; an earth electrode 2c for grounding a fluid
under measurement, which is connected to a common potential of a
convertor in the electromagnetic flow meter; a magnetizing coil
3; and a magnetizing circuit 4 for supplying the magnetizing coil
3 with a magnetizing current.
The electrodes 2a, 2b are connected to differential inputs of a
pre-amplifier A1 having a high input impedance, and are connected
to and disconnected from the common potential through a series connection
of a short-circuit resistor R.sub.S1 and a short-circuit switch
S1 and a series connection of a short-circuit resistor R.sub.S2
and the short-circuit switch S1 respectively.
An output E1 of the pre-amplifier A1 after a direct-current (DC)
offset voltage is removed therefrom by an offset compensation circuit
5 is sampled by a sampling circuit 6 which integrates an input
signal only during a predetermined period, and connected to four
holding circuits 11 12 13 14 for holding a sampled output every
sampling period.
The offset compensation circuit 5 is composed of an inverting amplifier
51 having resistors Ra, Rb and an operational amplifier OP1 connected
as illustrated; an integrator 52 having a resistor R1 a capacitor
C1 and an operational amplifier OP2 connected as illustrated; and
a switch S2 operated in synchronism with the short-circuit switch
S1 connected as illustrated.
The sampling circuit 6 is composed of an inverting amplifier A2
having a gain of -1 switches S3 S4 S5 a capacitor C3 an operational
amplifier OP3 and two resistors, not labelled, connected as illustrated.
The holding circuit 11 is composed of a switch P1 a capacitor
C3 and an amplifier, not labelled, connected as illustrated. The
holding circuit 12 is composed of a switch P2 a capacitor C4 and
an amplifier, not labelled, connected as illustrated. The holding
circuit 13 is composed of a switch P3 a capacitor C5 and an amplifier,
not labelled, connected as illustrated. The holding circuit 14 is
composed of a switch P4 a capacitor C6 and an amplifier, not labelled,
connected as illustrated.
A differential amplifier A3 is supplied with outputs of the holding
circuit 11 and the holding circuit 12 and outputs a flow signal
Vo.sub.1. A differential amplifier A4 is supplied with outputs of
the holding circuit 13 and the holding circuit 14 and delivers
an output Vo.sub.2 when the short-circuit switch S1 is turned on.
A processing circuit 8 calculates the electric conductivity of
a fluid under measurement from the ratio of the flow signal Vo.sub.1
which is the output of the differential amplifier A3 to the output
Vo.sub.2 of the differential amplifier A4 and outputs the calculated
electric conductivity.
A timing circuit 7 supplies the magnetizing circuit 4 with a timing
signal, turns on and off the short-circuit switch S1 and the switches
S2 S3 S4 S5 at predetermined timing, respectively, and turns
on the respective switches P1-P4 of the four holding circuits 11-14
at predetermined timing to hold the output of the sampling circuit
6 on the capacitors C3-C6 respectively.
The amplifiers, not labelled, connected to the capacitors C3-C6
constituting the respective holding circuits 11-14 for holding the
output of the sampling circuit 6 are buffer amplifiers having a
high input impedance.
FIGS. 7-9 are waveform and timing charts showing the operations
of the respective circuits in the electromagnetic flow meter of
FIG. 6.
FIG. 7-(A) is a waveform representing a magnetic field produced
by a magnetizing signal, wherein two cycles of square-wave magnetization
are applied in a period Ta, followed by a non-magnetization period
Tb having the same length (time) as a period Ta.
The period Ta consists of a total of two cycles of magnetization
periods, i.e., a cycle Ta1 in which the short-circuit switch S1
is in off state and a cycle Ta2 in which the short-circuit switch
S1 is in on state (FIG. 9).
The period Tb differs from the period Ta only in that the magnetizing
magnetic field is removed, with other conditions in the period Tb
equal to those in the period Ta. Specifically, the short-circuit
switch S1 is off during a first cycle Tb1 of the magnetization and
on during a second cycle Tb2 of the magnetization, subsequent to
the cycle Tb1.
In this embodiment, the cycles Ta1 Ta2 Tb1 Tb2 are all determined
to have the same length.
A signal proportional to a flow velocity of a fluid is generated
across the electrodes 2a, 2b in FIG. 6 by the action of the magnetizing
magnetic field as represented by FIG. 7-(A), and amplified by the
pre-amplifier A1 to produce the output E1 illustrated in FIG. 7-(B).
During the period Ta1 the short-circuit switch S1 is off so that
the pre-amplifier A1 has a high input impedance. The value El.sub.OFF
of the output E1 generated during the period Ta1 is a signal proportional
to a flow rate.
During the period Ta2 the short-circuit switch S1 is on, and the
inputs to the pre-amplifier A1 are grounded through the short-circuit
resistors R.sub.S1 R.sub.S2 to present a low input impedance, so
that the pre-amplifier A1 generates the output E1 having the value
E1.sub.ON which is smaller than the value E1.sub.OFF generated during
the period Ta1.
One object of the present invention is to accurately calculate
the electric conductivity of a fluid under measurement from the
ratio of the outputs E1.sub.OFF and E1.sub.ON.
The above described is the operation of the electromagnetic flow
meter performed during the former period Ta. Since a non-magnetization
condition is present during the latter period Tb, the signal proportional
to a flow rate is not generated.
FIG. 7-(B) illustrates the flow proportional signal generated across
the electrodes 2a, 2b, i.e., the output signal of the pre-amplifier
A1 during the periods Ta, Tb. A solid line in FIG. 7-(C) represents
"DC offset voltage fluctuation" V.sub.DC generated by
turning on and off the short-circuit switch S1.
The "DC offset voltage fluctuation" V.sub.DC presents
completely the same waveform during the magnetization period Ta
and during the non-magnetization period Tb. This is because the
"DC offset voltage fluctuation" V.sub.DC results only
from an electrochemical balance between the electrodes 2a, 2b and
the earth electrode 2c and a fluid by the on/off operation of the
short-circuit switch S1 and is not related to the presence or absence
of the magnetizing magnetic field.
A broken line in the period Ta in FIG. 7-(C) illustrates the DC
offset voltage fluctuation V.sub.DC multiplexed with the flow signal
of FIG. 7-(B). The signal actually generated across the electrodes
2a, 2b. i.e., the output of the pre-amplifier A1 has a waveform
represented by the broken line in the period Ta and the solid line
in the period Tb in FIG. 7-(C).
FIG. 7-(D) illustrates an output waveform of the offset compensation
circuit 5 in FIG. 6 which is generated when the flow proportional
signal of FIG. 7-(B) is supplied thereto.
The offset compensation circuit 5 of FIG. 6 has the same configuration
as the prior art offset compensation circuit 5 described with reference
to FIG. 1. As is also described above, since the switch S2 is off
during the period T1 as illustrated in FIG. 9 when the switch S2
is turned on at the beginning of the period T2 after an input to
the offset compensation circuit 5 is amplified by -G, a compensation
operation acts to lead the output of the offset compensation circuit
5 to zero. Subsequently during periods T3-T8 the input is amplified
as previously described.
A solid line in FIG. 8-(E) represents the output of the offset
compensation circuit 5 generated when the "DC offset voltage
fluctuation" represented by the solid line in FIG. 7-(C) is
applied thereto.
During the period T1 in which the switch S2 is off, input fluctuation
in the positive direction represented by the solid line in FIG.
7-(C) is multiplied by -G. Since the switch S2 is on during the
next period T2 the compensation operation acts to lead the output
of the offset compensation circuit 5 to zero. However, since the
input represented by the solid line in FIG. 7-(C) continues to exponentially
increase through the period T2 the output is led to zero at a slightly
lower speed. Since the switch S2 is off again during the next period
T3 an increase of the input in the positive direction only is multiplied
by -G, so that the output of the offset compensation circuit 5 increases
in the negative direction. Then, when the switch S2 is on again
during a period T4 the compensation operation acts to lead the
output to zero.
At the beginning time t4 of a period T5 the input represented
by the solid line in FIG. 7-(C) changes with an instantaneous level
difference in the negative direction, and then exponentially increases
in the negative direction over time. Since the input is multiplied
by -G, the output of the offset compensation circuit 5 presents
positive values during the period T5. Subsequently during periods
T6-T8 the offset compensation circuit 5 performs similar operations
to those during the periods T2-T4.
Since the "DC offset voltage fluctuation" has the same
waveform in the non-magnetization period Tb as in the period Ta,
the output waveform of the offset compensation circuit 5 in the
period Tb is also the same as the output waveform in the period
Ta.
A broken line shown only in the period Ta in FIG. 8-(E) represents
the output waveform multiplexed with the output waveform of. FIG.
7-(D) corresponding to the flow proportional signal. Therefore,
the broken line in the period Ta and the solid line in the period
Tb in FIG. 8-(E) represent the actual output waveform of the offset
compensation circuit 5 which indicates that the DC offset voltage
is multiplexed with the flow proportional signal.
FIG. 8-(F) illustrates an output waveform of the sampling circuit
6 in FIG. 6 which is generated when the flow signal of FIG. 7-(D)
is supplied thereto. In this example, a sampling period of the sampling
circuit 6 is assumed to have the same time length as the on-period
of the switch S2. The flow proportional signal of FIG. 7-(D) passes
through the inverting amplifier A2 having a gain of -1 and is integrated
while the switch s4 is on during the period T2 as illustrated in
the timing chart of FIG. 9. Thereafter, while the switch S3 is on
during the period T4 as illustrated in the timing chart of FIG.
9 the flow proportional signal is directly integrated without passing
through the amplifier A2. In this way, the sampling circuit 6 generates
an output ES1 at the end of the period Ta1 (FIG. 8-(F)).
Similarly, the flow proportional signal passes through the inverting
amplifier A2 and integrated during the period T6 in which the switch
S4 is on as illustrated in the timing chart of FIG. 9 and subsequently
integrated during the period T8 in which the switch S3 is on as
illustrated in the timing chart of FIG. 9. In this way, the sampling
circuit 6 generates an output Es2 at the end of the period Ta2 (FIG.
8-(F)).
The outputs Es.sub.1 and Es.sub.2 are outputs proportional to the
flow signal. On the other hand, the flow signal presents a zero-value
during non-magnetization periods Tb1 Tb2.
FIG. 8-(G) illustrates the output of the sampling circuit 6 corresponding
to the "DC offset voltage fluctuation" V.sub.DC represented
by the solid line in FIG. 8-(E). The "DC offset voltage fluctuation
signal"V.sub.DC is sampled similarly to the flow signal. The
sampling circuit 6 generates outputs E.sub.D c1 E.sub.D c2 corresponding
to the periods Ta1 Ta2 respectively, and also generates the same
outputs E.sub.D c1 E.sub.D c2 corresponding to the periods Tb1
Tb2 in the non-magnetization period Tb as the outputs corresponding
to the periods Ta1 Ta2. The respective outputs Es.sub.1 Es.sub.2
E.sub.D c1 E.sub.D c2 are reset respectively at the timing at which
the switch S5 is turned on as illustrated in the timing chart of
FIG. 9.
The actual output of the sampling circuit 6 corresponding to the
period Ta1 is the sum of Es.sub.1 and E.sub.D c1 which is held in
the holding circuit 11 in FIG. 6 at the timing of the switch P1
illustrated in FIG. 9. The actual output of the sampling circuit
6 corresponding to the period Tb1 is only E.sub.D c1 which is held
in the holding circuit 12 at the timing of the switch P2 illustrated
in FIG. 9. Subsequently, the differential amplifier A3 calculates
the following subtraction:
wherein Es.sub.1 +E.sub.D c1 is the output of the sampling circuit
6 in the period Ta1 and E.sub.D c1 is the output in the period
Tb1.
This subtraction removes the influence of the "DC offset voltage
fluctuation" to only leave the net flow proportional signal
Es1.
The flow proportional signal Es.sub.1 i.e., Vo.sub.1 is the output
signal of the electromagnetic flow meter when the short-circuit
switch S1 is in off state, and is the same as an output signal of
a common electromagnetic flow meter which does not have the short-circuit
switch S1 and the short-circuit resistors Rs.sub.1 Rs.sub.2.
The output signal of the sampling circuit 6 corresponding to the
period Ta2 is Es.sub.2 +E.sub.D c2 which is held in the holding
circuit 13 at the timing of the switch P3 illustrated in FIG. 9.
The output signal of the sampling circuit 6 corresponding to the
period Tb2 is only E.sub.D c2 which is held in the holding circuit
14 at the timing of the switch P4 illustrated in FIG. 9. Then, the
differential amplifier A4 calculates the following subtraction similarly
to the above:
where Es.sub.2 +E.sub.D c2 is the output of the sampling circuit
6 in the period Ta2 and E.sub.D c2 is the output in the period
Tb2.
In this way, the influence of the "DC offset voltage fluctuation"
is removed to only leave the flow proportional signal Es.sub.2.
This flow proportional signal Es.sub.2 i.e., Vo.sub.2 is the output
signal of the sampling circuit 6 when the short-circuit switch S1
is on, i.e., when the pre-amplifier A1 presents a low input impedance,
and serves as information for calculating the electric conductivity
of a fluid under measurement (hereinafter referred to as the "electric
conductivity information").
The output Vo.sub.1 (=Es.sub.1) of the differential amplifier A3
may be converted into a current output ranging from 4 to 20 mA by
a voltage-to-current convertor, not shown, or displayed as an accumulated
value of a flow under measurement by an accumulation display.
The output Vo.sub.2 is utilized to calculate the electric conductivity
of the fluid under measurement together with the output Vo.sub.1
in the processing circuit 8. The electric conductivity of the fluid
thus calculated may be outputted to the outside as an electric conductivity
signal or utilized to correct errors possibly introduced into the
flow signal due to changes in the electric conductivity to provide
more accurate flow measurements.
It should be noted that the embodiment so far described with reference
to FIGS. 6-9 is intended to explain the basic principles of the
present invention, and several problems are still left unsolved
for practical use.
A first problem, due to the operation principle of the offset compensation
circuit, will be explained below with reference to FIG. 13.
The offset compensation circuit operates on the assumption that
a square-wave signal is continuously inputted thereto. When the
offset compensation circuit is supplied with the square-wave signal
only during two cycles Tc.sub.1 -Tc.sub.4 as illustrated in FIG.
13-(A), it generates a half output during the period Tc.sub.1 and
also generates a half output even during a period Tc'.sub.1 in which
no input signal is supplied thereto, as illustrated in FIG. 13-(B).
With this respect, the basic operation of the offset compensation
circuit centered on the period Tc.sub.2 is to add the input during
the period Tc.sub.2 and the input during the previous period Tc.sub.1
by a feedback operation of an integrator (such as 52 in FIG. 6)
with opposite polarities to each other to remove a DC offset voltage.
Therefore, since no input signal is supplied to the offset compensation
circuit in a period Tc".sub.4 previous to the period Tc.sub.1
the offset compensation circuit generates an output in the period
Tc.sub.1 which is one half of the output in the periods Tc.sub.2
Tc.sub.3. In addition, since an input of the period Tc.sub.4 is
added with the negative polarity in the period Tc'.sub.1 in which
no input signal is supplied to the offset compensation circuit,
a half output is also generated in the period Tc'.sub.1.
As a second problem associated with the "DC offset voltage
fluctuation" accompanied by the on/off operation of the short-circuit
switch S1 the level difference .delta.2 produced when the short-circuit
switch S1 is turned off and the level difference .delta.2' produced
when the short-circuit switch S1 is turned off at the next time
are not completely the same, as the prior art example illustrated
in FIG. 3-(B). Thus, each time the short-circuit switch S1 is turned
off, a least difference is found in the level difference.
Due to such least difference, the correction relying on the subtraction,
implemented in the embodiment of FIG. 6 cannot completely remove
the DC offset voltage fluctuation.
An embodiment illustrated in FIGS. 10-12 is a practical one which
can solve the two problems left unsolved in the embodiment of FIGS.
6-9 i.e., the problem due to the operation principle of the offset
compensation circuit and the problem of the "DC offset voltage
fluctuation" which varies every time the short-circuit switch
is turned off. FIG. 10 illustrates an electric circuit diagram of
an electromagnetic flow meter according to this embodiment, and
FIGS. 11-12 are waveform charts representing the operations of components
in the electromagnetic flow meter.
Since the basic configuration and operation of this embodiment
is the same as those of the foregoing embodiment described with
reference to FIGS. 6-9 different portions will be mainly explained.
FIG. 11-(A) illustrates a magnetic field produced by a magnetizing
square wave, wherein a magnetization period Ta consists of magnetization
cycles Ta1 Ta2 in which a short-circuit switch S1 is off and magnetization
cycles Ta3 Ta4 in which the short-circuit switch S1 is on, i.e.,
a total of four cycles, as illustrated in FIG. 12. A subsequent
non-magnetization period Tb has the same length as the magnetization
period Ta and consists of cycles Tb1 Tb2 in which the short-circuit
switch S1 is off and cycles Tb3 Tb4 in which the short-circuit
switch S1 is on, i.e., a total of four cycles. A full cycle of measuring
operations consists of the magnetization period Ta and the non-magnetization
period Tb. FIG. 11-(B) illustrates a flow signal generated when
the magnetic field of FIG. 11-(A) is applied. A solid line in FIG.
11-(C) shows a waveform representing "DC offset voltage fluctuation"
which occurs when the short-circuit switch S1 is turned on and off.
A broken line in the period Ta shows a waveform representing the
"DC offset voltage fluctuation" multiplexed with the flow
signal. FIGS. 11-(A), 11-(B), 11-(C) correspond to FIGS. 7-(A),
7-(B), 7(C), respectively, and have the same waveforms as those
figures.
FIG. 11-(D) illustrates an output waveform of an offset compensation
circuit 5 in FIG. 10 when the input illustrated in FIG. 11-(B) is
supplied thereto. While FIG. 11-(D) corresponds to FIG. 7-(D), FIG.
11-(D) represents the characteristic of the offset compensation
circuit with the input signal starting from a zero level, so that
the output value in the former half of the period Ta1 is one half
of the output value in the former half of the period Ta2. Also,
the output value in the former half of the period Ta3 is larger
than the output value in the former half of the period Ta4. Although
no input is supplied during the former half of the period Tb1 one
half of the output generated in the period Ta4 is present in the
former half of the period Tb1.
The reason for the foregoing operation has been previously explained
with reference to FIG. 13. An off-period and an on-period of the
short-circuit switch S1 respectively include the two cycles Ta1
Ta2 and the two cycles Ta3 Ta4 in the magnetization period Ta,
and also include the two cycles Tb1 Tb2 and the two cycles Tb3
Tb4 also in the non-magnetization period Tb, respectively, so that
even if output waveforms different from ideal ones are generated
in the former halves of the cycles Ta1 Ta3 Tb1 Tb3 ideal outputs
are generated in the second cycles Ta2 Ta4 Tb2 Tb4.
FIG. 12-(E) illustrates an output waveform of the offset compensation
circuit 5 when the DC offset voltage fluctuation illustrated in
FIG. 11-(C) is inputted thereto, wherein a solid line represents
the output waveform when only the "DC offset fluctuation"
is inputted, and a broken line represents the output waveform when
the "DC offset fluctuation" multiplexed with the flow
signal voltage is inputted, similarly to the case of FIGS. 7 8.
FIG. 12-(F) illustrates a waveform generated when the flow output
of FIG. 11-(D) is sampled by a sampling circuit 6 in FIG. 10 at
the timing of switches S4 S3 illustrated in FIG. 12.
While the operation of the sampling circuit 6 is similar to that
described with reference to FIGS. 7-9 the sampling circuit 6 neglects
the flow output of the first cycles immediately after the short-circuit
switch S1 is turned on and off, and samples the flow output of the
second cycles Ta2 Ta4 to generate a sampled value Es1 in the period
Ta2 and a sampled value Es.sub.2 in the period Ta4.
Similarly, FIG. 12-(G) illustrates a waveform generated when the
output of the offset compensation circuit 5 supplied only with the
"DC offset voltage fluctuation" represented by the solid
line in FIG. 12-(E) is sampled by the sampling circuit 6 at the
timing of switches S4 S3 where sampled values E.sub.D c1 and E.sub.D
c2 are generated in the periods Ta2 Ta4 respectively.
Also in the non-magnetization period Tb, the sampling circuit 6
neglects the flow output of the first cycles immediately after the
short-circuit switch S1 is turned on and off, and samples the flow
output of the second cycles Tb2 Tb4 to generate sampled values
E.sub.D c1 E.sub.D c2 in the periods Tb2 Tb4 identical to the
sampled values in the period Ta2 Ta4 respectively.
In the embodiment illustrated in FIG. 10 the held values Es1
Es.sub.2 and E.sub.D c1 E.sub.D c2 indicated by hatched portions
in FIG. 12-(F) and FIG. 12-(G) are analog-to-digital (A/D) converted
by an A/D convertor 10 which is operated at the timing of an A/D
conversion control signal P10. Then, the A/D converted values are
fetched in MPU 9.
The MPU 9 internally performs the following calculations:
wherein Es.sub.1 +E.sub.D c1 is an output in the period Ta2 E.sub.D
c1 an output in the period Tb2 Es.sub.2 +E.sub.D c2 an output in
the period Ta4 and E.sub.D c2 an output in the period Tb4.
In the above equations, Vo.sub.1 is an output signal proportional
to the flow rate, and Vo.sub.2 is a value measured when the short-circuit
switch S1 is on, i.e., when the pre-amplifier A1 presents a low
input impedance, and serves as information for calculating the electric
conductivity of a fluid under measurement. The MPU 9 calculates
the electric conductivity of the fluid from the ratio of Vo.sub.2
to Vo.sub.1.
In the embodiment illustrated in FIGS. 10-12 the input signals
in the first cycles immediately after the short-circuit switch S1
is turned on and off are not employed, and the input signals in
the second cycles are sampled. Therefore, even if the level difference
.delta.2 and the next level difference .delta.2' do not match completely
as illustrated in FIG. 3-(B), the measurement is not affected by
variations in the level difference.
Since the electromagnetic flow meter of the present invention is
configured as described above, the electric conductivity of a fluid
can be correctly measured even if the magnitude of "DC offset
voltage fluctuation" due to the on/off operation of the short-circuit
switch (S1) varies by the action of foreign substances attached
on the electrodes or the like.
The present invention can also improve the accuracy of correcting
the influence of changes in the electric conductivity of a fluid
exerting on flow measurements.
Further, since the input signals in the first cycles immediately
after the short-circuit switch S1 is turned on and off are not employed
and the input signals in the second cycles are sampled, even if
a level difference .delta.2 and the next level difference .delta.2'
does not match completely as illustrated in FIG. 3-(B), the measurement
is not affected by variations in the level difference.
|