Abstrict A flow meter comprises electrodes (16 18), induction windings
(12 14), a winding driver circuit (20), and a processor circuit
(22). The winding excitation current is a combination of a current
of period T.sub.1 and a current of period T.sub.2 where T.sub.2
>T.sub.1. The processor circuit (22) processes the voltage signal
sent by the electrodes and generates intermediate measurement signals
of periods T.sub.1 and T.sub.2. These signals are then used to generate
a correction signal for removing the relaxation effect from the
measurement signal S.
Claims We claim:
1. An electromagnetic flow meter comprising:
a measuring tube for passing the liquid whose flow rate is to be
measured;
induction windings have an excitation current passing therethrough
for setting up a magnetic field perpendicular to the direction of
liquid flow;
diametrically-opposed electrodes for sensing the voltage signal
which results from the flow of liquid through the magnetic field;
a circuit for feeding current to said induction windings and comprising
means for generating a periodic excitation current of period T,
each period T being the result of a combination in time of n periods
of a first excitation current (I.sub.1) of period T.sub.1 and m
period of a second excitation current (I.sub.2) of period T.sub.2
where T.sub.2 >T.sub.1 with the period T.sub.2 being long enough
to ensure that portion of the corresponding sensed voltage signal
is free from the relaxation phenomenon; and
a processor circuit comprising means for generating a first intermediate
measurement signal (M.sub.1) of period T.sub.1 and having a short
response time, a second intermediate measurement signal (M.sub.2)
corresponding to the voltage signal of period T.sub.2 and being
free from measurement drift, and for combining and filtering said
first and second intermediate measurement signals in order to generate
a measurement signal S being both short in response time and free
from drift.
2. A flow meter according to claim 1 wherein the first and second
excitation currents are currents in which the periods are constituted
by alternating positive-going and negative-going pulses, each of
duration T.sub.1 /2 or T.sub.2 /2 as the case maybe, and wherein
said combination consists in alternating n periods of the first
current with m periods of the second current so as to obtain the
relationship:
3. A flow meter according to claim 2 wherein m is equal to 1
the period T.sub.1 lies in the range 100 milliseconds to 5 milliseconds,
and the period T.sub.2 lies in the range 50 milliseconds to 1 second.
4. A flow meter according to claim 1 wherein the first excitation
current is a sinewave current and wherein each period of said second
excitation current consists in a positive-going half cycle and a
negative-going half cycle of duration t'=T.sub.2 /2 said combination
consisting of superposing said first and second currents to obtain
the relationship:
5. A flow meter according to claim 4 wherein the period T.sub.2
lies in the range 50 milliseconds to 1 second, and wherein the period
T.sub.1 lies in the range 100 milliseconds and 5 milliseconds.
6. A flow meter according to claim 1 wherein said processor circuit
comprises means for generating a correction signal based on said
intermediate measurement signals M.sub.1 and M.sub.2 said correction
signal being substantially free from noise, together with means
for generating said measurement signal s from said first intermediate
measurement signal M.sub.1 and said correction signal, said measurement
signal being substantially free from drift.
7. A flow meter according to claim 6 wherein said correction signal
which is substantially free from noise is representative of the
relaxation error measured at predetermined instants relative to
the voltage signal of period T.sub.1 and said first intermediate
measurement signal M.sub.1 is representative of said voltage signal
of period T.sub.1 at said predetermined instants.
8. A flow meter according to claim 1 wherein said processor circuit
comprises means for filtering the first intermediate measurement
signal M.sub.1 with a response time t.sub.1 means for generating
a correction signal equal to the difference between the first and
second intermediate measurement signals M.sub.1 and M.sub.2 means
for filtering said difference signal with a response time of t.sub.2
where t.sub.2 is long enough to ensure that the filtered correction
signal is substantially free from noise, and means for subtracting
said filtered difference from the filtered first intermediate signal,
thereby obtaining the measurement signal (S) having a response time
t.sub.1 and being substantially free from drift.
9. A flow meter according to claim 1 wherein said processor circuit
comprises means for filtering both intermediate measurement signals
M.sub.1 and M.sub.2 with a response time t.sub.2 and for generating
a correction signal equal to the difference between said filtered
signals, where t.sub.2 is sufficiently long to ensure that said
filtered signals are substantially free from noise, means for filtering
the first intermediate measurement signal M.sub.1 with a response
time t.sub.1 means for generating the difference between the correction
signal and said first intermediate measurement signal M.sub.1 as
filtered with a response time of t.sub.1 thereby ensuring that
the measurement signal S has a response time of t.sub.1 and is substantially
free from drift.
Description The present invention relates to an electromagnetic flow meter,
and more particularly to a circuit for generating the magnetic field
and for processing the voltage sensed by the flow meter electrodes.
BACKGROUND OF THE INVENTION
In an electromagnetic flow meter of the type described in French
Pat. No. 2 405 466 the liquid whose flow rate is to be measured
flows along a portion of duct which is fitted with induction windings
to set up a magnetic field perpendicularly to a portion of duct,
and a pair of diametrically opposed electrodes for sensing the voltage
which results from the liquid flowing through the magnetic field.
This voltage is proportional to the average speed of the liquid
and thus to its average flow rate. The voltage can be processed
to obtain an electric signal representative of the liquid flow rate.
The magnetic field may be set up either by means of a sinewave
current or else by means of a pulsed direct current.
When using a sinewave current, the frequency of the signal is generally
about 30 Hz to about 50 Hz. The signal sensed on the electrodes
comprises: a flow rate signal which is in phase with the modulation
of the electric current, a parasitic induction signal which is in
quadrature with the modulation of the current; and a second parasitic
signal whose phase and amplitude vary as a function of the nature
of the fluid and the surface state of the electrodes.
Such a system suffers from drift even when the flow rate is zero.
This drift may be as much as several percent depending on the type
of fluid and on the electrodes. The zero point adjustment must be
performed on the final installation by the user. However, with this
type of drive to the magnetic field, it is easy to filter out very
low frequency noise (<10 Hz) created by particle-laden resistive
liquids and by liquids having marked acid or basic properties. Given
the frequency of the base signal (30 Hz to 50 Hz), effective filtering
can be performed with a short time constant, for example less than
three seconds. Such a time constant is compatible with the flow
meter being used as a link in a regulation servo system, for example.
When controlled by a pulsed current, the modulation frequency is
generally chosen to lie in the range 1 Hz to 10 Hz. Accompanying
FIGS. 1a and 1b show the operation of prior art flow meters using
pulsed current control.
FIG. 1a shows the intensity I of the control current (or field
strength B of the resulting magnetic field) as a function of time
t. FIG. 1b shows the corresponding voltage signal (V) sensed on
the electrodes of the flow meter as a function of time, for a given
flow rate of a given liquid.
Each pulse in the voltage signal comprises: a first zone A corresponding
to a parasitic induction effect which occurs when the direction
of the magnetic field is reversed; followed by a zone B corresponding
to a relaxation effect produced in the fluid-electrode interface
at the moment the magnetic field is reversed. As shown by the curve
in FIG. 1b, this parasitic effect wears off slowly to leave a pure
flow rate signal which corresponds to a zone C. In order to obtain
a signal representative of the flow rate, the signal must be observed
in zone C of pulse. The flow rate signal is then proportional to
U.sub.1 -U.sub.2. This system has a highly stable zero point, and
in addition, any drift from zero is self-compensating since the
measurement signal comes from the difference of two voltages, thereby
eliminating a parasitic DC voltage referred to as the asymmetry
voltage. FIG. 1b does not show the asymmetry voltage. The above-mentioned
French patent describes in detail how the asymmetry voltage may
be eliminated. In contrast, this feed system makes it difficult
to properly eliminate low frequency noise of the type present in
non-homogenous sludge. Given the frequency of the signal to be filtered
(1 Hz to 10 Hz), the time constant of the filter circuit may be
as much as several tens of seconds. This makes it very difficult,
if not impossible, to use such a flow meter in a regulation control
loop, since obtaining a noise-free measurement signal gives rise
to a response time which is too long.
Preferred embodiments of the present invention provide an electromagnetic
flow meter which combines the advantages of the two prior art principles
in order to measure flow rates without zero drift and with low noise
combined with a rapid response time. Such a flow meter is thus easier
to use in a regulation chain, for example.
SUMMARY OF THE INVENTION
The present invention provides an electromagnetic flow meter comprising:
a measuring tube for passing the liquid whose flow rate is to be
measured;
induction windings having an excitation current passing therethrough
for setting up a magnetic field perpendicular to the direction of
liquid flow;
diametrically-opposed electrodes for sensing the voltage signal
which results from the flow of liquid through the magnetic field;
a circuit for feeding current to said induction windings and comprising
means for generating a periodic excitation current of period T,
each period T being the result of a combination in time of n periods
of a first excitation current (I.sub.1) of period T.sub.1 and m
periods of a second excitation current (I.sub.2) of period T.sub.2
where T.sub.2 >T.sub.1 with the period T.sub.2 being long enough
to ensure that portion of the corresponding sensed voltage signal
is free from the relaxation phenomenon; and
a processor circuit comprising means for generating a first intermediate
measurement signal (M.sub.1) of period T.sub.1 and having a short
response time, a second intermediate measurement signal (M.sub.2)
corresponding to the voltage signal of period T.sub.2 and being
free from measurement drift, and for combining and filtering said
first and second intermediate measruement signals in order to generate
a measurement signal S being both short in response time and free
from drift.
It can thus be seen that the first intermediate signal M.sub.1
has sufficient pulses per period T for effecting filtering to be
performed in a limited number of periods T, and that each pulse
of the voltage signal of period T.sub.2 is sufficiently long to
include a portion which is free from relaxation phenomena.
In a first embodiment of the first and second excitation currents
are currents in which the periods are constituted by alternating
positive-going and negative-going pulses, each of duration T.sub.1
/2 or T.sub.2 /2 as the case may be, and said combination consists
in alternating n periods of the first current with m periods of
the second current so as to obtain the relationship:
In a second embodiment, the first excitation current is a sinewave
current and each period of said second excitation current consists
in a positive-going half cycle and a negative-going half cycle of
duration t'=T.sub.2 /2 said combination consisting in superposing
said first and second currents to obtain the relationship:
Preferably the period T.sub.1 lies in the range 5 milliseconds
to 100 milliseconds, m is equal to 1 and the period T.sub.2 lies
in the range 50 milliseconds to 1 second.
In a preferred embodiment, said processor circuit comprises means
for generating a corrections signal based on said intermediate measurement
signals M.sub.1 and M.sub.2 said correction signal being substantially
free from noise, together with means for generating said measurement
signal S from said first intermediate measurement signal M.sub.1
and said correction signal, said measurement signal being substantially
free from drift.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments of the invention are described by way of example with
reference to the accompanying drawings, in which:
FIGS. 1a and 1b are described above and show a prior art method
of processing the signal sensed by the electrodes in an electromagnetic
flow meter;
FIG. 2 is a block diagram of the first embodiment of a circuit
in accordance with the invention for processing the voltage signal
sensed by the flow meter electrodes;
FIG. 2a shows a variant of the FIG. 2 circuit;
FIG. 3 is a waveform diagram respectively showing the magnetic
field provided by the induction windings in accordance with a first
implementation of the invention, the corresponding voltage signal
sensed at the electrodes, and the control signals used in the FIG.
2 circuit;
FIG. 4 is a block diagram showing a variant circuit for processing
the voltage signal in accordance with a first embodiment of the
invention;
FIG. 5 shows other possible waveforms for the signal controlling
the flow meter magnetic field:
FIG. 6 is a block diagram of a second embodiment of a flow meter
in accordance with the invention; and
FIG. 6a is a waveform diagram of the control signals used in the
FIG. 6 embodiment.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
A first embodiment of an electromagnetic flow meter in accordance
with the invention is described with reference to FIGS. 2 and 3.
As can be seen in FIG. 2 the flow meter comprises a cylindrical
measuring tube 10 through which there flows the liquid whose flow
rate is to be measured. The tube 10 has two induction windings 12
and 14 disposed diametrically opposite each other in order to produce
a magnetic field perpendicular to the direction of liquid flow,
i.e. perpendicular to the axis XX' of the tube 10. The tube also
includes two diametrically opposed electrodes 16 and 18 which are
in contact with the liquid flowing along the tube. The electrodes
16 and 18 serve to sense the electromotive force (emf) created by
the flow of liquid through the magnetic field produced by the induction
windings 12 and 14.
The flow meter also includes an electronic circuit constituted
by a circuit 20 for feeding the induction windings 12 and 14 a
circuit 22 for processing the voltage signal sensed by the electrodes
16 and 18 and a sequencer circuit 24 for controlling the circuits
20 and 22.
In this first embodiment of a flow meter, the winding feed circuit
20 is constituted by a voltage controlled constant current generator
26 capable of producing a constant positive or negative current,
and a generator control circuit 30 for reversing the direction of
current flow through the induction windings 12 and 14.
The processor circuit 22 comprises a preamplifier 32 whose inputs
are connected to the electrodes 16 and 18 and a sampling and calculation
circuit 34 connected to the output from the preamplifier 34.
The output from the sampling circuit 34 is connected to a selection
and switching circuit 36 for controlling filters described below.
The selector circuit 36 has two outputs 36a and 36b which are respectively
connected to the inputs of two filters respectively referenced 40
and 42 and having time constants t.sub.3 and t.sub.2 which are as
close to each other as possible. These filters are lowpass filters.
The outputs from these filters 40 and 42 are connected to the inputs
of a subtractor 44. The output 36a from the selector 36 is also
connected to the input of a thirdlow pass filter 46 having a time
constant t.sub.1 where t.sub.1 is much less than t.sub.2 and t.sub.3.
The outputs from the subtractor 44 and from the filter 46 are connected
to the inputs of a subtractor 48.
Before describing the structure of the sequencer circuit 24 reference
is made to the waveform diagrams shown in FIG. 3.
Waveforem 3(a) shows the variation as a function of time in the
current applied to the induction windings 12 and 14 and is thus
substantially representative of the variations as a function of
time in the magnetic fields created by the induction windings.
The excitation current has a period T which is constituted by two
alternating series of pulses I.sub.1 and I.sub.2 which are all of
equal amplitude. The first series of pulses I.sub.1 is formed, as
shown in FIG. 3 by five periods T.sub.1 of alternating rectangular
pulses. The second series I.sub.2 is constituted by a single period
T.sub.2 of alternating pulses which are likewise rectangular. Thus:
More generally, the period T comprises 2n successive half-cycle
cycle pulses of period T.sub.1 and 2m successive half-cycle pulses
of period T.sub.2 giving the relationship:
where T.sub.1 <T.sub.2.
Waveform 3(b) represents the voltage sensed by the electrodes.
As already explained, the sensed voltage pulses are affected by
a parasitic relaxation effect. This effect may affect the entire
duration of the pulses of period T.sub.1. In contrast, the period
T.sub.2 is chosen to be long enough to ensure that the end of each
pulse is free from the relaxation phenomenon.
The sequencer circuit 24 serves firstly to control the application
of a pulse current having the waveform shown in waveform 3(a) to
the induction windings, and secondly to generate control signals
for the processor circuit 22.
Waveforem 3(c) is representative of a signal S.sub.1 for controlling
the current applied to the induction windings. The signal S.sub.1
has a period T and is constituted by five pulses of duration T.sub.1
/2 separated by time intervals of equal duration T.sub.1 /2 and
by a single pulse of duration T.sub.2 /2 and a single interval of
equal duration T.sub.2 /2. The signal S.sub.1 is high when the current
is positive and zero when the current is negative. Waveforem 3(d)
shows a signal S.sub.2 which is used to control the instants at
which the voltage signal delivered by the electrodes 16 and 18 is
measured. The signal S.sub.2 has a period T. Each period comprises
a first series of ten half-cycle pulses (2n) separated by time intervals
T.sub.1 /2. Each of the pulses in this portion of the signal S.sub.2
is offset relative to the corresponding pulse in the signal S.sub.1
by a length of time b.sub.1 =aT.sub.1 (where a<1). Each period
of the signal S.sub.2 also comprises two pulses (2m) separated by
time intervals equal to T.sub.2 /2. These pulses in the signal S.sub.2
are offset relative to the corresponding pulses in the signal S.sub.1
by a length of time b.sub.2 =bT.sub.2 (where b is less than 1).
The values a and b are chosen to that a pulse of the signal S.sub.2
is present at the end of each pulse of the voltage signal. The number
b is determined more exactly so that the corresponding pulse in
the signal S.sub.2 is present for that portion of the voltage pulse
which is free from the relaxation phenomenon.
Waveform 3(e) is representative of a third control signal S.sub.3
of period T which is used to distinguish between that portion of
each period T where the pulses of period T.sub.1 are present from
that portion of the same period T where the pulses of period T.sub.2
are present. Each period of the signal S.sub.3 comprises a single
positive going portion of duration 5T.sub.1 (nT.sub.1) and a negative
going portion of duratin T.sub.2 (mT.sub.2).
The structure of the sequencer circuit 24 is described in greater
detail with further reference to FIG. 2.
The circuit 24 includes a generator 50 which has outputs 50a, 50b,
50c, 50d, and 50e permanently delivering signals of period T.sub.1
T.sub.2 nT.sub.1 mT.sub.2 and the shifted control signal S.sub.2.
The generator 50 includes an oscillator 52 which delivers a signal
of period T', where T' is much less than T.sub.1. The output from
the oscillator 52 is connected to the input of two frequency dividers
54 and 56 connected in cascade which deliver signals of period T.sub.1
and T.sub.2 and to the inputs of three counters 58 60 and 62 which
respectively deliver the signal S.sub.2 and the signals of periods
nT.sub.1 and mT.sub.2. The outputs 50a to 50e from the generator
50 are connected to respective inputs of a logic circuit 64 which
has three outputs 64a, 64b, and 64c on which it delivers the control
signals S.sub.1 S.sub.2 and S.sub.3 respectively. Such a logic
circuit is conventional and there is no need to describe it in greater
detail.
The signal S.sub.1 is applied to the input of the control circuit
30 of the current generator, and to the control input 34a of the
circuit 34. The signal S.sub.2 is applied to the sampling cntrol
input 34b of the circuit 34. Finally, the signal S.sub.3 is applied
to the control input 36c of the selector 36.
The circuit shown in FIG. 2 operates as follows: the constant current
delivered by the generator 26 is applied to the induction winding
with alternating polarity and with the polarity-reversal instants
being under the control of the signal S.sub.1. This current has
the waveform shown in waveform 3(a). In a particular implementation,
the period T.sub.1 is 32 milliseconds (ms) and the period T.sub.2
is 131 ms. The overall period T is thus equal to 291 ms.
The voltage sensed by the electrodes 16 and 18 has the waveform
shown in waveform 3(b). Each period T comprises five positive and
negative half-cycle pulse pairs of period T.sub.1 and a single positive
and negative half-cycle pulse pair of period T.sub.2 and each of
these pulses may be degraded by the relaxation effect and by noise.
The sensed voltage is amplified by the preamplifier 32. The sampler
takes a sample of the signal at the end of each positive or negative
pulse at instants which are defined by the signal S.sub.2 and the
signal S.sub.1 is used to cause differences to be taken between
sampled values. Since the switch 36 is controlled by the signal
S.sub.3 the samples taken from the portion of the signal having
a period T.sub.2 and constituting the second intermediate measurement
signal M.sub.2 are applied to the input of the filter 40. As explained
above, these samples suffer from noise but they are free from the
relaxation phenomenon. One such sample is taken per period T, i.e.
every 291 ms. Similarly, the samples taken from the signal portion
of period T.sub.1 and forming the intermediate measurement signal
M.sub.1 are applied to the inputs of the filters 42 and 46 having
time constants t.sub.2 and t.sub.1. These samples are affected both
by noise and by the relaxation effect. There are five samples per
period T.
The time constants t.sub.3 and t.sub.2 are identical and equal
to 30 seconds. The common time constant value is sufficiently long
to provide good noise filtering even though there is only one sample
per period T. The output from the subtractor 44 delivers a correction
signal S.sub.z which provides a signal which is representative of
the error due to the relaxation effect and which is free from noise.
The signal delivered by the filter 46 which has a considerably
smaller time constant t.sub.1 for example a time constant of three
seconds, is well filtered but suffers from the relaxation effect.
The output from the subtractor 48 thus delivers a measurement signal
S which is free both from the relaxation effect and from noise,
and which is usable in a regulation system since overall it behaves
as though filtering is taking place with the shorter time constant
t.sub.1 which is equal to 3 seconds.
FIG. 2a shows a variant embodiment of the processor circuit 22
shown in FIG. 2.
The circuit referenced 70 comprises, like the circuit 22 the preamplifier
32 the calculation and sampling circuit 34 and the selector circuit
36 having two outputs 36a and 36b. The first intermediate measurement
signal M.sub.1 of period T.sub.1 appears on the output 36a which
is connected to one of the input terminals of the subtractor circuit
74 and to the input of a lowpass filter 72 which is identical to
the filter 46 and which has the same time constant t.sub.1. The
second intermediate measurement signal M.sub.2 of period T.sub.2
appears on the output 36b of the selector 36 which is connected
to the second input terminal of the subtractor 74. The output from
the subtractor 74 delivers a correction signal which is applied
to the input of a lowpass filter 76 having a time constant t.sub.2
which is identical to that of the filter 40. The outputs from the
filters 72 and 76 are connected to the inputs of a subtractor 78
which delivers the measurement signal S with the above-described
characteristics.
FIG. 4 shows a variant of the first embodiment of the electronic
circuit associated with the measurement sensor. In this variant,
the signal is processed digitally instead of being processed by
analog means, as in the FIG. 2 circuit.
The flow meter shown in FIG. 4 comprises a sensor per se constituted
by a measuring tube 200 with induction windings 202 and 204 and
with electrodes 206 and 208 for sensing the voltage signal.
The electronic circuit is essentially constituted by a current
source 210 for feeding the induction windings 202 and 204 an impedance-matching
preamplifier 212 which receives the voltage signal present on the
electrodes 206 and 208 an analog-to-digital converter 214 which
converts the voltage signal into digital form and which also provides
the sampling function, and a microprocessor 216 which performs both
the sequencing and the signal processing functions. The circuit
may also include a digital-to-analog converter 218 for delivering
an output signal, for example an electric current, which is proportional
to the fluid flow rate as calculated by the microprocessor 216.
The microprocessor 216 is advantageously a Motorola type 6803 microcontroller
which generates the signals S.sub.1 and S.sub.2 shown in FIG. 2
digitally, and which applies them to the current driver 210 and
to the analog-to-digital converter 214 respectively. The program
for generating the signals S.sub.1 and S.sub.2 from a clock signal
is conventional and does not need describing in greater detail.
The microprocessor also performs digital filtering on the portions
of the voltage corresponding respectively to the periods T.sub.1
and T.sub.2 and it combines the filtered portions in the same manner
as does the analog circuit shown in FIG. 3. A particular advantage
of digital processing is that is makes it possible to define filter
parameters to a high degree of accuracy. It is thus possible to
obtain time constants t.sub.2 and t.sub.3 which are very close to
each other.
Using a microprocessor for digitally processing, and in particular
for filtering, digitized signals is conventional and there is no
need for a more detailed description here. These techniques are
described in particular in "Journal of the Audio Enginnering
Society", October 1979 volume 27 number 10 pages 793 to
802 and in "Theory and Application of Digital Signal Processing"
by Lawrence R. Rabiner, published in 1975 by Prentice-Hall.
In the above description, the magnetic field control signal is
constituted by combining two signals I.sub.1 and I.sub.2 each comprising
alternating pulses and having respective periods T.sub.1 and T.sub.2.
The signals are combined by causing groups of n periods of the signal
I.sub.1 of period T.sub.1 to alternate with groups of m periods
of the signal I.sub.2 of period T.sub.2.
Waveform diagrams 5(a), 5(b), and 5(c) show other possible combinations
of the signals I.sub.1 and I.sub.2 of respective periods T.sub.1
and T.sub.2 and suitable for generating the magnetic field.
Waveforms 5(a) and 5(b) show two other control signal waveforms
obtained by the same combination of signals having periods T.sub.1
and T.sub.2. In waveform 5(a) each period T.sub.1 of the signal
is constituted by a pulse of duration T.sub.1 /2 followed by a zero
level for the same duration T.sub.1 /2. Each period T.sub.2 of the
signal is likewise constituted by a half-cycle pulse followed by
a half-cycle zero level. In waveform 5(b) each cycle of either period
includes both a positive-going pulse and a negativegoing pulse,
but the pulses are separated from one another by short durations
of zero level.
In waveform 5(c) the signal I.sub.1 of period T.sub.1 is a sinewave
signal of constant amplitude, while the signal I.sub.2 of period
T.sub.2 is an alternating pulse signal with each positive-going
and negative-going pulse having a duration of T.sub.2 /2. In addition,
T.sub.2 is an integer multiple of T.sub.1. These two signals are
combined by being superposed in order to obtain the control signal.
In this case it will be understood that the overall period T of
the control signal is equal to the period T.sub.2.
In all cases the signal of period T.sub.2 includes a pulse of duration
t', where t' is long enough to ensure that the voltage sensed at
the end of the pulse is free from relaxation effects.
FIG. 6 shows a second embodiment of a flow meter in accordance
with the invention and which makes use of a magnetic field generated
by a control signal of the type shown in waveform 5(c). It comprises
a measurement tube 10 having two induction windings 12 and 14 and
two electrodes 16 and 18. The flow meter also includes a driver
circuit 102 for generating the current which is applied to the induction
windings 12 and 14 and a circuit 104 for processing the voltage
signal which appears at the electrodes 16 and 18 of the flow meter.
The current generator circuit 102 comprises an oscillator 106 which
delivers a sinewave signal of period T.sub.1. The output from the
oscillator 106 is connected firstly to a first input 108a of a current
generator 108 and secondly to the input of a shaper circuit 110.
The output 110a from the shaper circuit 110 delivers a signal comprising
alternating rectangular pulses having the same period T.sub.1 as
the sinewave signal delvere by the oscillator 106. The output 110a
is connected firstly to the input 112a of a counter 112 and secondly
to a first input 114a of a logic circuit 114. The output 112b from
the counter 112 delivers a signal comprising alternating rectangular
pulses of period T.sub.2 where T.sub.2 is a multiple of T.sub.1.
The output 112b is connected firstly to to a second input 108b of
the current generator 108 and secondly to a second input 114b of
the logic circuit 114. The current generator 108 responds to the
signals applied to its first and second inputs 108a and 108b by
delivering a current having waveform 5(c) , and this current is
applied to the induction windings 12 and 14. On its output 114c,
the logic circuit 114 delivers a logic signal Cd having the waveform
shown in Figure 6a, i.e. comprising a short positive-going pulse
and a short negative-going pulse located at the ends of the positive-going
and negative-going portions respectively of the signal of period
T.sub.2 delivered by the counter 112. Each of these "short"
pulses lasts for a duration equal to two periods of the sinewave
signal.
The processor circuit 104 comprises a preamplifier 116 which receives
the voltage sensed by the electrodes 16 and 18 of the flow meter.
The output from the preamplifier 116 is connected firstly to the
input of a synchronous demodulator 120 and secondly to the input
of a sampler and subtractor 122. The control input 120a of the demodulator
120 receives the signal delivered by the shaper 110 and the control
input 122a of the sampler 122 receives the signal delivered by the
logic circuit 114. Thus the output from the demodulator delivers
a first intermediate measurement signal derived from the voltage
sensed by the electrodes and corresponding to the sinewave signal
of period T.sub.1. The output from the circuit 122 delivers a second
intermediate measurement signal which corresponds to the difference
between a positive sample of the complete signal sensed during a
complete period of the sinewave signal and a negative sample of
the complete signal sensed during a complete period of the sinewave
signal. Since both sampling instants are located at the ends of
pulses in the signal of period T.sub.2 the values of these samples
are not degraded by the relaxation effect. The circuit 104 finally
includes low pass filters 40', 42', and 46', analogous to the filters
40 42 and 46 of FIG. 2 together with subtractors 44' and 48'
identical to the circuits 44 and 48 of FIG. 2. The subtractor 48'
thus delivers a flow rate measuring signal S at its output, which
signal is unaffected by the phenomena of drift or noise.
This second embodiment has the advantage, for given periods T.sub.1
and T.sub.2 of reducing the value of the overall period T since
in this case T=T.sub.2 rather than T=nT.sub.1 +mT.sub.2. However,
the circuits are rather more complex since they must process both
logical signals and sinewave signals.
In the above description, the numbers of periods m and n in the
signals are integers, however there is no need for these numbers
to be integers, so long as both m and n are not less than unity
and n is greater than m.
Similarly, a good quality measurement signal (S) is obtained for
a period T.sub.1 lying in the range 100 milliseconds to 5 milliseconds,
and for a period T.sub.2 lying in the range 50 milliseconds and
1 second, with the 1 second value being acceptable in particular
when using a current waveform of the type shown in FIG. 5(c) .
It can thus be seen that in all embodiments of the invention, the
measurement signal S is obtained by generating: two intermediate
measurement signals M.sub.1 and M.sub.2 which correspond respectively
to the periods T.sub.1 and T.sub.2 of the signal sensed by the electrodes;
a correction signal based on the signals M.sub.1 and M.sub.2 and
representative solely of the relaxation phenomenon; and finally
a measurement signal S derived from the intermediate measurement
signal M.sub.1 and the correction signal. |