 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The present invention relates to current sources for loads of a
considerable range of impedance values which include an inductive
component and, more particularly, to a current source for such loads
through which the direction of current flow must alternate.
Current sources which provide a relatively constant current, at least for
successive periods of time, are needed in many applications including ones
which require establishing relatively constant magnetic fields for such
periods of time. Such currents can be difficult to keep constant for such
successive periods of time if the current flow direction must be reversed
between each of these periods and they are of relatively short duration.
This is especially so where the periods of desired current constancy are
to occur periodically, with a flow direction reversal between each, and at
a frequency which is significant with respect to various time constants
occurring in the system in which this current is provided.
One such system in which these conditions arise is electromagnetic
flowmeters. In these systems, a magnetic field is established across a
metering tube through which is flowing a liquid, or liquid-like, substance
which exhibits at least some electrical conductivity. Conductive media
flowing through a magnetic field lead, in accord with electromagnetic
theory, to the establishment of an electromotive force or voltage
perpendicular to both the flow direction and the magnetic field direction,
a voltage which is proportional to the average velocity of flowing fluid.
The provision of electrodes at the locations in which the voltage is
primarily developed permits obtaining a signal which is linearly
representative of the velocity of the liquid from which its liquid flow
can be determined.
However, if the magnetic field is constant in magnitude and direction, the
resulting constant polarity voltage portion due to the fluid flow induced
signal cannot be separated from the portion due to the electro- chemical
potential of the flowing fluid and the sensing electrodes together.
Further, the resulting direct signal current established in the flowing
fluid, transverse to the direction of its flow, can lead to polarizing the
two sensing electrodes over time thereby adversely affecting the output
voltage signal representation of flow. To avoid this result, the magnetic
field is usually applied to the flow tube alternately in opposite
directions to balance out such transverse current flows and so avoid a net
polarization of the sensing electrodes.
The frequency of reversals, or alternations of direction, of the magnetic
field through the flow tube has a bearing on performance of the
electromagnetic flow measurement system. On the one hand, a higher
frequency of reversal will further separate that frequency from the noise
in the signal taken from the sensing electrodes that is of a type which
can be described as 1/f noise. On the other hand, the signal transmission
leads carrying the signal from the sensing electrodes acts as a
transmission line, and can be relatively long if the data capture site is
a substantial distance from the flow measurement site. In these
circumstances, the transmission line distributed capacitance and the
resistance of the fluid will have the electrical characteristics of a
low-pass filter. Thus, at some point, increasing the frequency of
reversals forming the basis of periodic variation in the electrical
signals from the sensing electrodes, will lead to reduced amplitudes in
such signals obtained from these electrodes at the data capture location
because of the filtering action of the connecting signal line.
In electromagnetic flowmeter systems, there is a desire to have a constant
current flowing in magnetic field coils during the times of obtaining
values representing flow from the sensed signal in any period of the
alternation of the applied magnetic field. If the magnetic field is
constant during such obtaining of sense signal values, there will not be
much inductive pickup occurring in system portions therearound including
in the apparatus used in obtaining the flow representation signal, or
sense signal, from the sensing electrodes. Thus, less noise or offset will
be present in this sense signal.
However, difficulty arises as the frequency of reversals of the applied
magnetic field increases. The magnetic field strength B in the fluid
results from various currents flowing in the electromagnetic flowmeter
metering system including (i) the current applied to the coils from a
current source to provide the desired magnetic field, and (ii) the
resulting eddy currents which are induced to flow in the conductive
metering tube and the magnetic materials in the magnetic return circuit.
The magnetic field strength B in the flowing fluid thus approaches a
steady value exponentially after a current reversal, due to the shielding
effects of the eddy currents. Thus, changes in the B field in the flowing
fluid material lag the changes in the applied current giving rise to them,
and do so for a greater fraction of each of the periods of field direction
alternation the more frequently the reversals occur, i.e. the shorter the
period.
Hence, the amount of time that the B field is constant in each period
becomes smaller and smaller as the frequency of the reversals increases
thereby leading to less and less time to acquire a sensed value in the
sense signal from the sense electrodes which is substantially free of
inductive noise as a result of the inductive pickup noise having had time
to subside. At some point of increasing frequency, this leads to an
increase in the noise in the signal provided by the sense electrodes, and
can lead to a substantial decrease in the magnitude of the sense signal
because the B field remains substantially reduced by the eddy current in
each period.
This latter condition is worsened if the applied current provided in the
electromagnetic flowmeter system also has a relatively long rise time
after each reversal of its direction of current flow as required to
provide the corresponding reversal of the applied B field. After a current
direction reversal thereof, delay in reaching a desired constant current
value leads to shorter times also for the resulting B field generated by
it and the induced eddy currents to come to a constant value in any period
after a field direction alternation. Thus, there is a desire to shorten
the rise time of the current used to provide the B field in an
electromagnetic flowmeter after a flow direction reversal.
SUMMARY OF THE INVENTION
The present invention provides a current source electrically energized by a
variable excitation or supply for establishing a relatively constant
current between current direction reversals in selected loads of widely
differing impedances. The current source comprises a potential regulator
for adjustably regulating supplied potential to provide a current through
a current controller, a load and a current sensing means which supplies a
signal to the current controller. There is a further power regulator which
receives a signal from the current controller and provides an output to
the voltage regulator to adjust its output. The current to be supplied to
the load can be caused to alternately switch current direction through
that load by a switching commutator arrangement. A current rise time
regulator is provided to increase the rise time of currents provided after
a commutation of the load.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1A and 1B show a schematic circuit diagram of a circuit according to
the present invention, and
FIGS. 2A through 2E show waveforms characteristic of the circuit of FIG. 1.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIGS. 1A and 1B can be joined along line L--L' to form a schematic circuit
diagram of an embodiment of the present invention, used here in an
electromagnetic flowmeter system. The circuit is supplied a constant
polarity voltage between a pair of terminals, a positive voltage of
typically 40.0 V being applied at a positive voltage supply terminal, 10,
with respect to a circuit common or ground reference voltage which is
electrically connected to a ground reference terminal, 11.
The load element for the current source circuit of FIG. 1 comprises an
electromagnetic flowmeter coil, 12, and a cable represented by resistance
12A, which are connected in a commutator bridge arrangement formed by four
switching metal-oxide-semiconductor field-effect transistors (MOSFET's),
13, 14, 15 and 16. This commutation arrangement is operated by a
commutator current direction switching logic and switching transistor
driver arrangement, 17. Arrangement 17 first switches transistors 13 and
16 into the "on" condition with transistors 14 and 15 in the "off"
condition followed by reversing these conditions of each of the pairs of
transistors, and thereafter alternating between these conditions for these
transistor pairs. Thus, in the first condition during the first half of a
period for a cycle in this periodic switching, current can flow into the
source of transistor 13, and then out its drain through coil 12 into the
drain of transistor 16 followed by exiting through the source of
transistor 16. In the second half of this period of the cycle, current
flow through transistor 14 from source to drain, and then in the opposite
direction of flow through coil 12 from the first half period, and then
through transistor 15 from drain to source.
Thus, coil 12 periodically has a current applied therethrough in opposite
directions to provide a periodically varying sense signal and to avoid the
polarization of the sensing electrodes in that electromagnetic flowmeter
system metering tube, 12', on which coil 12 is installed in the
electromagnetic flowmeter system. Switching logic and driver means 17 can
be formed of any suitable circuitry which provides the properly timed
gating voltage pulses to transistors 13, 14, 15 and 16 to result in the
described switching sequence. The frequency of current flow direction
reversals set by means 17 is selectable so that it can be chosen to be
best for the flow measuring conditions encountered, particularly the
parameters of the fluid being measured. These parameters of such fluid can
differ greatly from fluid to fluid, the most significant typically being
the electrical conductivity of a fluid and its 1/f noise characteristics.
A series connection of a capacitor, 18, and a resistor, 19, are together
placed in parallel with the load comprising coil 12 and cable 12A to
provide a discharge path for the inductance of coil 12 during the moments
of switching when all of the commutator transistors could be momentarily
in a partially "off" condition as the magnetic field about coil 12
collapses to provide a temporary discharge path of a selected time
constant for this energy. Typical component values are 0.22 .mu.F for
capacitor 18, and 8.7K.OMEGA. for resistor 19, selected to keep inductive
spikes from coil 12 from exceeding the ratings of the commutator switches.
The load will have an effective series resistance associated with it due to
the internal resistance of the coil 12 and that of the interconnection
transmission line 12 A between coil 12 and the drains of transistors 15
and 16 serving as the current source load terminals. Since an
electromagnetic flowmeter will be installed in many different situations
so that the length of this interconnecting transmission line can vary
greatly, the amount of series resistance due to such interconnections will
be quite different from one situation to the next. A wide range of
temperatures in some measuring situations, and from one measuring
situation to the next, will further extend the resistance variation range.
Coil 12 will itself have a wide range of inductance and internal resistance
from one measuring situation to the next. Accommodating the sense
circuitry connected to the metering tube sense electrodes typically
requires that the signal at these sense electrodes be substantially the
same magnitude from one measuring situation to the next for the same flow
velocity of fluid through the metering tube using the same applied
current. Since the metering flow tube can have a diameter varying between
a fraction of an inch to a few feet from one measuring situation to the
next, there must be a wide variation in the inductance and internal
resistance of coil 12 for use with these various metering flow tubes.
Thus, the potential provided across the load terminals of the commutator
arrangement to which coil 12 is connected must vary significantly from
situation to situation to provide essentially the same absolute value
current flow in coil 12 in each such situation. As indicated above,
similar absolute value current flow is required in each measuring
situation to assure that magnetic fields developed thereby are of the
magnitudes desired to provide a signal indicative of the flow rate through
the flow tube which is generated on a consistent basis in each situation
for the sense signal acquisition circuitry, 12", connected to a pair of
sensing electrodes, 12'", in tube 12'. Further, as indicated above, the
currents must be maintained constant in each flow direction through coil
12 for a sufficient portion of the time during flow in each direction in
each alternation period to permit an accurate reading to be obtained in
this the same signal acquisition circuitry connected to the electrodes in
the flow tube.
The need to achieve rapid current magnitude increases after each applied
current flow direction reversal at the highest reversal frequencies if
relatively large effective resistance is in series with coil 12, and coil
12 has a relatively large inductance, is a situation demanding a
relatively large voltage being provided at the commutator arrangement lead
terminals. Yet, efficient use of electrical power is required of the
circuit supplying voltage at these terminals in other measuring situations
in which lower applied current flow direction frequencies are used or when
smaller values of effective resistance is in series with coil 12 or coil
12 has smaller inductances, or in situations in which a combination of
these reduced values are present. Thus, the remaining circuitry shown in
FIG. 1 is directed toward providing this result.
Because of the need for a variable potential at the commutator load
terminals to which coil 12 is connected in different situations, and yet
an essentially constant valued current flow between them in alternate
directions on alternate half cycles in every situation, the circuit of
FIG. 1 must provide for such a varying magnitude voltage but with a well
controlled current flow for each such voltage. This could be accomplished
by just changing a voltage supply means output voltage in a single
feedback loop. However, to maintain satisfactory power efficiency there is
a need for a switching regulator as the supply of voltage. Such switching
regulators are well known to have delays therein in adjusting output
voltage values at the output thereof in response to changing adjustment
signals at an input thereof. This means a relatively slow loop response in
adjusting currents when current value errors arise because of changing
conditions in which the desired fixed current value flow must be met in
alternate half cycles.
The circuit of FIG. 1 has, instead, two feedback loops for controlling the
voltage and the current substantially separately. A switching regulator is
used here also for maintaining power efficiency, this regulator being
based in part on a commercially available monolithic integrated circuit
with the portions therein substantially used in the circuit of FIG. 1
shown in block diagram form within a dashed line box, 20. The numbers
thereabout preceded by the letter p are the pin numbers for the commercial
integrated circuit available from Unitrode Integrated Circuits under the
designation UC 494A.
In block 20, an error amplifier, 21, receives signals indicating error in
the regulator output voltage as measured by external circuit portions
connected to its inverters and non-inverting inputs. External feedback
elements including a pair of resistors, 22 and 23, and a capacitor, 24,
are used to set the gain of the circuitry involving error amplifier 21 as
well as to set its frequency response as will be described later.
The error signal, after undergoing amplification provided by amplifier 21
and its associated circuitry, is provided to the non-inverting input of a
comparator, 25. Further supplied to the inverting input of comparator 25
is the output of a fixed frequency oscillator, 26. Comparator 25 compares
the filtered and amplified error voltage, supplied thereto from the output
of error amplifier 21, with the fixed frequency periodic sawtooth output
voltage provided thereto by oscillator 26 to generate a variable-width
pulse output signal at the output of comparator 25.
That is, a pulse width modulated signal is provided at the output of
comparator 25 to a digital circuit arrangement, 27, providing certain
logic functions which result at the output of logic circuit 27 in a
switching signal at the base of an npn bipolar transistor, 28, having its
emitter connected to ground reference terminal 11. Transistor 28 is
switched from the "on" condition to the "off" condition, and back,
alternately, in accordance with the signal provided at the output of
comparator 25 in such a manner that the relative time that transistor 28
is in the "on" condition versus the "off" condition is adjustable by the
pulse width of the comparator 25 pulse width modulated output signal.
The fixed frequency of oscillator 26 is selected by the choice of values
for a resistor, 29, and a capacitor, 30, connected thereto in the
integrated circuit within box 20 and to ground reference terminal 11.
Values chosen to set the oscillator frequency at about 40 KHz are
13K.OMEGA. for resistor 29 and 2200 pF for capacitor 30.
The fixed frequency oscillating signal from the output of oscillator 26 is
also supplied to a further comparator, 31, at its inverting input. A
voltage reference, 32, provides a fixed output voltage of typically 5.0 V,
a portion of which is provided to the non-inverting input of comparator 31
through a voltage divider arrangement connected between the output of
voltage reference 32 and ground reference terminal 11. This voltage
divider is comprised of a pair of resistors, 33 and 34, and a capacitor,
35, which is parallel with resistor 33. Typical values for these
components are 390 K.OMEGA. for resistor 33 and 100 K.OMEGA. for resistor
34, with capacitor 35 having a value of 10 .mu.F which acts to delay the
decrease in voltage supplied at the input to comparator 31 to delay
starting the operation of the switching regulator at initial application
of circuit power.
The voltage developed at the juncture of resistors 33 and 34 provided to
the non-inverting input of comparator 31 determines the minimum amount of
"off" time for transistor 28 limiting the amount of time it is in the "on"
condition to prevent damage to the power supply components due to the need
to dissipate heat from current flowing therethrough for too long a period
of time. The output of comparator 31 is also supplied to steering logic 27
to accomplish this purpose in controlling transistor 28.
Transistor 28 controls a further transistor, 36, external to the integrated
circuit in box 20 which is a pnp bipolar transistor. The emitter of
transistor 36 is connected to positive voltage supply terminal 10, and the
collector of transistor 36 is connected to an inductor, 37, which in turn
is connected to a capacitor, 38, that is connected between inductor 37 and
ground reference terminal 11. Inductor 37 and capacitor 38 are also
external to the integrated circuit in box 20. The juncture of inductor 37
and capacitor 38 forms the output of the switching regulator of FIG. 1,
and together provide a low-pass filter to average the voltage pulses
provided by switching "on" and "off" of transistor 36 under the control of
transistor 28 and the rest of the integrated circuit in box 20. Inductor
37 typically has a value of 220 .mu.H, and capacitor 38 has a typical
value of 10 .mu.F.
The collector of transistor 36 is also connected to a diode, 39. Diode 39
has its cathode connected to the juncture of the collector of transistor
36 and inductor 37, and has its anode connected to ground reference
terminal 11. Diode 39 provides a path for current through inductor 37 when
transistor 36 is in the "off" condition in switching back and forth
between this condition and the "on" condition as directed by transistor
28.
Transistor 28 controls such switching of transistor 36 through a pair of
resistors, 40 and 41. If transistor 28 is in the "off" condition, no
significant current flows in either resistors 40 or 41 so that the base of
transistor 36 is at approximately the voltage appearing on positive
voltage supply terminal 10. If, on the other hand, transistor 28 is in the
"on" condition, current will be drawn by its collector through resistors
40 and 41 sufficiently lowering the voltage of the gate of transistor 36
to switch it into the "on" condition.
Thus, constant polarity or direct voltage is provided at the juncture of
inductor 37 and capacitor 38 as a source of electrical energy for
operating coil 12 as will be described below. The value of this voltage is
controlled by signals occurring at the inputs to error amplifier 21, a
capability which will be made use of as will be described below.
The application of this regulated voltage from the juncture of inductor 37
and capacitor 38, and the current supplied therefrom as a result, is
controlled in its application to the commutating arrangement described
above, including coil 12, by a further p-channel MOSFET, 42, having its
source connected to this regulated voltage supplied at the output of the
switching regulator of the juncture of inductor 37 and capacitor 38.
Transistor 42 is controlled in turn by a first feedback loop in its
supplying of current to the commutation arrangement at the regulated
voltage value provided at this switching regulator output. This feedback
loop further includes a diode, 43, having its anode connected to the drain
of transistor 42 and its cathode connected to the commutation arrangement,
specifically with its cathode connected to the sources of transistors 13
and 14 together. The loop continues on the other side of the commutating
arrangement with a current sensing resistor, 44, connected between the
sources of transistors 14 and 16 together and ground reference terminal
11. A typical value for resistor 44 will be 2.0 .OMEGA..
This feedback loop is completed by a feedback amplifier arrangement driving
the gate of transistor 42. This feedback amplifier arrangement includes an
operational amplifier, 45, which is of a high voltage type so that its
output can rise to the maximum positive voltage level which appears at the
switching regulator output at the juncture of inductor 37 and capacitor
38, this voltage being around 38.0 V. The non-inverting input of amplifier
45 is electrically connected through a resistor, 46, typically of 10 k
.OMEGA. to the junction of resistor 44 and the sources of transistors 15
and 16 together. Resistor 46 is used in setting the frequency response of
the amplifier 45 circuit in connection with a minor feedback loop about
amplifier 45 as will be described below. Thus, the current through coil 12
provides a voltage across resistor 44, which voltage is supplied to
amplifier 45 as a measure of the current through resistor 44 and so
through coil 12.
The inverting input of amplifier 45 is connected through another resistor,
47, also typically of 10 k.OMEGA., to a reference voltage. Resistor 47 is
used in balancing offsets in amplifier 45. The reference voltage is
approximately 1.0 V and is supplied at the juncture of a pair of voltage
divider resistors, 48 and 49. Resistor 48 has as typical resistance value
of 5.11 k.OMEGA., and resistor 49 has a typical resistance value of 1.28
k.OMEGA.. Resistor 48 has its opposite end connected to the output of
voltage reference 32 supplying 5.0 V as indicated above. The opposite end
of resistor 49 is connected to ground reference terminal 11.
Because a major or primary feedback loop has been provided around amplifier
45, as has been indicated above, the effect is to keep the potential
difference between the inverting and non-inverting inputs of amplifier 45
approximately zero in value. Thus, the feedback loop will act to provide a
current through resistor 44 which provides approximately a 1.0 V voltage
drop thereacross. The resulting current through coil 12, 1.0 V divided by
2.0 .OMEGA., is thus set to be 0.5 A in the present example shown in FIG.
1.
The output of amplifier 45 is connected to the gate of transistor 42
through a parallel arrangement involving a resistor, 50, and a capacitor,
51, which together are in series with a further resistor, 52. The gate of
transistor 42 is also connected to the regulated voltage supplied at the
juncture of inductor 37 and capacitor 38 by a further resistor, 53. In
parallel with resistor 53 are a pair of back-to-back Zener diodes, 53' and
53", which prevent the voltage on the gate of transistor 42 from differing
from the voltage on its source by more than 18.0 V in either polarity to
prevent exceeding the gate-to-source breakdown limits of transistor 42.
Typical component values are 22 k.OMEGA. for resistor 50, 1 k.OMEGA. for
resistor 52, and 100 k.OMEGA. for resistor 53. Capacitor 51 is typically
0.1 .mu.F. These component values are chosen for a particular type of
amplifier 45 and transistor 42, and would change with other choices for
such amplifiers and transistors because of their differing
characteristics.
Resistor 33, with respect to the series connection of resistors 50 and 52,
forms a voltage divider arrangement across the gate of transistor 42
between the regulated voltage supply and the output of amplifier 45. This
arrangement allows choosing the amplifier 45 output to be at a nominal
voltage point compatible with its range of output voltage and with
operating transistor 42 when it is to be switched into the "on" condition.
Because amplifier 45 is a high voltage amplifier, the output voltage value
can rise to the output of the voltage regulator to restrict conductance
through transistor 42, but can also drop to switch transistor 42 strongly
into the "on" condition.
Because of the need to keep this feedback loop operating rapidly to restore
current flows through coil 12 after the current has been switched off
therein in preparation for reversing the flow through this coil, capacitor
51 is used across resistor 50 to increase the rate that the output signal
of amplifier 45 is applied to the gate of transistor 42. Resistor 52
limits this rate somewhat, but allows a significantly greater signal to
initially be applied to the gate of transistor 42 than would be possible
if resistor 50 was not bypassed at the beginning of bringing current
through coil 12 up to the desired constant value in the next half cycle
after each switching of current flow direction in coil 12.
However, this feedback loop cannot be allowed to operate so rapidly as to
have oscillations occur after each switching of current flow direction
through coil 12. This is because such oscillations or "ringing" after a
sharp switching pulse would be introduced into a second feedback loop (to
be described) which can have adverse consequences. As a result, a local,
or minor, feedback loop arrangement is provided around amplifier 45 to
shape the frequency response of the local feedback loop around amplifier
45 so as to shape the frequency response of the major feedback loop in
which amplifier 45 is present as described above.
This local feedback loop extends through transistor 42 and then through a
resistor, 54, in series with a capacitor, 55, to the non-inverting input
of amplifier 45. A further capacitor, 56, is parallel with this series
combination of resistor 54 and capacitor 55. Thus, this entire combination
of passive components is connected between the drain of transistor 42 and
the non-inverting input of amplifier 45. Typical component values are 1.8
M.OMEGA. for resistor 54, 0.1 .mu.F for capacitor 55, and 820 pF for
capacitor 56. This choice of values allows the major feedback loop to be
approximately critically damped in permitting only a small amount of
damped oscillation or overshoot of relatively short duration after a
switching of the current direction through coil 12.
The performance of this major feedback loop is indicated in the waveforms
shown in FIGS. 2A, 2B and 2C. FIG. 2A shows the switching pulses applied
to the gate marked A in FIG. 1B of the commutating transistor 14 as the
timing base used by logic and driver arrangement 17 to operate the
commutator arrangement, in this instance the voltage applied to the gate
of transistor 13. There is a voltage level transition in this waveform
every half cycle. Since this controls the switching of the commutator
arrangement, and so the current reversals through coil 12, this commutator
timing base is a convenient choice for the timing base of the system of
FIG. 1. The timing base on the left in FIG. 2A is identical to the one on
the right but the two of them are provided because alternative load
examples are presented side-by-side in the graphs of FIG. 2.
In FIG. 2A(i), the | | |
