Abstrict A novel, low-friction, low-inertia flexible diaphragm containing
a magnet formed as an integral part of the diaphragm is mounted
in a chamber (preferably cylindrical). The chamber comprises part
of the flow meter in a flow control system. One or more sensing
devices are mounted on the walls of the chamber to sense the instantaneous
position of the diaphragm. A novel electronic control circuit processes
the data from the one or more sensing devices to provide measures
of the flow rate during the displacement of the diaphragm along
the cylinder. Signals are generated by the electronic control circuitry
for switching a pair of three-way valves (one valve comprising the
input valve and the other valve comprising the output valve) such
that during one-half of a cycle the intput valve transfers fluid
into the cylinder on one side of the diaphragm and, during the other
half of the cycle, transfers fluid into the cylinder on the other
side of the diaphragm. The output valve is switched synchronously
with the input valve to transmit fluid from the other side or one
side of the diaphragm to the output valve. The electronic control
system includes means for amplifying the output signals from the
one or more sensing devices ("sensors") to provide one
or more signals representing the position of the diaphragm as a
function of time, means for converting the output signals from these
amplifiers to digital form, and computation means for operating
on the digitized output signals from the sensing devices to provide
control signals for controlling a second input valve thereby to
control the flow rate of the fluid being metered to within a desired
range.
Claims We claim:
1. Structure comprising:
a chamber containing a first and a second opening through which
fluid can pass;
a flexible diaphragm located in said chamber so as to divide said
chamber into two portions, a first chamber accessed through said
first opening and a second portion being accessed through said second
opening;
a magnet attached to said flexible diaphragm;
means for directing fluid whose flow is being measured through
said first opening into said first chamber while withdrawing the
fluid whose flow is being measured through said second opening from
said second chamber and, into response to a control signal, reversing
the chambers in which the fluid is inserted and from which the fluid
is withdrawn, the movement of fluid into one chamber and out of
the other chamber causing said flexible diaphragm to move into the
chamber from which the fluid is being withdrawn;
means for continuously producing an output signal representative
of the position of said magnet;
means for converting said output signal to a sequence of digital
signals;
means for processing said sequence of digital signals to produce
a second signal representative of the flow rate of said fluid into
and out of said chamber;
means for comparing said flow rate to a reference flow rate to
produce a control signal representative of the difference between
said measured flow rate and said reference flow rate; and
means, responsive to said control signal, for changing the flow
rate of said fluid.
2. Structure comprising:
a chamber containing a first and a second opening through which
fluid can pass;
a flexible diaphragm located in said chamber so as to divide said
chamber into two subchambers, a first subchamber which is accessed
through said first opening and a second subchamber which is accessed
through said second opening;
a magnet attached to said flexible diaphragm, said magnet being
formed as an integral part of, and being completely enclosed by,
the material of said flexible diaphragm;
means for directing fluid whose flow is being measured through
said first opening into said first subchamber while withdrawing
the fluid whose flow is being measured through said second opening
from said second subchamber and, in response to a control signal,
reversing the subchambers into which the fluid is inserted and from
which the fluid is withdrawn, the movement of fluid into one subchamber
and out of the other subchamber causing said flexible diaphragm
to move into the subchamber from which the fluid is being withdrawn;
means for producing an output signal representative of the position
of said magnet;
means for converting said output signal to a sequence of digital
signals;
means for processing said sequence of digital signals to produce
a second signal representative of the flow rate of said fluid into
and out of said chamber;
means for comparing said flow rate to a reference flow rate to
produce a control signal representative of the difference between
said measured flow rate and said reference flow rate; and
means, responsive to said control signal, for changing the flow
rate of said fluid.
3. Structure as in claim 2 including:
means for sensing the approach of said magnet to one or the other
ends of said chamber and for switching the flow of fluid into said
chamber from said first subchamber to said second subchamber and
for switching the flow of fluid from said chamber from said second
subchamber to said first portion.
4. Structure as in claim 3 including reference means representing
the relationship between the position of said magnet and the volumetric
displacement of said diaphragm, thereby to enable the determination
of the volumetric displacement of said diaphragm from the output
signal from said means for producing.
5. Structure as in claim 4 including:
means for varying the position at which the motion of said flexible
diaphragm is reversed by determining the flow rate of said fluid
and selecting a reversal point to maximize the sensitivity of the
output signal from said means for producing as a function of the
displacement of said flexible diaphragm.
6. Structure as in claim 4 including:
means for reversing the direction of motion of said flexible diaphragm
in response to the expiration of a predetermined time.
7. Structure as in claim 2 wherein said chamber is cylindrical.
8. Structure which comprises
a chamber containing two openings therein;
a flexible diaphragm containing as a part thereof a magnet, said
diaphragm separating said chamber into a first and a second portion
and sealing said first portion from said second portion so as to
prevent leakage of fluid from said first portion to said second
portion or vice versa;
means for continuously producing an output signal representative
of the position of said diaphragm;
means for selectively controlling the flow into said first portion
and out of said second portion and for reversing the flow as desired
so that the fluid flows into the second portion and from the first
portion;
means for processing the output signal from said means for producing
including
switching path means for selectively accessing the output signal
from said means for producing
conversion means for converting said output signal from said means
for producing to a first digital signal;
buffer means for storing said first digital signal from said conversion
means;
means for inputting to said system a second digital signal representing
a desired fluid flow rate;
means for storing said second digital signal representing said
desired fluid flow rate;
processing means for operating on said first and second digital
signals to produce an estimate of the flow rate of fluid into and
out of said chamber;
means for storing a signal representative of the flow rate calculated
by said processing means; and
driver circuit means periodically actuated by said output signal
from said processing means to move a control valve either open or
closed in response to the output signal stored in said means for
storing.
9. Structure as in claim 8 including
means for storing the output signal from said processing means
representing the calculated flow rate; and
means for displaying the calculated fluid flow rate.
10. The method of measuring a fluid flow rate using a chamber containing
a first opening and a second opening through which fluid can pass
with a flexible diaphragm located in said chamber so as to divide
said chamber into two subchambers, a first subchamber which is accessed
through said first opening and a second subchamber which is accessed
through said second opening, and a magnet attached to said flexible
diaphragm, comprising:
directing the fluid whose flow is being measured through said first
opening into said first subchamber while withdrawing the fluid whose
flow is being measured through said second opening from said second
subchamber and, in response to a control signal, reversing the subchambers
in which the fluid is inserted and from which the fluid is withdrawn,
the movement of fluid into one subchamber and out of the other subchamber
causing said flexible diaphragm to move into the subchamber from
which the fluid is being withdrawn;
producing a continuous output signal representative of the position
of said magnet;
converting said output signal to a sequence of digital signals;
processing said sequence of digital signals to produce a second
signal representative of the flow rate of said fluid into and out
of said chamber;
comparing said flow rate to a reference flow rate to produce a
control signal representative of the difference between said measured
flow rate and said reference flow rate; and
changing the flow rate of said fluid in response to said control
signal.
11. Structure as in claim 2 wherein said means for producing comprises
means for continuously producing an output signal representative
of the instantaneous position of said magnet, thereby to enable
a measure of the instantaneous flow rate of the fluid entering said
chamber to be obtained during the trave of said diaphragm from any
one position to any other position in said chamber.
Description 1. Field of the Invention
This invention relates to an electronic flow control system using
an electronically-controlled positive displacement flow meter.
2. Prior Art
Positive displacement flow meters are well known. Thus British
Pat. No. 1051710 published Dec. 21 1966 discloses a positive
displacement flow meter utilizing a cylinder wherein a reciprocating
piston is controlled to move from one end to the other of the cylinder
in response to the alternate passage of the fluid whose flow is
being measured into the cylinder at one or the other end of the
piston. As fluid under pressure enters one end of the cylinder via
an inlet pipe, the piston is pushed along the cylinder and the fluid
which entered the cylinder at the other end of the piston as a result
of the previous stroke is forced into the outlet pipe. Valves in
a well known arrangement allow fluid to alternately enter one end
of the cylinder and be withdrawn from the other end of the cylinder
and vice versa.
Other positive displacement flow meters are shown in U.S. Pat.
No. 2772664 issued Dec. 4 1956 to Jones, et. al., U.S. Pat. No.
3181360 issued May 4 1965 to Hederhorst, and U.S. Pat. No. 3657925
issued Apr. 25 1972 to Gross. The '710 '664 '925 and '360 patents
all disclose reciprocating pistons as the positive displacement
member. However, the '360 patent discloses in addition, the use
of a flexible diaphragm 18 to seal a rigid piston 16 (FIG. 1 of
the '360 patent), which travels between two chambers in a cylinder,
to the cylinder wall. As one chamber is filled with gas or fluid,
a like amount of gas or fluid is expelled from the other chamber.
Reversal of the chambers in which gas is inserted and from which
the gas is removed causes the diaphragm to reciprocate back and
forth across the chamber. The number of strokes of the piston-diaphragm
combination (with the diaphragm made of Teflon or rubber), is representative
of the volume of gas passing through the meter.
U.S. Pat. No. 3974825 issued Aug. 17 1976 discloses a pump using
a flexible, pneumatically-driven diaphragm for pumping blood in
an artificial heart.
Each of the above structures has certain disadvantages. A piston
introduces inertia and friction into the flow system thereby affecting
the flow to be measured. Moreover, the fluid whose flow is being
measured often leaks past the piston. In addition, the frequency
response of the system is limited by the inertia of the piston.
For accurate flow control, a low inertia, low friction, accurate
flow meter is required as an essential part of the control system.
SUMMARY OF THE INVENTION
This invention overcomes certain of the disadvantages of the prior
art positive displacement flow meters. In accordance with this invention,
a novel, low-friction, low-inertia flexible diaphragm containing
a magnet formed as an integral part of the diaphragm is mounted
in a chamber (preferrably cylindrical). The chamber comprises part
of the flow meter in a flow control system. One or more sensing
devices are mounted on the walls of the chamber to sense the instantaneous
position of the diaphragm. A novel electronic control circuit processes
the data from the one or more sensing devices to provide measures
of the flow rate during the displacement of the diaphragm along
the cylinder. Signals are generated by the electronic control circuitry
for switching a pair of three-way valves (one valve comprising the
input valve and the other valve comprising the output valve) such
that during one-half of a cycle the input valve transfers fluid
into the cylinder on one side of the diaphragm and, during the other
half of the cycle, transfers fluid into the cylinder on the other
side of the diaphragm. The output valve is switched synchronously
with the input valve to transmit fluid from the other side or one
side of the diaphragm to the output valve.
The electronic control system includes means for amplifying the
output signals from the one or more sensing devices ("sensors")
to provide one or more signals representing the position of the
diaphragm as a function of time, means for converting the output
signals from these amplifiers to digital form, and computation means
for operating on the digitized output signals from the sensing devices
to provide control signals for controlling a second input valve
thereby to control the flow rate of the fluid being metered to within
a desired range.
As a feature of the invention, an output flow rate is not determined
unless and until the diaphragm has traveled a selected distance
or until a maximum time has elapsed, thereby to ensure that a minimum
volume of fluid has entered the positive displacement flow meter
portion of the control system and displaced the diaphragm at least
a selected amount. By dividing this displacement by the time over
which it occurs, the volumetric flow rate is obtained.
DESCRIPTION OF THE DRAWINGS
FIG. 1 illustrates schematically the chamber in which is located
the positive displacement diaphragm used for measuring flow rate.
FIGS. 2a and 2b show diaphragms connected to the chamber by bellows
and an O-ring type seal, respectively.
FIG. 3 shows schematically the connection of the sensing element
on the end of the chamber containing the diaphragm through an electronic
controller to an electronically controlled valve for maintaining
the proper fluid flow.
FIG. 4 shows schematically the relationship of the diaphragm, flow
chamber, sensing elements, the microcomputer used to compute the
control signals used to control the fluid flow rate, the fluid control
valve and the structure for controlling the fluid flow into the
flow chamber.
FIG. 5a shows the reed-switch flow reversal control circuit useful
with this invention.
FIG. 5b shows a circuit useful with the circuit of FIG. 5a for
producing an output voltage proportional to flow rate.
FIG. 6 shows a typical output voltage versus displacement curve
used with this invention for calibrating the output voltage from
a sensing element on the end of the cylinder to the position of
the diaphragm within the cylinder.
FIG. 7 shows the microcomputer based flow controller structure
connected to the sensing element on the end of the chamber containing
the diaphragm.
FIG. 8 shows the relationship of the various circuit boards used
to process the signals from the sensor and to control the settings
of the three-way flow valves 1 and 2 (FIGS. 1 and 7) which allow
the fluid to pass into one or the other of the chambers in the cylinder,
and which control the setting of the fluid control valve.
FIG. 9 shows a sensor circuit useful with this invention.
FIG. 10 shows the motor and solenoid driver circuits used with
this invention.
FIGS. 11a and 11b show schematically the thumbwheel digit input
circuit and the display circuits used with this invention;
FIG. 12 shows the logic diagram for the computer program used in
the microcomputer shown in FIG. 7 as part of the structure of this
invention; and
FIG. 13 shows schematically the method and structure for fabricating
the diaphragm with an enclosed magnet used with this invention.
DETAILED DESCRIPTION
While this invention will be described in conjunction with specific
components in the electronic control circuitry, this description
is exemplary only and is not intended to limit the scope of the
invention.
Turning now to FIG. 1 a flow chamber 12 containing a right chamber
12a and a left chamber 12b separated by diaphragm 4 is shown schematically.
Diaphragm 4 has mounted on its center as an integral part thereof
a magnet 5. Magnet 5 is completely coated with the material of which
diaphragm 4 is constructed to protect magnet 5 from the fluid (possibly
corrosive) being metered. On the left face of chamber 12 is a sensor
6b and on the right end of chamber 12 is a second sensor 6a. If
desired, only one sensor (either 6a or 6b) can be used. While shown
as reed switches, preferably, these sensors comprise Hall effect
devices of the type known as LOHET (for "linear output, Hall-effect
transducer") sensors such as described in more detail in Electronic
Design 19 dated Sept. 27 1979 on page 23. This article is incorporated
herein by reference. Other Hall effect sensors can, of course, also
be used with this invention.
The input flow is transmitted through three-way valve 1 (shown
schematically) arranged to allow the flow to pass into line 9 connected
directly to left chamber 12b. The entry of fluid into left chamber
12b displaces diaphragm 4 to the right thereby expelling fluid in
the right chamber 12a through line 10. Three-way outlet valve 2
is, in this mode, adjusted to allow the fluid flowing from line
10 to pass directly into the output line.
During the entry of fluid into left chamber 12b, diaphragm 4 and
magnet 5 move to the right toward Hall effect sensor 6a. The output
signal from Hall effect sensor 6a is a function of the position
of magnet 5 which in turn, is proportional as a function of time,
to the rate of fluid flow through line 9 into left chamber 12b.
As magnet 5 moves closer to sensor 6a due to the displacement of
diaphragm 4 to the right, the Hall effect sensor 6a produces an
output signal uniquely related to the position of magnet 5. This
position as a function of time is a function of the flow rate. Sensor
6a produces an output signal which is monitored in a manner to be
described later by the electronic control circuits. As magnet 5
reaches its rightmost position, a control signal is generated switching
valves 1 and 2 such that the input gas now flows through line 7
into rightmost chamber 12a while the fluid in leftmost chamber 12b
is expelled from this chamber through line 8 and outlet valve 2.
Consequently, diaphragm 4 is forced to the left and magnet 5 now
travels away from Hall effect sensor 6a and toward Hall effect sensor
6b. The output voltage from Hall effect sensor 6b is also uniquely
related to the position of magnet 5. Since each sensor's output
voltage is uniquely related to the position of the diaphragm, by
measuring time independently, the sensor output voltage and thus
the diaphragm position is determined as a function of time which
allows the flow rate to be calculated given a calibration curve
of volume versus diaphragm position (as shown in FIG. 6).
The use of a flexible diaphragm as part of a fluid meter is based
on our discovery that, contrary to expectation, the diaphragm position
as a function of volume of fluid in each chamber is predictable
and repeatable each cycle despite the fact that the diaphragm is
flexible and "loose" (i.e., the diaphragm is designed
to interfere minimally with the fluid flow). Thus, diaphragm 4 serves
to separate the left chamber 12b from the right chamber 12a while
supporting magnet 5 the position of which is used to measure the
fluid flow rate. The calibration of the position of magnet 5 to
the output signals from sensors 6a and 6b is essential to the proper
operation of this system. As is apparent from FIG. 1 diaphragm
4 can easily be replaced should it wear out by separating the two
halves of cylinder 12 removing the old diaphragm and inserting
a new diaphragm. Typically, to ensure a leakproof connection, diaphragm
4 has two "O-rings," an inner "O-ring" 210 and
an outer "O-ring" 211 as shown in FIG. 13 (which shows
the die used to fabricate the diaphragm).
It is apparent from the above description that the position of
diaphragm 4 as a function of time is directly proportional to the
volumetric flow rate of the fluid. By knowing the temperature and
the pressure of the fluid the mass flow rate can be determined in
a well known manner.
When sensors 6a and 6b comprise reed switches, the output signals
from these switches occur with a frequency directly proportional
to the flow rate and are used directly to actuate electronic circuitry
for both valve actuation and feedback or other fluid control purposes.
On the other hand, when the output sensors 6a and 6b are Hall effect
devices, a continuous output signal is generated from each of these
sensors. This output signal is related to the flow rate. By knowing
the calibration curve of output signal level versus displacement
of the diaphragm 4 and magnet 5 the volumetric flow rate can be
calculated continuously as a function of the output signals from
sensors 6a and 6b.
Diaphragm 4 is preferably made of an elastomeric material (viton
is preferred but neoprene, silicon rubber and butyl rubber can also
be used, as appropriate) which is suitable for use for temperatures
up to 150.degree. C. (180.degree. C. for viton). For higher temperatures
metal bellows can be used in place of the elastomeric diaphragm.
The choice of the diaphragm 4 material depends on the particular
fluid that is to be used, and the preferred operating temperatures.
The differential pressure across the diaphragm is small in comparison
to the working fluid pressures. Preferably this pressure difference
is negligible. Viton is appropriate for temperatures between -40.degree.
C. and 180.degree. C. and pressures between vacuum and 300 atmospheres.
The diaphragm with encapsulated magnet is made by a compression
or transfer molding process. In this process (illustrated in FIG.
13), a metal die 200 containing upper section 200a and lower section
200b, is fabricated to the exact dimensions of the diaphragm. Die
section 200b includes a central cavity 200c to house the magnet
212 (corresponding to magnet 4 in FIG. 1). A thin disc of the elastomer
material 213 having the same or slightly larger diameter as the
magnet disc 212 (approximately 1/2 inch and preferably formed of
Alnico 8 or Cemarium cobalt) is first inserted into the magnet cavity
200c and the magnet 212 is placed on top of this thin disc. A pre-weighed
amount of the elastomer raw material 214 is then put on top of the
magnet 212 the material is heated to about 400.degree. F. when
Viton is the material and pressure is applied to it by the upper
half 200a of the die 200 to mold and cure the elastomer 214 to the
desired diaphragm shape. The pressure is merely that sufficient
to achieve the desired result. The compression is conducted at an
elevated temperature whose magnitude depends on the particular elastomer
and results in elastomer material 214 and 213 assuming a substantially
uniform thickness (typically 0.015" to 0.020" when Viton
is the material) and consistency throughout the die 200. This operation
ensures complete encapsulation of the magnet by the elastomer and
excellent control of the diaphragm thickness. A Viton diaphragm
typically has a durometer of fifty (50).
A particularly suitable valve for use as valves 1 and 2 (FIGS.
1 and 7) is the D30 three-way valve made by Precision Dynamics Company.
This valve switches within about eight milliseconds and is fabricated
out of 303 and 430 stainless steel with viton elastomer for the
seals and O-rings. Other valves are also appropriate depending on
design requirements.
The flow measurement by this device is reproducable well within
one percent. Diaphragms or bellows can be easily replaced before
fatigue sets in.
FIGS. 2a and 2b show two configurations for the diaphragm inside
the cylinder. In FIG. 2a, a diaphragm has a magnet mounted internally
to it which is protected by the material of the diaphragm. The diaphragm
is then connected and sealed to the ends of two bellows. Each bellows
is capable of contracting or expanding in response to lateral movement
of the diaphragm in response to fluid entry into one or the other
chambers of the cylinder. Thus when fluid enters the left chamber
of the cylinder, the diaphragm and the magnet move to the right
and the fluid in the right chamber is expelled. When fluid enters
the right chamber, the magnet and the diaphragm move to the left
expelling the fluid in the left chamber. A magnetic sensor mounted
on the cylinder detects the motion of the magnet and thereby produces
an output signal proportional to the position of the diaphragm.
Typically the bellows comprises either a metal or polymer material
and the diaphragm likewise comprises a polymer or metal material.
FIG. 2b shows a different construction wherein the diaphragm is
attached between two portions of the cylinder. Again, a magnet 5
is mounted on the diaphragm 4 and coated with the diaphragm material
thereby to protect it from the fluid whose flow is being measured.
The entry of fluid to the left chamber 12b displaces diaphragm 4
and magnet 5 to the right as with the structure of FIG. 2a and similarly
the entry of fluid to the right chamber displaces the diaphragm
4 and magnet 5 to the left, again as with the structure of FIG.
2a. The diaphragm is constructed such that its motion to the right
or left occurs with very little friction or resistive force. Thus
the flow of the fluid is not disturbed by the presence of the diaphragm.
In addition, the weight of the diaphragm is kept very low thereby
minimizing the inertia of the diaphragm. Magnetic sensors 6b and
6a are mounted on the left and right walls of the cylinder to detect
the movement of the diaphragm-mounted magnet 5 to or from a given
wall. In both FIGS. 2 a and 2b a conventional set of two three-way
flow valves (such as shown in FIG. 1) is used to first route the
fluid into one chamber and out of the other chamber and then to
reverse this pattern.
Contrary to the prior art meters using a free flowing piston, wherein
the fluid whose flow is being measured can often leak from the left
chamber to the right chamber and vice versa, this invention uses
a flexible, low inertia diaphragm to seal one chamber from the other
and thereby to prevent leakage while at the same time providing
a relatively instantaneous measure of fluid flow rate.
FIG. 3 shows the structure of FIG. 1 (or the chambers of FIGS.
2a and 2b), in combination with an electronic controller 30 a DC
stepping motor 31 and a clutch or coupling 32 joining the stepping
motor to a valve 33 (typically a needle flow valve) for controlling
the flow of fluid. The DC stepping motor 31 adjusts the position
of the valve 33 until the flow rate detected by the meter corresponds
to a flow rate setpoint input to the electronic controller 30. As
will be shown later, the electronic controller 30 preferably comprises
a microprocessor controlled digital circuit with an analog-to-digital
converter, a sampling circuit, buffer stores, memory and selected
input and display elements. The electronic controller 30 also controls
the setting of the two three-way valves 1 and 2 to ensure that the
fluid flow into the meter and from the meter is reversed at appropriate
times to obtain maximum accuracy in the flow measurements. The system
provides accurate and responsive feedback (or feed forward if desired)
control. In the simplest form, the flow into one side of the cylinder
12 is reversed when the diaphragm reaches its extreme position one
way or the other. The flow rate can, in this embodiment, be simply
measured by the time it takes for the diaphragm to travel between
the two extremes. That is, the flow rate is equal to the volume
between the two extreme positions of the diaphragm 4 divided by
the time taken by the diaphragm to travel between these two positions.
FIGS. 4 and 5a show an embodiment of this invention using a reed
switch to detect the maximum displacement of the diaphragm adjacent
each of the two walls of the cylinder 12. The reed switch signal
is transmitted to a microcomputer control 30 which then calculates
from the time taken for the two reed switches 46a and 46b to be
sequentially actuated, the measured flow rate. This measured flow
rate is compared to a reference flow rate set into memory and the
difference is used to set a control valve 33 to bring about the
proper flow rate. Solenoids and solid state relays are also actuated
by the microcomputer control 30 to reverse the fluid flow into cylinder
12 at the maximum displacement points of the diaphragm 4.
FIG. 5a shows a circuit schematic of the structure used to respond
to the change in state of the reed switches 46a and 46b due to the
approach of magnet 5. A flip flop comprising two 7400 NAND gates
41 and 42 has as one input signal to each of the NAND gates the
output signal representing the state of a given reed switch. Thus
normally, when reed switch 46a is open, the signal level on input
lead 41a to NAND gate 41 is at a high level corresponding to the
5 volt DC supply voltage. The other input lead 41b is coupled to
the output lead 41d from the other NAND gate 42. The output signal
on lead 41d is normally high level for at least one low level input
signal. Thus, with the signal on output lead 41c low level, the
signal on input lead 42b is low level and the signal on output lead
41d from gate 42 is high level, thereby holding the output signal
from gate 41 at low level. When reed switch 46a is closed, the input
signal on lead 41a to NAND gate 41 goes low thereby driving the
output signal from NAND gate 41 to a high level. This high level
output signal is transmitted to input lead 42b of NAND gate 42 and
drives the output signal from NAND gate 42 to a low level thereby
latching up the output of NAND gate 41 to a high level. The high
level output signal from NAND gate 41 is amplified by amplifier
43 and turns off transistor Q10 (2N2904) thereby changing the current
through a coil to close two relay control switches thereby activating
solenoids to change the settings of the two three-way valves 1 and
2 shown in FIG. 1 and thus to reverse the flow. When diaphragm 4
arrives at the other side of the cylinder 12 reed switch 46b closes
thereby driving the output signal from NAND gate 42 from low level
(corresponding to two high level input signals) to high level thereby
driving the output signal from NAND gate 41 to low level. This low
level output signal is also amplified and used to turn on transistor
Q10 thereby again activating a relay to reverse the two three-way
flow valves 1 and 2 (FIG. 1) and thereby again to reverse the fluid
flow.
FIG. 6 shows a typical curve of output signal from a Hall effect
device (such as sensor 6b or 6a in FIG. 2b) versus position of the
flow diaphragm 4. This curve is substantially linear over short
portions but gradually flattens out as the diaphragm 4 moves away
from the Hall sensor. The slope of voltage versus position (i.e.,
displacement) is negative such that the output voltage as a function
of the distance of the diaphragm from the sensor increases with
decreasing distance of the diaphragm from the sensor. Operation
of the sensor in a region of substantially steep slope gives greater
sensitivity and accuracy to the measurement of the fluid flow than
does operation in a region of flatter slope. As will be discussed
shortly, this feature is used to enhance the accuracy of certain
flow measurements.
FIG. 7 discloses the preferred embodiment of this invention using
a microcomputer-controlled digital circuit to provide proper feedback
signals to control the flow rate. The microcomputer 81 operates
on an output signal produced from the Hall effect sensor 6b mounted
on the wall of the chamber 12 containing the flexible diaphragm
4 with magnet 5 formed as an integral part thereof.
FIG. 7 shows in block diagram form the electronic control circuitry
used to process the information produced by the sensor 6b attached
to chamber 12 containing the flexible diaphragm 4 used to measure
flow rates in accordance with the invention. The output signal from
Hall effect sensor 6b is transmitted through sensor interface circuit
73 to analog switch 74. Analog switch 74 comprises in one embodiment
a well-known device for selecting a particular signal path in response
to digital input signals (input to switch 74 on lines 74a from control
logic 76a) for the purpose of passing a selected signal from either
a pressure transducer 72a and a temperature transducer 72b (for
the purpose for allowing the mass flow rate to be calculated from
the volumetric information produced from the signal generated by
sensor 6b attached to flow meter chamber 12) or from sensor interface
circuit 73.
The output signal from analog switch 74 is transmitted to an analog-to-digital
converter 75 of well known construction. Preferably, this converter
comprises a twelve bit converter such as is used with the Analog
Devices RTl-1220 Data Acquisition Board 84 (FIGS. 7 and 8). The
output signal from A-to-D converter 75 comprises a digital signal
representing any one of 4096 possible signal levels. Of course,
by selecting an A-to-D converter containing a different number of
bits, a different level of accuracy can be obtained.
The output signal from A-to-D converter 75 is then transmitted
to input buffer 76e. Input buffer 76e buffers the output signal
from A-to-D converter 75 and holds this information until a request
for this information is received from microcomputer 81. Microcomputer
81 comprises a Mostek MK79612 CPU and Timer Board which contains
a CPU corresponding to the well known Z80 microprocessor. The MK79612
is shown in more detail in the Mostek publication entitled "MD
Series Microcomputer Modules, Operations Manual for MDX-CPU 1"
and subtitled "Z80 Central Processor Module MDX-CPU 1",
copyright 1978 by Mostek Corporation. This document is also incorporated
herein by reference. FIG. 8 shows the relationship of the Mostek
MK79612 to the remainder of the circuit, the components of which
are also shown in more detail in FIG. 7.
The control logic 76a, input buffer 76e, A-to-D converter 75 and
analog switch 74 comprise the Analog Devices RTI 1220 data acquisition
board 84 (shown as such in FIG. 8).
The output latches 76b, 76c and 76d comprise the Pro-Log 7601 TTL
I/O board 83 also shown in FIG. 8. The program memory of the microcomputer
81 comprises the Mostek MK79604 Eprom/UART board 82 shown as such
in FIG. 8.
The data bus interconnecting the CPU timer and the Eprom/UART portions
of the structure to the Pro-Log 7601 TTL I/O board and the Analog
Devices RTI 1220 data acquisition board comprises the STD bus, a
standard bus used by Mostek and Pro-Log for interconnecting the
components of a typical eight bit microcomputer system using an
eight bit microprocessor (such as the Z80). The operation of the
STD bus is described, for example, in the Pro-Log publication copyrighted
1979 entitled "Series 7000 STD Bus, Technical Manual."
The operation of this bus is thus well know in the art and will
not be described in detail. The above cited Pro-Log technical manual
is incorporated herein by reference.
Control logic 76a generates several sets of output signals. First
this logic generates a set of signals for controlling the setting
of analog switch 74. These signals are transmitted on lead 74a.
Logic 76a is driven by signals taken off the STD bus and generated
by microcomputer 81.
Output latch 76b (part of TTL I/O Board 83) generates a signal
to actuate driver circuit 79b which in turn produces signals which
operate three-way valves 1 and 2. Driver circuit 79b is shown in
more detail in FIG. 10 and will be described below.
Additional signals from output latch 76b also actuate driver circuits
79b which in turn drive stepper motor 79c. Stepper motor 79c controls
the setting of control valve 79d which controls the flow rate though
valves 1 and 2. Thus the output of latch 76b (derived from microcomputer
81) controls the setting of valve 79d in response to the flow rate
measured from the signals produced by sensor 6b.
The reference signal used to determine the magnitude and direction
of change in the position of valve 79d is derived by comparing the
signal representing the flow rate computed from the signals from
sensor 6b with another signal placed in input buffer 76c and derived
from the setting of thumbwheel 77. The setting of thumbwheel 77
can be varied as desired by the operator thereby changing as desired
the setting of control valve 79d. Basically, thumbwheel 77 converts
a setting visible to the user to four binary-coded decimal digits
(corresponding to 16 bits). These signals are held in input buffer
76c and are used by microcomputer 81 to calculate the proper control
signal to be transmitted through output latch 76b to control the
setting of valve 79d.
Output latch 76d is configured to retain the output signals from
microcomputer 81 corresponding to the latest reading of the flow
rate. The signals retained in latch 76d activate display 78 to display
the latest value of flow rate calculated by microcomputer 81.
Structures corresponding to those shown in FIG. 7 are shown in
FIG. 8. Those elements shown in block diagram form in FIG. 8 function
as described above in connection with FIG. 7 and thus will not be
described in more detail here.
FIGS. 11a and 11b show in more detail the thumbwheel circuitry
110a through 110b and the display circuitry 114a through 114b used
with this invention. The thumbwheel is capable of inputting four
digits into the circuit. Each digit circuit 110b, 110c and 110d
comprises a replica of the specific circuit 110a shown figuratively
for digit one. This circuit is of a type well known in the arts
and thus will not be described in detail except to say that if the
thumbwheel is set, for example, on a 7 the switches within the
thumbwheel corresponding to the 4 2 and 1 level signals (connected
to the 5 V supply voltage through resistors 112b, 112c and 112d,
respectively), are closed thereby to provide an output signal for
the TTL input at a level corresponding to the complement of 7 (negative
logic). Any other decimal digit from 0 to 9 is selected by closing
the corresponding switches to provide an output signal of the proper
level.
The display likewise comprises a four digit display. Each display
114a, 114b, 114c and 114d comprises a well known decoder driver
for taking a BCD input and converting it to a digital number ranging
from 0 to 9. Typically, each display comprises an HP 7304 display
of a type well known in the art.
FIG. 10 shows in more detail a typical drive circuit of the type
used to drive the stepper motor 79c and the solenoid driven three-way
valves 1 and 2 (the latter two valves being used as described above
to reverse the flow in chamber 12). The driver circuits are driven
by signals on input leads d0 through d5 derived from the TTL I/O
card (FIGS. 7 and 8). The particular stepper motor windings to be
energized depend upon the setting of the stepper motor (a record
of which is recorded in the RAM memory of microcomputer 81) and
the direction in which it is desired to move the stepper motor.
The stepper motor will move a standard distance (typically (1/200)th
of a revolution) once each step. A timer in the microcomputer allows
the generation of a new step after a fixed time has elapsed. In
the embodiment of this invention, this time is 7.5 milliseconds.
The timer on the microcomputer board 81 interfaces with the microcomputer
central processing unit (CPU) via interrupts. The timer generates
periodically signals which are transmitted to the CPU on a separate
line thereby to activate an interrupt circuit within the CPU through
an interrupt pin. This signal then activates the CPU to calculate
the next setting for the stepper motor; that is the CPU determines
whether the stepper motor should be driven such that the control
valve 79d is either opened or closed based upon the last flow rate
reading present in the microcomputer when compared to the thumbwheel
setting. Typically, the stepper motor is driven by a sequence of
signals on input leads D0 through D3 corresponding to hexadecimal
5 9 6 A. Thus if the setting of the stepper motor corresponds
to a 9 then the hexadecimal encoded binary 6 transmitted on input
leads D0 through D3 will activate the stepper motor to move one
setting in the proper direction to open. Should the stepper motor
be instructed to close rather than open, then a hexadecimal encoded
5 would be transmitted in binary form on leads D0 through D3. In
this latter case wherein the most significant digit corresponds
to a zero, the second most significant digit corresponds to a one,
the third most significant digit corresponds to a zero and the fourth
most significant digit corresponds to a one (corresponding to the
signals 0101 on leads D3 through D0 respectively), then the "1"
on lead D0 will drive the output signal from inverter 101 to a low
level thereby turning on transistor Q1 and energizing winding W1.
Diode D1 prevents voltage pulses generated by changing the current
through winding W1 from burning out transistor Q1 and resistors
102 and 103 comprise pull-up and base-drive resistors, respectively.
FIG. 9 shows the circuitry associated with the Hall-effect sensor
6b. The magnetic flux from the magnet 5 mounted within, and as an
integral part of, diaphragm 4 is converted to a voltage and amplified
to produce an output signal. Hall-effect sensor 6b preferably comprises
part number 633SS2 made by Microswitch, Inc., a division of Honeywell.
The output signal from microswitch 90 is transmitted via twisted-pair
cable 94 to Analog Devices data acquisition board 84 (FIGS. 7 and
8). There this output signal activates the processor to operate
in a way previously described. Pull-down resistor 92 (100 ohms)
is connected between the fifteen volt power supply in series with
a 5.1 volt zener diode 91 to ground. Variable resistor 93 (100K)
is connected across the node between resistor 92 and diode 91 and
ground. Resistor 93 is used to produce a bias offset for the other
lead in the twisted pair 94 transmitting the output pulse from microswitch
90 to the data acquisition board 84. This bias offset comprises
a way of compensating for the fact that the curve of flux versus
voltage for the Hall-effect sensor does not pass through the origin.
By adjusting the setting of variable resistor 93 the input signal
to the control circuitry is offset (i.e., zero adjusted) to provide
a signal within a desired range of magnitude.
Operation of the above described structure is implemented by a
software program. The Microl high-level language version of this
program is attached to this application as Appendix A. This program
uses the Microl Language which is described in the User's Manual,
Microl Language, dated Mar. 8 1979. This Manual is incorporated
herein by reference.
FIG. 12 illustrates the logic flow diagram of the computer program
devised to operate microcomputer and timer 81 (FIGS. 7 and 8) in
accordance with this invention. The computer program is designed
to operate on the output signals from sensors 6b as processed by
the data acquisition board 84 to determine flow rates. To do this
one of several algorithms is employed. In the preferred mode of
operation the flow rate is measured on a relatively continuous basis
as a function of the displacement of diaphragm 4 across chamber
12 with time. This preferred mode does not depend upon the reversal
of the direction of motion of diaphragm 4 to measure the flow rate
but does set an optimum reversal point for changing the direction
of motion of diaphragm 4 to ensure that at least a minimum voltage
difference from sensor 6b as a function of the location of diaphragm
4 is obtained in a given time. In the limit, the minimum voltage
difference per sample can be expanded to a value corresponding to
the reversal point of the diaphragm. In this situation, the preferred
mode reduces to the second, reversal dependent mode. In essence,
the software emulates the operation of the reed switch mode, but
in addition provides an autoranging capability (to be described
later).
The second mode of operation contemplated for use with this invention
is a reversal dependent mode. The program parameters are set to
cover the highest and lowest expected flow rates (which correspond
to the shortest and longest times for diaphragm 10 to complete one
cycle). The setting of the reversal position for reversing the direction
of motion of diaphragm 4 closer to its central neutral position
results in a decrease in the cycle time. In this mode of operation,
the reversal point is set to ensure at least one sample within some
specified time for the lowest expected flow rate. In one embodiment,
this specified time is on the order of two (2) seconds. This time
must be much greater than eight milliseconds, the switching time
of the valves.
In describing the operation of the program of this invention, certain
conventions must be defined. Thus in the following description,
a "conversion" comprises one look at the voltage produced
by sensor 6b reflecting the position of diaphgram 4.
The average of four conversions comprises one "reading."
The signal representing the average of these four conversions is
a digital signal as is the signal representing each conversion.
Continuous readings are then made of the output signal from sensor
6b on a periodic basis until a change in voltage from 6b in excess
of a minimum voltage change (DELMIN) is obtained. At this time,
the systems has completed one "sample."
The system actually takes four conversions at the start of operation
and continues taking groups of four conversions and averaging each
group of four. Since it takes about 25 microseconds for each A-D
conversion and there are four conversions per reading, 100 microseconds
are required at a minimum for one reading. After each sample, the
program tests to determine whether the direction of movement of
the diaphragm should be reversed. Every seven and one-half milliseconds
or thereabouts there is an interrupt and for a few microseconds
the program then decides whether stepper motor 79c should be instructed
to step valve 79d in one or another direction.
Every flow rate calculated from each sample is sent to a memory
location called "FLOW." By comparing the calculated flow
rate to the set point (placed in thumbwheel 77 FIG. 7) using the
relationship "set point minus flow rate," a difference
signal is obtained. Multiplying this difference signal by some non-negative
control function gives a gain for use in determining the proper
change in position of valve 79d. This gain is placed in "STPSIZ."
The non-negative function is in turn a function of "set point
minus flow rate" and can also be a function of one or more
previous "set point minus flow rate" measurements. A separate
interrupt program called "s-timer" looks every 71/2 milliseconds
at STPSIZ to decide whether to step up or to step down the valve.
In the above calculation, the program uses the subroutine "GETFLO"
to calculate the flow. Subroutine "NEWVLV" corresponds
to the control function which generates the new valve position.
"STPSIZ" corresponds on a one-to-one basis to the steps
on a valve. It then takes 7.5 milliseconds for the valve 79d to
make one step. The program does not sample while the valve is opening
or closing.
Once sampling is started, a sample is always obtained before the
direction of motion of diaphragm 4 is reversed and before it is
necessary to reverse the motion of diaphragm 4. An important part
of the system is the calibration table calibrating the volumetric
displacement of the diaphragm to the output signal from sensor 6b.
To ensure an accurate conversion of the output signal from sensor
6b into a flow rate, the volumetric displacement of the diaphragm
4 represented by any given signal from sensor 6b must be determined
accurately. Techniques for doing this are well known. One such technique
comprises the discharge of a fluid from the output line into a vertical
stand pipe in which the discharged fluid or gas displaces a colored
liquid. By calibrating the stand pipe in terms of milliliters or
cubic centimeters, for example, the volumetric displacement of diaphragm
4 can be measured with great accuracy. FIG. 6 shows a curve of voltage
from sensor 6b versus volumetric displacement of diaphragm 4.
While the structure in FIG. 7 has been described as using one Hall
effect device sensor, using two Hall effect devices, one on each
side of chamber 12 yields twice the sensitivity to the measurement
of flow rate. By using two Hall effect devices, the range of the
devices is doubled. The diaphragm 4 can then travel a maximum excursion
distance in chamber 12 and during its travel in the left portion
of the chamber, sensor 6b is used while during its travel in the
right portion of chamber 12 sensor 6a is used. In this manner, both
sensors are operated in their range of maximum sensitivity and thus
a sensor operating in its maximum range of sensitivity is used over
the full travel of the diaphragm 4.
The thumbwheel 77 (FIGS. 7 and 8) is used to place into the system
a "set point" which determines the flow rate to be allowed
by valve 79d. The set point is placed into the thumbwheel. The system
then measures the difference between a new set point and the old
set point. The program then loops back through the set point change
logic and produces an output signal proportional to the difference
between the new set point and the old set point. If there is no
change in the set point during this loop back the system then looks
at a measured variable called "sampling."
The concept of "sampling" means taking a reading, calculating
a diaphragm displacement from the reading, starting a timer within
the system, measuring the output voltage from sensor 6b and testing
to determine whether the minimum change in voltage is greater than
a selected value. In the preferred embodiment, the minimum voltage
change required to be detected from sensor 6b before a flow rate
is calculated is 200 millivolts. Thus in reading voltage when sampling,
a sufficient voltage change is allowed to ensure that the minimum
voltage change has occurred or a maximum time has elapsed without
having this minimum voltage change occur. If in this maximum time
the system does not record a minimum voltage change (DELMIN) the
system then assumes zero flow and opens the valve 79d (FIG. 7).
If the proper minimum voltage change (DELMIN) is obtained, the
system then takes the measured voltage from sensor 6b and refers
to the proper place on the voltage-displacement calibration curve
(FIG. 6) to calculate the displacement change over time. From the
displacement change over time and the time, the average flow rate
over this time is calculated. Referring now to the calibration curve
(FIG. 6) it is apparent that as the flow rate becomes lower, the
displacement in a given time becomes less and the reversal point
on the curve must move from right to left. That is, the diaphragm
travels a smaller distance in a given time for a low flow rate than
for a high flow rate and thus to maximize the sensitivity of voltage
versus displacement, one must operate on the left-most portion of
the curve rather than on the flatter right portion of the curve.
The right most (farthest from sensor 6b) reversal point is selected
to ensure that magnet 5 does not hit the wall and that there is
time to generate a voltage change equal to DELMIN between the reversal
point and the wall. The reversal point closest to sensor 6b is selected
by a formula C.sub.1 -flow.times.C.sub.2 where C.sub.1 and C.sub.2
are selected constants equal to 3800 and 5 respectively, in the
program of Appendices A and B.
The logic flow diagram shown in FIG. 12 describes the logic of
the software program shown in Appendices A and B. The program is
implemented in Microl, a block structured language for implementing
programs on the Z80. This language is written in PASCAL and modeled
after PASCAL. The Microl compiler produces the Z80 assembly language
which is assembled and then linked with other modules to form the
final object code. Microl is a publicly available high level language
and is described in a document entitled "User's Manual Microl
Language, Mar. 8 1979." This document is herein incorporated
by reference.
Referring to FIG. 12 at the start of the program, power on reset
results in the program going to "JRESET" which initializes
the processor (CPU) and timer, stepper motor and valve. Essentially
the program closes the valve and turns off the timer. The program
then jumps to the main control algorithm represented in the Microl
procedure by MlSTRT.
The main control algorithm calls these other Microl procedures
to implement the computation of flow or displacement or new valve
setting. MlSTRT and these other Microl procedures call numerous
short assembly language procedures to implement low level functions
directly on the hardware (such as moving the stepper up or down
one step or outputting to the display LEDs from a specified register).
The program includes routines for handling two vectored interrupts
(i.e., means in the hardware and program for jumping to different
locations in memory when the two interrupts occur) generated by
the timer. One interrupt is activated by the timer to keep track
of elapsed time during a sample by incrementing a register or memory
location every one-half microseconds. Another interrupt is activated
by the stepper timer which generates an interrupt every 71/2 milliseconds
and vectors to the interrupt servicing procedure (STIMER) to determine
from memory location (STPSIZ) whether to move the stepper motor
up or down or not at all. Once the program has been initialized,
the program checks to determine whether or not the system is sampling.
If the answer is "no," the program then checks to determine
whether the flow into chamber 12 should be reversed. This test,
which comprises measuring the output of sensor 6b to determine whether
diaphragm 4 has reached its minimum or maxinum point of excursion,
yields either a "yes" or "no." If the answer
is yes, the flow should be reversed. The program then initiates
the subroutine which switches the solenoids which drive valves 1
and 2 (FIGS. 1 and 7) through driver circuit 79b and initiates a
delay to ensure that the flow has actually reversed and the transients
in the system have settled before starting to sample to measure
flow rate during the reverse motion of the diaphragm. If the logic
indicates there is no need to reverse flow, the system then determines
whether or not there has been a set point change. If "yes,"
the valve 79d is adjusted by measuring the difference between the
new set point and old set point. If the answer is "no,"
the system moves directly to the logical blocker stepper motor.
If the valve was adjusted in response to a set point change, after
the valve adjustment is completed the system also moves to the stepper
motor logic. The stepper motor logic determines whether or not the
stepper motor has moved to its desired position. If the answer is
"no," then the system goes back to the initial sampling
logic block. If the answer to "stepper motor done?" is
"yes," the system begins a new sample and returns to sampling.
If the output of the system sampling block is "yes,"
the system determines whether the output of the sampling compared
to a reference sample is greater than DELMIN. If the answer is "yes,"
the system determines whether or not the time is greater than the
minimum time. If the answer is "yes," then the system
computes a new flow rate and, depending upon the time, adjusts the
auto range of the system. This last adjustment is an adjustment
to the reversal point of diaphragm 4 by changing the level of the
output signal from sensor 6b at which the direction of motion of
diaphragm 4 is reversed. Once auto range has been completed, a new
valve position is calculated from the flow rate measurement compared
to the set point and valve 79d is appropriately adjusted. Finally
an end of sample signal is produced which then causes the program
to initiate sampling again.
If on the other hand the output signal from the DELMIN logic determines
that the difference in the output signal from sensor 6b is less
than the minimum required, a test is run to determine if the time
between this sample and the previous sample is greater than the
maximum time. If the answer is "yes," then a new flow
rate is calculated and the auto range is adjusted as in the preceding
sequence. If the answer is "no," then the system tests
to determine if there has been a set point change. If the answer
is "yes," the valve is adjusted as described above. If
the answer is "no," the system loops back to run another
test to determine if the new sampling voltage is greater than DELMIN.
The particular lines on the flow sample in the microcomplier version
of the program shown in Appendix A which implement the particular
logic block in FIG. 12 are shown on FIG. 12.
Appendix B gives the assembly language procedures called for by
the Microl procedures in Appendix A.
As a feature of this invention, for applications not requiring
the precision and versatility of computer controlled logic circuitry,
several lower cost embodiments are possible using either the reed
switches or the Hall effect devices. A multiple segment down counter
can approximate the flow rate by counting down from a "maximum
flow" between flow reversals. The rate of down counting is
varied after a time interval lapses (the device will count slower
after each elapsed interval). By this method, the function f=(c/t)
is approximated by several straight line segments (where f equals
the flow rate, c equals the count and t equals the time interval
which has elapsed).
The method achieves moderate accuracy over a limited range with
good repeatability and produces a direct digital output for very
low cost when implemented using discrete logic.
A further improvement is to continuously vary the rate of down
counting by using, for example, a capacitive discharge to drive
a VCO whose oscillations are then used to drive the down count.
It may be necessary to use a capacitive charge if the VCO has a
negative voltage to frequency characteristic. In this case a baseline
(i.e., zero bias) count would also be taken and subtracted to produce
the direct digital representation of the flow.
The capacitive discharge itself can produce an approximation of
the flow as a voltage. Accuracy and range is limited but can also
be improved by multiple interval method described above with respect
to the countdown technique.
The above embodiments disclose the use of sensing devices on the
external walls of the chamber. Under some circumstances, the Hall-effect
sensor can be molded into the diaphragm (with lead wires also in
the diaphragm) and the magnets can be placed on the external walls
of the cylinder. Compensation for temperature can be done in either
the transducer or externally by a computer using a temperature input.
Likewise, compensation for pressure changes can also be done in
a similar manner.
FIG. 5b illustrates a circuit which produces an output signal inversely
proportional to time and therefore provides an output signal directly
proportional to the flow rate. In this circuit, a reference voltage
is applied to the negative input lead of operational amplifier 54
through input resistor 51. The positive input lead of operational
amplifier 54 is connected through resistor 57 to ground. The reference
signal is integrated by capacitor 53 in a well known manner when
the gate voltage on reset FET transistor 52 is such that this transistor
is turned off (i.e., non-conducting). The negative going output
signal from operational amplifier 54 across capacitor 53 is then
applied to the gate of depletion mode N-channel FET transistor 56.
This transistor is a linear FET device with the source-drain resistance
linearly proportional to the gate voltage. Thus as the gate voltage
decreases linearly with time, the source-drain resistance of this
transistor (a depletion mode device is normally conducting) increases
with gate voltage. Thus, the output voltage from operational amplifier
55 (the negative input lead of which is connected to a reference
voltage through input resistor R.sub.G and the positive input lead
of which is connected to a suitable reference) drops hyperbolically
with time (since its output voltage is proportional to 1/R.sub.F
where R.sub.F is the source to drain resistance of linear FET 56).
Accordingly, at the time FET transistor 52 is turned on to discharge
capacitor 53 the output voltage from operational amplifier 55 is
proportional to the flow rate. Typically, FET transistor 52 is reset
at each flow reversal of fluid into cylinder 12. Thus the output
voltage from operational amplifier 55 at this time represents the
flow rate and can be sampled and suitably operated on by the other
components of this invention, in the manner described above. |