This invention relates to a device and a method by which to directly
measure based on pressure and/or temperature data and to control
the flow rate, viscosity, density, velocity, pressure or temperature
of a variety of fluids in motion with improved reliability and less
susceptibility to disturbance of the measuring and control system
while, at the same time, offering maximum flexibility for influencing
measurement and control data. Components of the flow meter system
are described and the methods for calibrating the system and determining
pour volumes are explained.
1. A device for continuously measuring multiple properties from
a variety of fluids in motion, comprising: a fluid inlet; an optimally
dimensioned fluid path; at least three sensors; a data acquisition
and analysis means for accurately determining multiple properties
of a variety of fluids in motion; and a fluid outlet, wherein said
at least three sensors measure the pressure and optionally the temperature
of said variety of fluids in motion; wherein at least two of the
said at least three sensors are pressure sensors and wherein data
acquired from said at least three sensors is analyzed in order to
calculate properties relating to the variety of fluids in motion,
said properties selected from the group comprising; viscosity, density,
velocity, flow rate, pressure and temperature.
2. A fluid flow measuring device comprising a recessed fluid pathway
to optimally receive a fluid in motion for precise measurements
of properties selected from the group comprising; viscosity, velocity,
density, temperature and pressure.
3. A method for continuously measuring properties of a variety
of fluids in motion comprising the steps of: pumping a fluid in
motion into a flow block comprising an optimally dimensioned recessed
flow path; sensing a variety of parameters of said fluid in motion
using a series of sensors optimally positioned within said recessed
flow path; acquiring data directly from the sensors and analyzing
said data using a matrix; and using said analyzed data to report
properties relating to said fluid in motion, said properties selected
from the group comprising; viscosity, density, velocity, flow rate,
pressure and temperature.
 Benefit of priority under 35 U.S.C. 119(e) is claimed herein
to U.S. Provisional Application No.: 60/432754 filed Dec. 11
2002. The disclosure of the above referenced application is incorporated
by reference in its entirety herein.
BACKGROUND OF THE INVENTION
 The invention relates to control and measurement of the
flow rate of a fluid.
BRIEF SUMMARY OF THE INVENTION
 Analog flow rate measuring and controlling units are known
with which differential pressure measurement is effected by way
of an orifice or other restriction in a flow channel to determine
the rate of flow. Following that, the value obtained by this measurement
is compared with a desired value in a calculating unit. If the actual
value differs from the desired value specified the calculating unit
emits a correcting signal for application to a proportional valve
unit which then initiates a correcting process to cause the measured
value of the flow rate to coincide with the desired value.
 A problem with the known flow rate measuring and controlling
means is that they are relatively inflexible and cannot readily
be matched to changes occurring in fluid properties. Processing
of the measurement and control data is substantially predetermined
by the system components. For a flowing fluid, the initial fluid
properties change with time due to the effect of external influences
(e.g., heat) which may occur in the course of flow. Flow meter components
such as sensors, analog amplifiers, analog comparators, and the
like are influenced by such drifts in properties. As a result, calculated
flow rates will substantially differ from actual flow rates.
BRIEF DESCRIPTION OF THE FIGURES
 FIG. 1 is an exploded view of the flow meter of the current
 FIGS. 2 a-c are top view, cross-sectional side view and
bottom view, respectively, of the flow block with recess.
 FIG. 3 is a bottom view of the flow block with recess highlighting
 FIGS. 4 a-b are schematic views of flow block with sensor
ports highlighting sensor port to sensor port reading models.
DETAILED DESCRIPTION OF THE INVENTION
 It is, therefore, an object of the instant invention to
provide a device and a method by which to measure and control the
flow rate of a fluid with improved reliability and less susceptibility
to disturbance of the measuring and control system while, at the
same time, offering maximum flexibility for influencing measurement
and control data.
 This document describes a Multiple Indicator Flow Meter
and accompanying system that can be used in a characterization process.
Components of the flow meter system are described and the methods
for calibrating the system and determining pour volumes are explained.
The Multiple Indicator Flow Meter of the current invention is useful
with a variety of flow systems; however, by way of example only
and not limitation, the current invention is described herein as
embodied in two types of flow meter systems used with a soda fountain.
First, the entire system is described using a "Gage Pressure
Type" of flow meter. This means the flow meter uses only absolute
or gage pressure sensors. Alternatively, variations of the meter
are described that compose a second embodiment of the flow meter.
This is a "Differential Pressure Type" meaning that the
meter incorporates differential pressure sensors where both sides
of the sensor are exposed to pressures within the meter. Other improvements
are also noted.
 As described herein and depicted in FIG. 1 the flow meter
system 2 comprises the following components and methods: a plastic
flow body comprising a base member 4 and a flow block 6; four MEMS
pressure sensors 8; a thermistor 10; a sensor housing 12; a circuit
board making contact with the pressure sensors and thermistor (Contact
Board) 14; a circuit board to drive the pressure sensors and thermistor
(Analog Test Board); a piston pump; a scale; and data acquisition/process
control hardware and software for delivering a method of calibration
and a method of evaluating data to determine pour volume. While
the flow meter system of the current disclosure comprises the above
elements, it is to be noted that a variety of alternatively designed
flow meter systems can employ the inventive sensor device. Such
variations do not depart from the spirit of the current invention.
 The flow body comprises two flat pieces made of polycarbonate
plastic, ceramic or other suitable material and termed the base
member 4 and the flow block 6. The base member 4 has two optimally
spaced holes, one for a fluid inlet 16 and the other for a fluid
outlet 18. Into the inlet hole 16 is glued an inlet tube to allow
for a fluid collection, and, similarly, to the outlet hole 18 is
glued an outlet tube to allow for fluid drainage. The inlet tube
is the smaller of the two shown in FIG. 1. At the outlet hole 18
a small section of thin-walled plastic honeycomb 20 is inserted
to reduce the development of vortices in the exiting fluid. In an
alternative embodiment, the honeycomb 20 is removed and outlet hole
18 comprises a rounded exit transition to reduce vortices. Other
methods for reducing vortices are well known in the art.
 The inlet and outlet tubes are glued onto the same side
of the Base Member 4 as is shown in FIG. 1. The inlet and outlet
holes are aligned with a fluid path recessed into a mating piece.
The mating piece is referred to as the flow block 6 and detailed
in FIG. 2. Flow block 6 is connected to base member 4 on the side
opposite the inlet and outlet tubes.
 A recess 22 is milled into flow block 6 such that when
the base member 4 and the flow block 6 are assembled, the recess
22 creates a fluid path from the inlet hole 16 to the outlet hole
18 of the base member 4. The recess 22 is shaped to generally form
an initial semi-circular entry that gradually tapers out towards
a thin rectangular cross-section for the fluid flow that forms a
larger semi-circular end (see FIGS. 2b and 2c). In a preferred embodiment,
flow block is about 2.5 inches long and about 1.0 inch wide and
about 0.35 inch high, and thus, the rectangular cross-section forming
recess 22 is about 1.8 inches long, about 0.5 inches wide and about
0.025 inches high. Small changes in the height of the recess 22
can cause significant error in the data acquired. Because ceramic
has a low thermal coefficient of expansion and modulus of elasticity,
relative to most plastics, in the preferred embodiment, a ceramic
flow block and base member is used to reduce changes in the shape
of the meter when temperature or pressure changes.
 The flow block 6 also contains four (4) sensor port holes
24 that allow the fluid from the fluid path to contact four (4)
MEMS pressure sensors 8 which are discussed below.
 The flow block 6 guides the fluid from the inlet tube 16
through a right angle turn, expanding to a rectangular cross sectional
area that is smaller than the circular cross sectional area of the
inlet tube 16. In a preferred embodiment, the inlet tube has about
a 0.25 inch ID, and the outlet tube, when used in conjunction with
honeycomb 20 is about 0.5 inch narrowing to about 0.25 inch. In
the preferred embodiment, as used on a soda dispensing system, the
rectangular cross section has a width that ranges from about 0.25
inch to about 1 inch with about 0.5 inch being most preferable;
a height that ranges from about 0.02 inch to about 0.065 inch with
0.025 being preferable; and a length that ranges from about 2.2
inches to about 1.4 inched. In an alternative embodiment, height
of recess 22 is stepped. In this embodiment, the height decreases
from about from about 0.04 inch to about 0.03 inch across the length
of said recess 22. The fluid velocity increases at this area of
recess 22. A pressure sensor port 24 is located at the beginning
and at the end of this transition detecting any change in pressure
caused by increased velocity and by turbulence.
 A temperature sensor 26 is also located on the fluid surface
of the Flow Block 6 in this area. It projects slightly, but does
not significantly interfere with the flow. The fluid then continues
across several recessed grooves 28 which cause increased turbulence
in the fluid. The turbulence causes pressure of the fluid to drop.
Another pressure port 24 is located after the grooves 28. The fluid
then meets a pitot-tube type port 30. Traveling around this port
it then exits through the honeycomb flow conditioner 20 and outlet
hole 18. In an alternative embodiment, the pitot tube can be replaced
by a step that reduces the cross-sectional area of the flow path.
In one example of this alternative embodiment, said step is a reduction
in height of recess 22 across its length. For example, said height
can decrease from about 0.04 inch near the inlet hole to about 0.03
inch at the outlet hole. This alternative feature is not as restrictive
as the pilot tube and is less prone to clog from debris. Like the
pilot tube, it has an element of density dependence because the
fluid velocity must increase.
 The height and width of the rectangular flow path formed
by recess 22 can be changed in order to better match the unique
pressure drops associated with a variety of different liquids. The
flow meter 2 described in the preferred embodiment is designed to
run with soda water or sugar syrup with no modifications required.
However, the current invention is capable of use with any flowing
liquids, and thus finds use and applicability to a variety of devices.
 One benefit of the invention is that the geometry of the
flow body is adjusted such that the same flow meter geometry can
be used for a variety of fluids. Thus, when the flow meter system
of the current invention is used with a soda fountain a flow meter
having the same geometry is used for either soda water or for syrup.
Thus the flow meter of the current invention is highly versatile,
and can be applied to a variety of fluids with only slight to no
modifications to geometry.
 Pressure limits in industry are given for the soda water
and the syrup in typical soda fountain applications (e.g. soda water
at 4 oz/s should have a maximum pressure drop of about 40 PSI, while
cold syrup at 1 oz/s should have a maximum pressure drop of about
20 PSI). This invention optimizes the overall pressure drops for
either case to the allowable maximums, thus creating the largest
pressure signals possible for both cases. This is done by adjusting
the height to width ratio of the cross section of the flow body
channel. For example, a thinner cross section will restrict the
viscous sugar syrup more than the soda water because of said syrup
viscosity. At maximum flow rates, a thin cross-section in the meter
would create a much higher pressure with the syrup than with the
soda water. Conversely, a more square shaped cross-section would
produce a much lower pressure with the syrup than the soda water.
The soda water is thinner, so the pressure drop is smaller than
that of sugar syrups at any given rate. But, the soda water runs
at a rate 4-5 times faster than sugar syrup, raising the pressure
drop of the water relative to the syrup. The cross section of the
rectangular recess 22 can be adjusted in both width and height to
create the desired pressure drops (e.g., 40 psi for soda water and
20 psi for syrup) for a desired fluid. Those of ordinary skill in
the art will readily adjust the dimensions of the current invention's
recess 22 to facilitate it use with a variety of fluids. Such adjustments
are well within the spirit of the current invention.
 The advantage of this technique is that is maximizes the
pressure signals. If this technique is not used, the signal under
one condition (like with cold syrup) could be very high, while the
signal under the other condition (like with soda water) would be
very weak, or visa-versa.
 It has been determined that the largest pressure drop with
soda water is at the 90 degree bend. Increasing the height or width
of the flow path will decrease the total pressure drop of the soda
water because it opens the flow path, and decreases the fluid velocity
change. The largest pressure drop using syrup is over the rectangular
cross-section because of its higher viscosity. Increasing the height
or width of the flow path will also decrease the total pressure
drop of the syrup. However, the relationship between the height
of the flow path and the fluid is much stronger with syrupy fluids
than with watery fluids. Small changes in the height will affect
the pressure drop with syrup much more than with water. This difference
allows the pressure limits to be matched to their required values,
(e.g., soda water to 40 psi and syrup to 20 psi). The height is
the most critical factor in the matching. The length of this restriction
can also be adjusted to increase or decrease the pressure drops.
Thus the current invention, therefore, creates a single path that
produces optimal pressure drops for a variety of fluids, (e.g.,
40 PSI pressure drop with soda water at 4 oz/s, and a pressure drop
of 20 PSI using cold sugar syrup at 1 oz/s).
 The shape of the inlet hole 16 radius on the base member
4 at the first 90 degree bend in the fluid path can be small or
more rounded. A small radius can increase any pressure drop significantly
due to turbulent flow. With a smaller radius, the pressure drop
becomes more viscosity dependent and less density dependent. A large
radius on this corner will lower the pressure, decrease turbulence
and leave a smaller signal, but it will be more dependent on density.
 A highly density dependent pressure drop occurs at the transition
between the first two sensors at the 90 degree bend, and a highly
viscosity dependent pressure drop occurs across the flat cross-section
area. Having one signal that is highly density dependent and one
that is highly viscosity dependent is a key element of the flow
meter design, discussed below. Thus, the flow meter of the current
invention is at least three meters in one, wherein every detected
pressure drop is an indicator. The use of the pressure drop across
the 90 degree transition, for example, could be used on it's own
to determine the flow rate; however, such single metered measurements
are often inaccurate. By placing another one right in line (pressure
drop across the thin area), the second meter can be used to both
check other conditions or fluid properties and correct the data
collected by the first meter. Under some circumstances, those two
might show a flow rate, but they may also be inaccurate, so the
third indicator (another pressure sensor or temperature) can be
used to check them and correct them. Thus, the flow meter of the
current invention has at least three sensors for measuring three
independent variables: flow, viscosity, and density, (temperature
is a function of viscosity and density for sugar syrups, so it is
not independent), as well as to correct data collected in the first
 Micro-Electromechanical Sensors (MEMS) pressure sensor elements
are preferably used for sensors 8. These elements consist of a silicon-based
MEMS pressure sensor die with a partially conductive gasket covering
the electrical contacts on one side, and another gasket on the opposite
side. The gasket and die assemblies are placed over each hole in
the Flow Block 6. Those of ordinary skill in the art are readily
familiar with MEMS and sensor technologies.
 The sensors are mounted on the opposite side of the Flow
Block 6. The port holes 24 are positioned at strategic points along
the fluid path in recess 22. Between these holes, restrictions on
the fluid flow cause pressure drops as discussed previously. The
pressure at each port hole 24 is detected by its respective pressure
 A sensor housing 12 was placed around the pressure sensor
8 assemblies. The geometry of this sensor housing 12 closely duplicates
the original MEMS pressure sensor housing. The sensor housing 12
holds the pressure sensor 8 assemblies in the right location over
the ports 24. The conductive pressure sensor gaskets are left exposed
to contact the contact board 14. The gaskets on the opposite side
seal against the flow block 6.
 A thermistor 10 is mounted in the flow block 6 so that it
is in contact with fluid in the inlet area. Small conductive contact
pins are soldered to the thermistor leads. These contact pins protruded
through the sensor housing 12 and are positioned to make electrical
contact with the contact board 14.
 Over the sensor housing 12 is placed a circuit board called
the contact board 14. The contact board 14 makes direct contact
with the conductive pressure sensor gaskets and the thermistor 10
pins. The contact board 14 is routed to connect the contact points
for the pressure sensors 8 and the thermistor 10 to cable leads
positioned on the opposite side of said contact board 14. A stiff
support is mounted over the contact board 14 to increase the rigidity
of the assembly.
 The base member 4 flow block 6 pressure sensors 8 thermistor
10 Sensor Housing 12 and Contact Board 14 along with assembly
hardware constitutes the preferred embodiment for the Flow Meter
Assembly. Said flow meter assembly is shown in FIGS. 1 and 2. One
of ordinary skill in the art will readily apply the teachings of
the current invention to a variety of flow measure systems. Such
applications and variations are well within the spirit of the current
 Cables connect the contact board 14 with an analog test
board. The analog test board consists of electronic circuits to
drive the four pressure sensors 8 and thermistor 10 and provide
corresponding voltage output signals. It also contains circuits
that generate analog differences between the pressure sensor voltage
signals. It is preferred that signal errors are eliminated using
the pressure differences obtained from an analog circuit, rather
than mathematically subtracting them in the digital domain because
the signals can change in the time that it takes to sample two pressure
sensors. Sampling one analog signal gives an instantaneous reading
of the pressure difference.
 In a preferred embodiment, a piston pump is connected to
the inlet hole 16 of the flow meter using stiff tubing. The pump
is driven by a stepper-motor, and is capable of delivering fluid
at precise rates. The piston pump is used for calibration and preliminary
volume tests. Other types of pumps can be used, and such use of
other pumps is well within the spirit of the current invention.
 Preferably, a scale having 0.1 gram resolution is placed
at the outlet of the flow meter assembly. A container is placed
on the scale and fluid from the flow meter falls freely into said
container. The scale should have a serial interface allowing for
data acquisition equipment to tare the scale (set the scale reading
to zero) and read the scale value after each pour or calibration
 Data acquisition and process control software and hardware
is installed and programmed on a PC. This equipment is used to drive
the piston pump, tare and read the scale, read the voltage values
from the analog test board and provide time stamps for each sample
of the data. Before each run, the scale is tarred. Next, the piston
pump is activated. Voltage samples are then taken across all channels
of the analog test board every 10 ms through the duration of the
run. After the run, the scale is read, and a text file is generated
 The file contains a log of the programmed settings, voltage
readings with corresponding time stamps, and the final scale reading.
 The system is calibrated to generate a flow rate value based
directly on the readings from the pressure sensors 8 and the thermistor
10. This is unlike the flow meters of the prior art wherein the
data acquired from pressure sensors is combined with a variety of
other data, such as inlet duct size diameter, renolds equations
and the like, for making calculations of flow rate. For the current
invention, a matrix is developed allowing the flow rate of a test
sample to be calculated directly from the pressure voltage. A given
fluid is pumped through the flow body 2 at a constant flow rate
using the piston pump. Pressure readings are sampled along with
temperature. The flow rate is then changed and the corresponding
pressures and temperature are again sampled. Changes in temperature
of the fluid creates pressure changes at a given flow rate, so several
temperatures are sampled at several flow rates for each fluid. The
process is repeated with different fluids, thereby generating a
matrix containing flow rate, pressure and temperature data for the
fluids. The pressures that are used in the matrix are the differential
pressures (differences between separate pressure ports 24 generated
on the analog test board). Two or more of these different values
are used to generate the flow rate algorithm.
 Data from the matrix of flow rates, pressures and temperatures
is then conditioned (outliers and repeated values were thrown out)
and processed using software. This generates a mathematical formula
for flow rate, with flow rate as a function of the sensor outputs.
Any combination of relationships can be used, as long as there are
at least three independent values:
Q=f(P1 P2 T) Eq. 1.
Q=f(P1 P2 P3) Eq. 2.
Q=f(P1 P2 P3 T) Eq. 3.
 Where Q is the flow rate, P1 is the first differential pressure
sensor, P2 is the second differential pressure sensor, P3 is the
third differential pressure sensor, and T is temperature of the
fluid. For syrups where viscosity, density and temperature are dependent
upon each other (one cannot change without one of the others changing),
temperature can be used as one of three variables as in Equation
1. This is the case with sugar containing syrups. For other fluids,
more information is required as in Equations 2 and 3.
 The same process is used to generate density or viscosity
information along like with flow rate, provided that this information
is available while calibrating as with a known flow rate. The density
of the fluid can be observed by dividing the mass of a pour (measured
by the scale) by the volume of the pour (derived from the distance
the piston pump traveled). Density is then substituted into the
.rho.=f(P1 P2 T) Eq. 4.
.rho.=f(P1 P2 P3) Eq. 5.
.rho.=f(P1 P2 P3 T) Eq. 6.
 Where .rho. is the density of the fluid.
 For viscosity, using the current set-up, only a qualitative
value of the viscosity can be generated by using a function of the
viscosity and density dependent pressure drops. Of course, if the
meter is intended to be a viscosity meter, and a viscometer is used
during the calibration process, the viscosity readings can be associated
with the sensor outputs just like flow rate, and a relationship
can be found in a similar manner, using equations:
.upsilon.=f(P1 P2 T) Eq. 7.
.upsilon.=f(P1 P2 P3) Eq. 8.
.upsilon.=f(P1 P2 P3 T) Eq. 9.
 Where .upsilon. is the viscosity of the fluid.
 Two basic methods of determining the flow rate can be used
to employ an empirical data set. The first is an interpolation algorithm;
the other is a direct mathematical formula. Interpolation can be
very accurate, regardless of the shape of the functions, but may
require data that is well outside the intended rage of operation.
Mathematical formulas can also be used. The shapes of the functions
are very important here. A wide variety of formula forms can be
used; however, in a preferred method a 2nd or 3rd order multivariate
polynomial is used. It is important to precondition the data (taking
the square root of some of the pressure values) before processing
the data into formulas.
 The shape of the pressure to flow rate relationship has
much to do with the Renolds number and with whether the flow is
laminar or turbulent or is going through transition. It is preferred
that the flow is either in a laminar state or a turbulent state
because of inconsistencies in the transitional areas.
 If a mathematical formula is used, it is unlikely that the
data from of soda water, diet syrup and sugar syrups under all temperatures
and flow rates will fit well using a single formula. Basic information
from the indicators can be used to classify the fluid. Separate
formulas can be generated during the calibration process, and selected
based on the fluid class. For example, when sugar syrups are in
the system, the ratio between the Port A-Port B pressure drop and
the Port B-Port C pressure drop is very different than when soda
water or diet syrups are flowing. (Port A, B, C and D refer to sensor
ports 24 and are generally aligned according to FIG. 3. The ports
are aligned by way of example only, and an alternative aligning
of the ports does not depart from the spirit of the current invention.)
To tell the difference between soda water and diet syrup, magnitude
of the Port A-Port B pressure drop can be used. It will be higher
with soda water because of the higher flow rate.
 Data is collected for individual pour tests using the pumps,
fluids, discrete flow rates and varying temperatures specified.
The data associated with each pour is stored as a text file in a
format similar to the file format used in the calibration procedure.
The text file is then evaluated using an algorithm; however, the
evaluation could also be accomplished on a dedicated microprocessor.
The algorithm calls on the pressures, temperatures, and the mathematical
flow rate formula to evaluate each set of sampled data. A flow rate
is then generated for the sample and the time for each sample is
recorded as well. The volume is determined by multiplying the flow
rate by the duration of the sampling time (-10 ms). These small
volumes are calculated for each sample, and then added to a cumulative
sum. At the end of the file, the sum represents the total volume
for the pour.
 To improve accuracy, a different mathematical formula is
generated for syrup classes, (diet syrup, sugar syrup, soda water)
in the calibration process. Since the pressure readings for these
three fluid types are in distinct ranges, widely separated from
each other, the algorithm looks at key indicators such as the ratio
between the pressure differentials and the magnitude of the pressure
differentials in order to determine which of the mathematical formulas
to use. Once chosen, the formula is then applied to the entire run
 In an alternative embodiment of Applicant's flow meter system,
the pressure sensors 8 are differential pressure sensors. Differential
pressure sensors, where fluid is applied to both sides of a strain
gage-type membrane, are much more resistant to damage by a pressure
spike because the force of the pressure is applied on both sides
of the meter. By using differential pressure sensors 8 the reading
between two ports that generate a pressure drop of 15 PSI, for example,
can be measured using a 15 PSI pressure sensor. If using gage pressure
sensors, all of the sensors should have a significantly higher rating
in order to withstand pressure spikes in the line. Pressure spikes
can be as much as several hundred PSI, which will damage most gage
sensors. Large pressure spikes can be further reduced with the uses
of a plenum. A plenum is a small camber filled with air or other
compressible material that is in contact with the inlet fluid. As
pressure spikes or vibrations travel toward the meter, they are
damped by the plenum. The plenum can absorb and release the fluid's
kinetic energy, thereby smoothing the fluid flow and preventing
damaging pressure spikes.
 Using the lowest possible rating of the differential pressure
sensors can maximize their sensitivity. This can be accomplished
by installing pressure sensors that are rated at the maximum pressure
drop that the pressure sensor will experience.
 With differential pressure sensors, the fluid must be routed
to the back of the pressure sensor. This routing can be done in
two basic ways. As shown in FIG. 4a, the differential pressure sensor
can be routed from Port A to Port B, then another from Port B to
Port C, and the last from Port C to Port D. In this alternative
configuration the pressure sensors 8 are each measuring the small
pressure drops from port to port. The sensitivity is maximized;
however, error is increased when combining the sensor readings for
 A further alternative method is to route the differential
pressure sensors from Port A to Port B, then another from Port A
to Port C, and the last from Port A to Port D. See FIG. 4b. In this
way the pressure sensors are each measuring the cumulative drop
from the first port. The last sensor measures the entire pressure
drop across the meter. This can be an advantage when the accuracy
of the meter relies upon mathematical calculations that require
the value for the entire pressure drop. Any time the values of the
pressure sensors need to be manipulated in the digital domain, errors
 By resetting the value of the pressure drop to zero when
there is no flow will help account for drift that is sometimes associated
with the less expensive sensors, thereby improving accuracy. Initially,
the voltage of the sensor is set to zero, but because of the more
complex nature of the algorithm, the pressure should be determined,
then it should be offset by a pressure value adjustment for the
 It is possible to detect the temperature of the fluid based
on other electronic values read from the pressure sensors. If these
values are sampled it may allow the elimination of the temperature
sensor. If this is the case, the pressure sensors should be placed
in close proximity to the flow path. That way they will be more
sensitive to the change in temperature of the fluid. This method
is not recommended if the temperature reading is to be used for
flow rate calculations because the response time of the pressure
sensor will most likely be too slow.
 A preferred method is to use properly temperature compensated
pressure sensors and remove the sensors from the flow path to further
reduce the possibility of errors due to rapid fluid temperature
changes. If the algorithm resets the pressure value to zero before
each run, the temperature effects will have a minimal impact on
accuracy so long as the pressure sensors are buffered from dramatic
changed in temperature during the run.
 A bubble in the routing tube can cause large errors with
rapidly transitioning flow rates. This is because the fluid will
have to move through the narrow routing path. The movement takes
time and will cause a lag in the time as the pressures in the flow
path and at the sensor equalize. With no bubble, fluid transmits
the pressure directly to the sensor with negligible movement.
 There are at least three ways to reduce errors caused by
bubbles in the routing. A first method is to locate the ports below
the flow path. A second method is to make the port routing relatively
large after the initial port hole. This way if a bubble does get
into the routing, its effects will be minimized. Third, a membrane
can be located between the flow path and the ports. It must be flexible
in order to effectively transmit the pressure. Those of ordinary
skill in the art will readily exercise a variety of methods for
eliminating or minimizing the errors caused by bubbles in the routing
line. Alternatively, elongating the pressure ports across the flow
path can reduce errors caused by localized swirling of the fluid.
This also allows fluid to travel in and out of the port more freely,
reducing errors cause by bubbles.
 The absolute line pressure can also affect the reading of
a differential pressure sensor. The direction and rate of change
of the reading can cause error as well (Hysteresis). Temperature
may also affect the reading of the pressure sensors. If the pressure
sensors need additional temperature compensation, this can come
from a value read from the sensor that is based on the temperature
of the pressure sensor. The corrected pressure can therefore be
represented by the following function:
P.sub.C=f(V.sub.P, V.sub.P/t, P.sub.LINE+P.sub.u/2 T)
 P.sub.C=the corrected differential pressure.
 V.sub.P=the voltage reading from the differential pressure
 V.sub.P/t=the change in the voltage reading from the differential
pressure sensor divided by a corresponding change in time, in other
words the rate of change of the differential pressure.
 P.sub.LINE=the line pressure.
 P.sub.U=the differential pressure based on a constant multiplied
 T=a reading that corresponds to the temperature of the pressure
sensor. If there is already sufficient thermal correction for the
sensor, this variable can be eliminated from the formula.
 Using these corrections, the accuracy of inexpensive sensors
can be significantly improved.
 For low flow fluids (e.g., diet syrup), the signal is smallest.
Redundant formulas based on more than one sensor can be used to
calculate the flow rate. This will work if the error associated
with the sensors is random and if both sensors used generate about
the same amount of error.
 During calibration, ideally a fluid is pumped through the
meter at a programmed, known rate. If the actual flow rate deviates
from the programmed flow rate (due to vibration, tube compliance
pump ramping plenum dampening, etc.) error will be introduced to
the system. This error can be greatly reduced by ensuring steady
state conditions or by generating a temporary flow rate estimate.
A temporary flow rate estimate can be generated by observing a single
pressure drop signal through the run and making a simple calibration
for that run base on it. Since the conditions (temperature and fluid)
are relatively stable for a single run, a simple estimate of the
flow rate can be generated based on a single pressure sensor reading.
This temporary flow rate estimate accurately tracks small unintentional
changes in flow rate. These estimated changes are then be associated
with the small variations of the readings at any point in time in
the run. The data set that is gathered will have more accurate flow
rate information, thus improving the accuracy of the final calibration,
when all of the data is used to crate a general formula or is used
in an interpolation data set.
 A variety of flow meter systems can be created incorporating
the any or all of the improvements mentioned above. Those of skill
in that art will readily make flow meter sensors incorporating the
current invention and any or all of the alternative configurations
described above or known in the art. Such flow meters are well within
the spirit of the current invention.