`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`_______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_____________
`
`MICRO MOTION, INC.
`Petitioner
`v.
`
`INVENSYS SYSTEMS, INC.
`Patent Owner
`
`Patent No. 7,571,062
`Issue Date: August 4, 2009
`Title: DIGITAL FLOWMETER
`_______________
`
`Case No. IPR2014-01409
`____________________________________________________________
`
`DECLARATION OF DR. MICHAEL D. SIDMAN REGARDING THE
`INVALIDITY OF CLAIMS 1, 12, 23, 24, 25, 29, 36 AND 43 OF U.S. PATENT
`NO. 7,571,062
`
`4852-9264-2334.1
`
`1
`
`Micro Motion 1064
`
`
`
`1.
`
`I, Dr. Michael D. Sidman, resident at 6120 Wilson Road Colorado
`
`Springs, CO, hereby declare as follows:
`
`2.
`
`I have been retained by Foley & Lardner LLP to provide my opinion
`
`concerning the validity of U.S. Pat. No. 7,571,062. I am being compensated for my
`
`time at the rate of $450/hour in preparing this declaration.
`
`I.
`
`QUALIFICATIONS
`
`3.
`
`I completed my undergraduate studies at Northeastern University,
`
`where I received a Bachelor’s and a Master’s degree in Electrical Engineering
`
`concurrently in 1975.
`
`4.
`
`I received my Ph.D. from Stanford University in 1986. My work at
`
`Stanford as a Digital Equipment Corporation Fellow and University Resident
`
`included developing a high-performance digital control system for a lightly-
`
`damped mechanism in the Stanford Aero/Astro Robotics Laboratory.
`
`5.
`
`My dissertation was entitled: Adaptive Control of a Flexible Structure.
`
`This research culminated in an adaptive control system that actively damps the
`
`vibrations of a lightly-damped mechanism, like a large space structure or disk drive
`
`actuator, whose resonant frequencies may be unpredictable or variable. The system
`
`performed on-line system identification of the frequencies of the mechanism’s
`
`mechanical resonances.
`
`4852-9264-2334.1
`
`2
`
`
`
`6.
`
`I have worked for over 35 years in the field of motor, motion and
`
`servo control systems, and specifically in the field of digital control and signal
`
`processing systems. I have researched the control and m’/echanical dynamics of
`
`high performance, precision digital servo systems such as found in a range of
`
`computer peripheral devices.
`
`7.
`
`Since 1992, I have been working as an independent engineering
`
`consultant. I am currently President of Sidman Engineering, Inc. I provide
`
`engineering design services to manufacturers worldwide, which span a range of
`
`industries. This work has included the following: (1) optimizing and simulating
`
`mechatronic systems; (2) developing comprehensive custom design and dynamic
`
`system simulation tools including computer models of motor, motion and servo
`
`control systems; (3) teaching on-site technical short courses to design engineers
`
`and scientists; and (4) consulting on high-performance digital servo systems design
`
`and problem resolution.
`
`8.
`
`The field of “mechatronics” encompasses mechanics, electronics and
`
`control systems technologies.
`
`9.
`
`A “servo” or “servomechanical” system is a control system that
`
`controls position, velocity or acceleration, often utilizing motion sensor feedback.
`
`10.
`
`Through Sidman Engineering, I provide interdisciplinary analysis and
`
`resolution of complex design issues. This may include providing clients with
`
`4852-9264-2334.1
`
`3
`
`
`
`customized, comprehensive computer based design tools and simulation models of
`
`a variety of dynamic systems, including electromechanical products and systems.
`
`These comprehensive models integrate actuator dynamics and electrodynamics,
`
`mechanical resonances, electronic circuitry, sensors, signal processing and
`
`filtering. In this role I have developed comprehensive servo system simulation
`
`models and design tools. The design tools I provide generally are used by product
`
`or system design engineers to understand system behavior and to optimize system
`
`parameters. As discussed below, I also provide on-site high level technical training
`
`courses for design engineers and scientists at companies. My business address is at
`
`6120 Wilson Road, Colorado Springs, Colorado, 80919.
`
`11. My commercial clients span the following industries and applications:
`
` Industrial and commercial: chemical process control, steel
`
`manufacturing, hydraulic control, commercial aviation, medical
`
`equipment, textile manufacturing, food processing, bicycle motor
`
`control, fuel cells.
`
` Computer peripherals and related test equipment: hard disk drives,
`
`optical disc drives, tape drives, printers, digital pens, robotics.
`
` Automotive: tire manufacturing & test, engine and vehicle
`
`dynamometers, electromechanical EGR valves, electric power assisted
`
`steering.
`
`4852-9264-2334.1
`
`4
`
`
`
` Chip design: motor, motion and digital servo control IC’s, DSPs and
`
`microcontrollers.
`
` Defense: aerospace, naval, optical reconnaissance, security scanning.
`
` Instrumentation: software, flow meters, optical position sensing,
`
`coordinate measurement machines.
`
` Telecommunications: digital signal processing, speech analysis,
`
`optical switching.
`
`This list is simply representative of my technical consulting activities to
`
`companies over a period of more than two decades.
`
`12. Before I became an independent engineering consultant, I spent 17
`
`years at Digital Equipment Corporation (DEC) in roles spanning product
`
`development, advanced development and research. I headed DEC’s Advanced
`
`Servo Development Group and Servo-Mechanical Advanced Development Group,
`
`both of which I founded. These groups developed and demonstrated technology
`
`involving, for example, position and velocity sensing, MEMS accelerometers,
`
`active vibration control, optimal seek control, piezoelectric head positioning
`
`actuators and DSP-based digital servo systems for hard disk drives. In a prior
`
`product design development role, I was the Project Engineer for DEC’s RK07 disk
`
`drive product.
`
`4852-9264-2334.1
`
`5
`
`
`
`13.
`
`I served as DEC’s representative to the Berkeley Sensor and Actuator
`
`Center (BSAC), which conducts industry-relevant interdisciplinary research on
`
`micro- and nano-scale sensors and moving mechanical elements and actuators
`
`constructed using integrated-circuit technology. I also served as DEC’s
`
`representative for servo and mechanical technology to the National Storage
`
`Industry Consortium (NSIC).
`
`14.
`
`I also sponsored applied research and/or researchers at Stanford
`
`University, U.C. Berkeley and the University of Colorado at Colorado Springs.
`
`15.
`
`I have taught numerous courses and seminars relating to the field of
`
`mechatronics to product and system design engineers and a graduate level course
`
`in Optimal Control at the University of Colorado in Colorado Springs.
`
`16.
`
`Through Sidman Engineering, I have provided my on-site, customized
`
`Digital Servo System Short Courses and my MATLAB/SIMULINK/Toolbox
`
`Laboratory Training Courses to product development and research engineers and
`
`scientists worldwide in a wide range of industries and government entities since
`
`1993. These courses may optionally include portions devoted to control systems
`
`and/or signal processing analysis and simulation. I developed these courses to
`
`enable attendees, who usually represent a range of technical disciplines, to model
`
`and simulate dynamic systems and the operation of products they are developing.
`
`4852-9264-2334.1
`
`6
`
`
`
`17.
`
`I became a Third Party Provider for The MathWorks, Inc. in 1993 and
`
`authored an invited feature article, entitled “Control Design Made Faster and More
`
`Effective,” for MATLAB News and Notes, Summer/Fall 1994.
`
`18.
`
`I am a member of professional organizations dedicated to control
`
`systems and mechatronic technology. I am a Senior Member of the Institute of
`
`Electrical and Electronics Engineers (IEEE) where I am a member of the Control
`
`Systems Society.
`
`19.
`
`I am also a member of the American Society of Mechanical Engineers
`
`(ASME), where I was Chairman of the Pikes Peak Section and member of the
`
`Dynamic Systems and Control Division (DSCD).
`
`20.
`
`I am the named inventor of eighteen U.S. patents relating to
`
`technologies including: analog and digital electronics, digital signal processing,
`
`high performance digital servo systems, adaptive runout control, active damping of
`
`mechanical vibrations, adaptive control system gain regulation, etc.
`
`21.
`
`I have published numerous articles relating to control systems, and
`
`specifically, articles relating to the motion sensing and control of precision
`
`actuators.
`
`22.
`
`In March, 1998 I taught a customized three day
`
`MATLAB/SIMULINK Laboratory Training Course directed specifically to
`
`Coriolis flowmeter resonant dynamics as well as the use of digital techniques in
`
`4852-9264-2334.1
`
`7
`
`
`
`Coriolis flowmeter design to ten Coriolis flowmeter engineers and designers, all of
`
`whom were employed by Micro Motion. Because this was a small group class, it
`
`was quite interactive, so I was able to get a good understanding of the level of skill
`
`of these Coriolis flowmeter engineers and designers in the use of digital signal
`
`processing techniques. I understand the inventors of the ’062 patent claim priority
`
`to a provisional patent application filed on November 26, 1997. Because I taught
`
`this three day course less than four months after the filing date of the provisional
`
`application to which the ’062 patent claim’s priority, this allowed me to get a good
`
`understanding of the level of skill of these Coriolis flowmeter designers and
`
`engineers at almost exactly the time that the provisional patent application was
`
`filed. Attached as Exhibit 1065 are some materials relating to this course,
`
`including an attendance list. As the list reflects, one of the ten attendees at this
`
`course was Mr. Michael Zolock, the inventor of one of the prior art patents I
`
`discuss later in this declaration.
`
`II.
`
`INTRODUCTION TO CORIOLIS FLOWMETERS
`
`A. Uses of Coriolis Flowmeters
`
`23. Gaspard-Gustave de Coriolis was a French mathematician,
`
`mechanical engineer and scientist who lived from 1792 to 1843. Coriolis studied
`
`forces as observed from a rotating frame of reference.
`
`4852-9264-2334.1
`
`8
`
`
`
`24. A Coriolis flowmeter is a measurement device used to measure mass
`
`flow rate and/or density of a material flowing through an oscillating conduit (i.e.,
`
`tube) in the flowmeter. (See, e.g., U.S. Pat. No. 4,872,351, Ex. 1020, 1:49-2:4.)
`
`25. Coriolis flowmeters include one or two conduits through which
`
`material flows. One conduit is described for simplicity, but the discussions herein
`
`are generally applicable to a system with two conduits. (See, e.g., U.S. Pat. No.
`
`4,823,614, Ex. 1021, 1:44-47, 2:51-55.)
`
`26. Coriolis flowmeters measure mass flow rate by sensing the effect of
`
`Coriolis forces on the material flowing through the vibrating conduit.
`
`27.
`
`The material flowing in the tube may be a single type of material
`
`(single phase) or multiple types of material mixed together (multi-phase).
`
`Examples of materials include a wide variety from gas to liquid to near-solid. (E.g.,
`
`U.S. Pat. No. 4,679,947, Ex. 1007, 1:41-43 (“two-phase flow”); U.S. Pat. No.
`
`4,823,614, Ex. 1021, 3:14-17 (“highly viscous fluids and thick slurries, for
`
`example, asphalt, latex paint and peanut butter”); U.S. Pat. No. 4,872,351, Ex.
`
`1020, 1:29-44 (sweetening agent, oil); U.S. Pat. No. 5,068,116, Ex. 1028, 1:18-20
`
`(water, syrup for beverage mixing); U.S. Pat. No. 5,143,257, Ex. 1022, 1:7
`
`(“medication and/or nutrients”); U.S. Pat. No. 5,148,945, Ex. 1023, 1:12-14
`
`(“ultra-high purity chemicals . . . such as in the manufacture of semiconductor
`
`wafers”).) Examples of multi-phase materials include a gas/liquid mixture, a
`
`4852-9264-2334.1
`
`9
`
`
`
`liquid/solid mixture, and a gas/liquid/solid mixture. (E.g., U.S. Pat. No. 5,224,372,
`
`Ex. 1024, 1:34-38 (“multiphase fluid emanating from oil and gas wells wherein
`
`essentially, mixtures of water, hydrocarbon liquids, such as crude oil; and gas are
`
`continually produced in varying proportions of the total fluid flowstream”); U.S.
`
`Pat. No. 5,317,928, Ex. 1025, 1:17-20; 1:25-30 (“Typically a two-component fluid
`
`mixture consists of either a solid component fully or partially dissolved within a
`
`liquid carrier fluid, or a liquid component mixed with a liquid carrier fluid” such as
`
`“sugar in water” in the “beverage industry,” “concentration of TiO2” in the “pulp
`
`and paper industry”).)
`
`28. As described below, Coriolis flowmeters rely on the principle that the
`
`mass flow rate of material passing through a sinusoidally oscillating conduit can be
`
`determined by sensing conduit twist induced by the Coriolis force. And, the natural
`
`frequency of conduit oscillation provides the basis for measurement of the density
`
`of the material inside the conduit. The Coriolis flowmeter’s electronics are
`
`responsible for initiating, sustaining and, in general, controlling the sinusoidal
`
`oscillation of the conduit(s). For example, through the use of the conduit motion
`
`sensors and an electromagnetic conduit driver or actuator, Coriolis flowmeter
`
`electronics provide the mechanical energy to vibrate the conduit and to regulate the
`
`amplitude of conduit oscillation. Coriolis flowmeters rely on the persistence of
`
`sustained sinusoidal oscillation of the conduits.
`
`4852-9264-2334.1
`
`10
`
`
`
`29.
`
`The flowmeter induces a Coriolis force on flowing material in the
`
`conduit by oscillating the conduit, and determines a property of the material based
`
`on information about the effect of the Coriolis force. (See, e.g., U.S. Pat. No.
`
`4,733,569, Ex. 1026, 1:27-36; U.S. Pat. No. 4,823,614, Ex. 1021, 1:47-61.).) For
`
`example, by measuring a phase difference in the sinusoidal oscillation of the
`
`conduit between two points on the tube, it is possible to determine the mass flow
`
`rate of the fluid flowing through the tube. Coriolis flowmeters were first
`
`commercialized by petitioner Micro Motion in the late 1970s and early 1980s. (See
`
`U.S. Pat. No. 5,373,745, Ex. 1003, 1:24-25 (“[Coriolis flowmeters were] first made
`
`commercially successful by Micro Motion, Inc. of Boulder, Colorado.”).)
`
`B. Components of Coriolis Flowmeters
`
`30. Coriolis flowmeters include the following basic components: a
`
`vibratable conduit (which can have various shapes and sizes) through which fluid
`
`flows; an electromechanical drive mechanism (including one or more
`
`electromagnetic drivers or actuators) for vibrating the conduit; one or more sensors
`
`that transduce the vibration of the tube; and electronics for controlling the drive
`
`mechanism and for analyzing signals from the sensors.
`
`31. Coriolis (and other) flowmeters were originally implemented with
`
`analog electronic components. (E.g., U.S. Pat. No. 2,865,201, Ex. 1004.) To do the
`
`necessary signal processing and control, such an analog flowmeter uses analog
`
`4852-9264-2334.1
`
`11
`
`
`
`components to process signals from the sensors and to control the drive
`
`mechanism. As digital electronic components became more readily available,
`
`flowmeters also incorporated digital components. (See, e.g., U.S. Pat. No. Re.
`
`31,450, Ex. 1005, which discloses a predominantly analog system incorporating
`
`some digital components.) Digital components include digital logic and
`
`programmable digital devices (e.g., microprocessors). (E.g., U.S. Pat. No.
`
`4,934,196 (“Romano”), Ex. 1006, Fig. 3; U.S. Patent No. 5,009,109 (“Kalotay”),
`
`Ex. 1008, Fig. 4; U.S. Pat. No. 5,050,439 (“Thompson”), Ex. 1027, 16:11-15.) A
`
`digital flowmeter may include analog and digital components. For example, a
`
`digital flowmeter may process signals from the sensors using digital components
`
`but control the drive signal using analog components. A digital flowmeter may
`
`alternatively control the drive signal using digital components.
`
`32.
`
`The flowmeter must process the sensor signals to extract information
`
`of interest from other information in the signals. Thus, all flowmeters, whether
`
`analog or digital, perform signal processing on the sensor signals. For example, in
`
`a Coriolis flowmeter, fluid flowing through an oscillating flowtube may cause a
`
`phase shift in the flowtube oscillation due to the Coriolis effect, and the flowmeter
`
`processes the sensor signals to extract the information related to the Coriolis effect
`
`from other information in the signals to determine mass flow rate or density. If the
`
`4852-9264-2334.1
`
`12
`
`
`
`signal processing is performed in digital components, then the signal processing is
`
`digital signal processing.
`
`C. Operating Principles
`
`33.
`
`The Coriolis effect caused by oscillating a conduit in which fluid is
`
`flowing results in a twist in the conduit, with a resulting phase shift in the
`
`oscillation between two points physically spaced apart on the conduit. The phase
`
`shift is proportional to the mass flow rate of the material.
`
`34. A tutorial showing the principles underlying the operation of Coriolis
`
`flowmeters may be found at
`
`www3.emersonprocess.com/micromotion/tutor/28_flowoprinccurvtubevib.html.
`
`35.
`
`The Coriolis effect as related to a Coriolis flowmeter was described in
`
`a 1990 article by the petitioner, where the term “flow tube” is synonymous with the
`
`term “conduit”:
`
`The Micro Motion flowmeter measures fluid mass in motion. A
`flowmeter is comprised of a sensor and a signal processing
`transmitter. Each sensor consists of one or two flow tubes enclosed in
`a sensor housing. The principle of operation is the same for all Micro
`Motion sensors.
`The sensor operates by application of Newton's Second Law of
`Motion: Force = mass x acceleration (F = ma). The flowmeter uses
`this law to determine the precise amount of mass flowing through the
`sensor tubes.
`
`4852-9264-2334.1
`
`13
`
`
`
`Inside the sensor housing, the flow tube is vibrated at its natural
`frequency (Figure 1) by an electromagnetic drive coil located at the
`center of the bend in the tube. The vibration is similar to that of a
`tuning fork, covering less than a tenth of an inch and completing a full
`cycle about 80 times each second.
`As the fluid flows into the sensor tube, it is forced to take on the
`vertical momentum of the vibrating tube. When the tube is moving
`upward during half of its vibration cycle (Figure 2), the fluid flowing
`into the sensor resists being forced upward by pushing down on the
`tube. Having the tube's upward momentum as it travels around the
`tube bend, the fluid flowing out of the sensor resists having its vertical
`motion decreased by pushing up on the tube (Figure 2). This causes
`the flow tube to twist (Figure 3). When the tube is moving downward
`during the second half of its vibration cycle, it twists in the opposite
`direction. This tube twisting characteristic is called the Coriolis effect.
`Due to Newton's Second Law of Motion, the amount of sensor
`tube twist is directly proportional to the mass flow rate of the fluid
`flowing through the tube. Electromagnetic velocity detectors located
`on each side of the flow tube measure the velocity of the vibrating
`tube. The two velocity signals are sent to the transmitter where they
`are processed and converted to an output signal proportional to the
`mass flow rate. Sensor tube twist is proportional to mass flow and is
`determined by measuring the time difference exhibited by the velocity
`detector signals. During zero flow conditions, no tube twist occurs and
`both sides of the tube cross the midpoint simultaneously. With flow, a
`twist occurs along with a resultant time difference between midpoint
`
`4852-9264-2334.1
`
`14
`
`
`
`crossing. This time difference appears as a phase shift between the
`two velocity signals and indicates mass flow.
`
`(“How the Micro Motion Mass Flow and Density Sensor Works”, Micro
`
`Motion, Inc., 1990, Ex. 1009, p. 1.)
`
`36. Note that in the Micro Motion article, only one conduit (flow tube) is
`
`shown. In a one-conduit system, velocity measurements would be made with
`
`respect to the housing. In a two-conduit system, velocity measurements are made
`
`4852-9264-2334.1
`
`15
`
`
`
`as relative velocity between the conduits. In either case, the concepts described
`
`below apply.
`
`37. As seen from the Micro Motion article, measurement of mass flow
`
`rate is based on oscillation of the conduit. Similarly, measurement of density is
`
`based on frequency of oscillation. Thus, oscillation of the conduit is necessary
`
`while accurately performing measurement.
`
`38. Measuring the phase shift described in the Micro Motion article is no
`
`different than measuring phase shift of input signals in any other system, and thus
`
`may be addressed by conventional signal processing techniques. Additionally,
`
`maintaining oscillation at a desired amplitude and/or frequency is addressed by
`
`conventional control theory techniques.
`
`D. Control Systems for Coriolis Flowmeters
`
`39.
`
`The control system of the Coriolis flowmeter provides a drive signal
`
`to an electromagnetic mechanism (driver) external to the conduit to initiate
`
`oscillation of the conduit. The conduit at some point generally settles into
`
`sinusoidal oscillation at the resonant frequency (or a harmonic) of the conduit. The
`
`resonant frequency of the conduit depends in part on properties of the material in
`
`the conduit, and may change if the properties of the material change. (See, e.g.,
`
`U.S. Pat. No. 4,823,614, Ex. 1021, 2:9-50 3:14-17.) For example, as the density of
`
`the material changes (e.g., during aeration, or during transitions from empty to full
`
`4852-9264-2334.1
`
`16
`
`
`
`and full to empty), the resonant frequency of the flowtube will also change, and the
`
`amount of energy required to keep the flowtube oscillating will generally change.
`
`Additionally, because the flowtube is a fixed volume, changes in density will result
`
`in changes in mass of the material in the flowtube, and corresponding changes in
`
`mass flow rate. The faster the density changes (e.g., in a rapid empty-to-full
`
`transition), the more quickly the mass within the flowtube will change.
`
`40. As disclosed, for example, in the many prior art patents listed above, a
`
`flowmeter incudes a control system that performs at least the following functions.
`
`41.
`
`The control system provides signals to the driver to add energy to the
`
`oscillation of the conduit as necessary to maintain oscillation of the conduit at a
`
`desired frequency (typically at or near the resonant frequency of the conduit).
`
`42.
`
`The control system receives signals from one or more motion sensor
`
`pick-ups mounted externally on the conduit to sense conduit oscillation. These
`
`transducers are located apart from each other on the vibrating conduit in order to
`
`provide information about conduit torsion induced by the Coriolis forces during
`
`flow conditions.
`
`43.
`
`The control system processes the signals received from the pick-ups,
`
`drives the conduits in synchronism with the resonant frequency of oscillation, and
`
`determines a property of the material (e.g., mass flow rate and/or density) from the
`
`pick-up signals. Information extracted from the pick-up signals (including the
`
`4852-9264-2334.1
`
`17
`
`
`
`property of the material) may be stored, displayed, and/or exported to another
`
`system by the flowmeter.
`
`44. A block diagram illustration of the relationship between the
`
`electronics and conduit of the flowmeter is shown in the figure below by way of an
`
`example which uses two pick-ups. The flowmeter electronics includes a drive
`
`control system.
`
`45.
`
`The control system of a Coriolis flowmeter therefore performs control
`
`functions – processes system input signals and controls system output signals – as
`
`described in well-known control theory dating back far before the 1990’s (i.e., long
`
`before the filing of the application leading to the patent that is the subject of this
`
`Inter Partes Review.) Further, the processing of signals was described by well-
`
`known signal processing techniques also dating back far before the 1990’s. Some
`
`elements of control theory and signal processing techniques will be discussed
`
`below.
`
`4852-9264-2334.1
`
`18
`
`
`
`III.
`
`CONTROL SYSTEMS BACKGROUND
`
`A. Open Loop Control
`
`46. A control system may be used to control the motion of a device or
`
`mechanism. Open loop control refers to a control action without feedback.
`
`47.
`
`In a Coriolis flowmeter, open loop control may be used to initially
`
`impart energy into a conduit to initiate oscillation of the conduit. For example, a
`
`drive signal having the appropriate spectral content could be used to initially excite
`
`resonant vibration of the conduit.
`
`48. Because an open loop control system cannot observe actual conduit
`
`motion or the effect of a control action, it cannot, for example, act to regulate or
`
`maintain oscillation amplitude of a conduit. Without feedback, there is no basis for
`
`the control system to modify the drive signal or adjust the control action. Thus,
`
`oscillation may decrease, to a stall condition, or oscillation may increase, resulting
`
`in amplifier saturation or even causing damage to the conduit.
`
`B. Closed Loop Control
`
`49. Closed loop control refers to the use of feedback as the basis for
`
`control action. For example, a feedback sensor or transducer such as a velocity
`
`sensor allows a feedback control system to monitor the effects of its control actions
`
`and to adjust the control action.
`
`50. A servo system is a motion control system. Most servo system are
`
`closed-loop systems which track or follow a desired setpoint or trajectory
`
`4852-9264-2334.1
`
`19
`
`
`
`command, sometimes referred to as a reference command level or signal. The term
`
`‘servo’ comes from the Latin word servus – meaning servant or slave. A closed
`
`loop motion control system may regulate position, velocity or acceleration.
`
`51.
`
`The following illustrates a simple example of a closed-loop control
`
`system. Closed-loop control relies on feedback (e.g., a feedback loop) to control a
`
`“plant” or mechanism in such a way as to make the plant’s output closely track a
`
`reference input signal. The “plant” in the figure below is the mechanism whose
`
`output is to be controlled. In the case of the Coriolis flowmeter, the plant would
`
`include the conduit driver and the dynamics of the conduit. The pick-ups in a
`
`Coriolis flowmeter correspond to the output measuring device in the figure below.
`
`52.
`
`In this closed-loop control system, the plant’s output is fed back for
`
`comparison with the reference input, and the difference, or error, is then used as
`
`the basis for a control or drive signal applied to the plant. The difference (error) is
`
`sensed by the ‘comparator’ (i.e., difference amplifier in this context) and amplified
`
`before being processed by a controller such as a PI controller, described below.
`
`Typically, a power amplifier (not shown) boosts the drive signal produced by the
`
`4852-9264-2334.1
`
`20
`
`
`
`controller. The blocks illustrated in the above figure form a control loop that
`
`provides a feedback control path which enables self-correction. The control system
`
`produces a dynamically controlled drive signal by way of a controller or control
`
`law that acts to minimize error. In some control systems, the reference input may
`
`be zero. Generally, “a feedback system has the ability to correct for load
`
`disturbances and inaccuracies in the controller.” (Introduction to Continuous
`
`and Digital Control Systems, Saucedo & Schering, Macmillan, 1968, Ex. 1029, p.
`
`3 (emphasis added).)
`
`i.
`
`Negative Feedback
`
`53.
`
`If the phasing of the feedback acts in such a way as to reduce error,
`
`instead of to accentuate it, then the closed loop control system would be said to be
`
`operating in a stable manner and operating with negative feedback. Stability of
`
`closed loop systems is a key consideration in control systems design. Oscillatory
`
`behavior is an indication that the control system is unstable.
`
`ii.
`
`Multi-Loop Feedback
`
`54.
`
`It is common to add an outer control loop to stabilize or enhance the
`
`dynamic performance of an inner control loop (or vice versa). The following is an
`
`example of multi-loop control in a servo system.
`
`4852-9264-2334.1
`
`21
`
`
`
`(Electromechanical Control Systems and Devices, Canfield, Robert E. Kreiger
`
`Publishing Company, Original Edition 1965, Reprint 1977, Ex. 1030, p. 42.)
`
`55. As described below, an outer control loop can be used to regulate the
`
`amplitude of oscillation of an inner control loop. Prior art patents describing
`
`flowmeters discuss implementing control loops having inner and outer control
`
`loops. (See, e.g., U.S. Pat. No. 4,524,610, Ex. 1031, Fig. 6; U.S. Pat. No. 4,655,089
`
`Ex. 1047, Fig. 9.)
`
`iii.
`
`Positive Feedback
`
`56.
`
`If the feedback signal is in phase with the error signal, the closed loop
`
`control loop is said to be operating with positive or regenerative feedback, i.e.,
`
`there is positive loop gain around the feedback loop. “Two alternating quantities
`
`are said to be ‘in phase’ when their maximum values occur at the same instant of
`
`4852-9264-2334.1
`
`22
`
`
`
`time.” (Dictionary of Mechanical Engineering, Fourth Edition, Nayler,
`
`Butterworth-Heinemann, 1996, Ex. 1013, p. 277.) When positive feedback occurs,
`
`disturbances or oscillations may naturally increase even without, for example, a
`
`reference input. An example of a closed loop system having positive feedback is
`
`the squealing sound produced by a loudspeaker to microphone feedback in a public
`
`address system.
`
`IV.
`
`SINUSOIDAL OSCILLATORS
`
`57. While most feedback control systems are designed to avoid positive
`
`feedback, electronic sinusoidal oscillators have long been designed to employ
`
`positive feedback to sustain oscillation. Coriolis flowmeters are, in fact, sinusoidal
`
`oscillators, in which the dynamic elements include the resonant mechanical
`
`dynamics of the conduit.
`
`58. With respect to a sinusoidal oscillation, such as that which is
`
`(desirably) established in a Coriolis flowmeter conduit, a regenerative positive
`
`feedback control loop was used in prior art flowmeters to maintain the amplitude
`
`of sinusoidal oscillation of the conduits at a desired value. (See, e.g., U.S. Pat. No.
`
`4,524,610, Ex. 1031, 7:45-50 (“a positive feedback circuit is interconnected
`
`between the detector coil 2a and the torsional driver coil la, and the power supplied
`
`to the driver coil la in order to maintain a predetermined amplitude of torsional
`
`oscillation of the tube 50 is measured”); U.S. Pat. No. 4,934,196, Ex. 1006, Fig. 4.)
`
`4852-9264-2334.1
`
`23
`
`
`
`59. Analog circuit engineers used positive (i.e., regenerative) feedback in
`
`sine wave oscillator designs at least back to the early 1970s to maintain oscillation
`
`amplitude, as described in the following excerpt:
`
`[The following figure] shows an amplifier, a feedback network,
`and an input mixing circuit not yet connected to form a closed
`loop.
`
`The amplifier provides an output signal x0 as a consequence of
`the signal xi applied directly to the amplifier input terminal. The
`output of the feedback network is xf = ßx0 = Aßxi, and the
`output of the mixing circuit (which is now simply an inverter) is
`x’f = -xf = -Aßxi
`From [the figure] the loop gain is
`Loop gain = x’f / xi = -x’f /xi = -ßA
`(Integrated Electronics: Analog and Digital Circuits and Systems, Jacob Millman
`
`and Christos Halkias, McGraw-Hill, 1972, Ex. 1032, p. 483.)
`
`60.
`
`In the figure above, if points 1 and 2 were connected and the
`
`relationship x’f = xi were true (i.e., the waveforms x’f and xi were identical in
`
`4852-9264-2334.1
`
`24
`
`
`
`amplitude, phase and frequency for a sinusoid), then loop gain -ßA would be equal
`
`to one (unity). Unity gain is called the Barkhausen Criterion, which expresses a
`
`necessary, but not sufficient, condition for sine wave oscillation in a positive
`
`feedback control loop. The Barkhausen Criterion is described in the following
`
`excerpt:
`
`The frequency at which a sinusoidal oscillator will operate is
`the frequency for which the total shift introduced, as a signal
`proceeds from the input terminals, through the amplifier and feedback
`network, and back again to the input, is precisely zero (or, of course,
`an integral multiple of 2π.) Stated more simply, the frequency of a
`sinusoidal oscillator is determined by the condition that the loop-gain
`phase shift is zero.
`Oscillations will not be sustained if, at the oscillator frequency,
`the magnitude of the product of the transfer gain of the amplifier and
`the magnitude of the feedback factor of the feedback network (the
`magnitude of the loop gain) are less than unity.
`(Integrated Electronics: Analog and Digital Circuits and Systems, Jacob Millman
`
`and Christos Halkias, McGraw-Hill, 1972, Ex. 1032, p. 484 (emphasis omitted and
`
`elsewhere