Direct digital speaker apparatus having a desired directivity pattern

REFERENCE TO CO-PENDING APPLICATIONS Priority is claimed from US provisional application No. 60/802,126 filed 22

May 2006 and entitled "An apparatus for generating pressure" and from a US provisional application No. 60/907,450 filed 2 April 2007 and entitled "Apparatus for generating pressure and methods of manufacture thereof and from US provisional application 60/924,203 filed 3 May 2007 and entitled "Apparatus and Methods for Generating Pressure Waves" .

FIELD OF THE INVENTION

The present invention relates generally to actuators and specifically to speakers.

BACKGROUND OF THE INVENTION

The state of the art for actuators comprising an array of micro actuators is believed to be represented by the following, all of which are US patent documents unless otherwise indicated:

2002/0106093: The Abstract, Figs. 1 - 42 and paragraphs 0009, 0023, and 0028 show electromagnetic radiation, actuators and transducers and electrostatic devices.

6,373,955: The Abstract and column 4, line 34 - column 5, line 55 show an array of transducers. JP 2001016675: The Abstract shows an array of acoustic output transducers.

6,963,654: The Abstract, Figs. 1-3, 7-9 and column 7, line 41 - column 8, line 54 show the transducer operation based on an electromagnetic force.

6,125,189: The Abstract; Figs. 1 - 4 and column 4, line 1 - column 5, line 46, show an electro-acoustic transducing unit including electrostatic driving.

WO 8400460: The Abstract shows an electromagnetic-acoustic transducer having an array of magnets.

4,337,379: The Abstract; column 3, lines 28-40, and Figs. 4, 9 show electromagnetic forces.

4,515,997: The Abstract and column 4, lines 16-20, show volume level. 6,795,561 : Column 7, lines 18-20, shows an array of micro actuators. 5,517,570: The Abstract shows mapping aural phenomena to discrete, addressable sound pixels.

JP 57185790: The Abstract shows eliminating the need for a D/A converter. JP 51120710: The Abstract shows a digital speaker system which does not require any D-A converter. JP 09266599: The Abstract shows directly applying the digital signal to a speaker.

6,959,096: The Abstract and column 4, lines 50-63 show a plurality of transducers arranged within an array.

Methods for manufacturing polymer magnets are described in the following publications:

Lagorce, L. K. and M. G. Allen, "Magnetic and Mechanical Properties of Micro- machined Strontium Ferrite/Polyimide Composites", IEEE Journal of Micro- electromechanical Systems, 6(4), Dec. 1997; and Lagorce, L. K., Brand, O. and M. G. Allen, "Magnetic micro actuators based on polymer magnets", IEEE Journal of Micro-electromechanical Systems, 8(1), March 1999.

United States Patent 4,337,379 to Nakaya describes a planar electrodynamics electro-acoustic transducer including, in Fig. 4A, a coil-like structure. United States Patent 6,963,654 to Sotme et al describes a diaphragm, flat-type acoustic transducer and flat-type diaphragm. The Sotme system includes, in Fig. 7, a coil-like structure.

Semiconductor digital loudspeaker arrays are known, such as those described in United States Patent document 20010048123, U.S. Patent 6,403,995 to David Thomas, assigned to Texas Instruments and issued 11 June 2002, U.S. Patent 4,194,095 to Sony, U.S. Patent 4,515,997 to Walter Stinger, and Diamond Brett M., et al, "Digital sound reconstruction using array of CMOS-MEMS micro-speakers", Transducers '03, The 12 ^th International Conference on Solid State Sensors, Actuators and Microsystems, Boston, June 8 - 12, 2003; and such as BBE's DS48 Digital Loudspeaker Management System.

YSP 1000 is an example of a phased array speaker manufactured by Yamaha.

The disclosures of all publications and patent documents mentioned in the specification, and of the publications and patent documents cited therein directly or indirectly, are hereby incorporated by reference.

SUMMARY OF THE INVENTION

Provided herewith, in accordance with certain embodiments of the present invention, is direct digital speaker apparatus receiving a digital input signal and generating sound accordingly, the apparatus comprising an array of pressure-producing elements such as but not limited to moving elements as described herein; and a controller operative to compute a timing pattern determining if and when each pressure- producing element is actuated so as to achieve a desired directivity pattern.

Further in accordance with a preferred embodiment of the present invention, at least one pressure-producing element is capable of producing positive pressure pulses and at least one pressure-producing element is capable of producing negative pressure pulses.

Still further in accordance with a preferred embodiment of the present invention, each pressure-producing element is operative to produce both positive pressure pulses and negative pressure pulses.

Also provided, in accordance with a preferred embodiment of the present invention, is a method for controlling direct digital speaker apparatus receiving a digital input signal and generating sound accordingly, the method comprising providing an array of pressure-producing elements, and computing a timing pattern determining if and when each pressure-producing element is operative to produce pressure pulses so as to achieve a desired directivity pattern.

Further in accordance with a preferred embodiment of the present invention, each pressure-producing element comprises a moving element, operating to travel alternately back and forth along a respective path Still further in accordance with a preferred embodiment of the present invention, the apparatus also comprises a user interface receiving a desired directivity pattern from a user.

Further in accordance with a preferred embodiment of the present invention, the directivity pattern is omni-directional defining a focal point. Still further in accordance with a preferred embodiment of the present invention, the directivity pattern is cylindrical defining a focal axis.

Further in accordance with a preferred embodiment of the present invention, the directivity pattern is unidirectional defining an angle of propagation.

Still further in accordance with a preferred embodiment of the present invention, the directivity pattern comprises a combination of a plurality of unidirectional directivity patterns.

Further in accordance with a preferred embodiment of the present invention, the array is centered at the focal point.

Still further in accordance with a preferred embodiment of the present invention, the array is centered at a projection of the focal point.

Further in accordance with a preferred embodiment of the present invention, the array is oriented symmetrically relative to the axis. Still further in accordance with a preferred embodiment of the present invention, the array is rectangular, defining four sides thereof, and the four sides include two sides parallel to the axis.

Additionally in accordance with a preferred embodiment of the present invention, the timing pattern comprises employing a suitable delay for at least some of the pressure-producing elements, using the formula: delay = [(d ² +r ² ) ^'5 -d]/c, where r = distance between the projection of the focal point onto the pressure-producing elements array and a given pressure-producing element, d = the distance of the plane of the array of the pressure-producing elements from the focal point of the omni-directional sound, and c = the speed of sound propagation through the medium in which the speaker is operating.

Still further in accordance with a preferred embodiment of the present invention, the timing pattern comprises employing a suitable delay for at least some of the pressure-producing elements, using the formula: delay = [(d ² +r ² ) ^{0 5} -d]/c, where r = distance between the projection of the focal axis onto the pressure-producing elements array and a given pressure-producing element, c = the speed of sound through the medium in which the speaker is operating, and d = the distance of the plane of the array of pressure-producing elements from the focal axis.

Further in accordance with a preferred embodiment of the present invention, the timing pattern comprises employing a suitable delay for at least some of the pressure- producing elements, using the formula: delay = x cos α where x = the distance from the plane defined by the pressure-producing elements array edge and a given pressure- producing element and α = the angle between the direction of the uni-directional propagation and the pressure-producing elements array plane.

Further in accordance with a preferred embodiment of the present invention, each of the pressure-producing elements is individually controlled.

Still further in accordance with a preferred embodiment of the present invention, the pressure-producing elements are moving elements, that produce pressure by virtue of their movement.

Still further in accordance with a preferred embodiment of the present invention, each moving element is responsive to alternating magnetic fields and wherein the apparatus also comprises at least one latch operative to selectively latch at least one subset of the moving elements in at least one latching position thereby to prevent the individual moving elements from responding to the electromagnetic force, and wherein the controller comprises a magnetic field control system operative to receive the clock and, accordingly, to control application of the electromagnetic force to the array of moving elements; and a latch controller operative to receive the digital input signal and to control the at least one latch accordingly. Further in accordance with a preferred embodiment of the present invention, the method also comprises reading in a desired directivity pattern provided by a user. Regarding terminology used herein:

Array: This term is intended to include any set of moving elements whose axes are preferably disposed in mutually parallel orientation and flush with one another so as to define a surface which may be planar or curved.

Above, Below: It is appreciated that the terms "above" and "below" and the like are used herein assuming that, as illustrated by way of example, the direction of motion of the moving elements is up and down however this need not be the case and alternatively the moving elements may move along any desired axis such as a horizontal axis.

Actuator: This term is intended to include transducers and other devices for inter-conversion of energy forms. When the term transducers is used, this is merely by way of example and it is intended to refer to all suitable actuators such as speakers, including loudspeakers. Actuator element: This term is intended to include any "column" of components which, typically in conjunction with many other such columns, forms an actuator, each column typically including a moving element, a pair of latches or "latching elements"

therefor, each latching element including one or more electrodes and insulative spacing material separating the moving element from the

Coil: It is appreciated that the alternating electromagnetic force applied to the array of moving elements in accordance with a preferred embodiment of the present invention may be generated by an alternating electric current oriented to produce a magnetic field gradient which is co-linear to the desired axes of motion of the moving elements. This electric current may comprise current flowing through a suitably oriented conductive coil or conductive element of any other suitable configuration.. The term "coil" is used throughout the present specification as an example however it is appreciated that there is no intention to limit the invention which is intended to include all apparatus for applying an alternating electromagnetic force e.g. as described above. When "coil" is used to indicate a conductor, it is appreciated that the conductor may have any suitable configuration such as a circle or other closed figure or substantial portion thereof and is not intended to be limited to configurations having multiple turns. Channels, also termed "holes" or "tunnels": These are illustrated as being cylindrical merely by way of example, this need not be the case.

Electrode: An electro-static latch. Includes either the bottom or top electro-static latch which latches its corresponding moving element by virtue of its being oppositely charged such that each latch and its moving element constitute a pair of oppositely charged electrodes.

Flexure: at least one flexible element on which an object is mounted, imparting at least one degree of freedom of motion to that object, for example, one or more flexible thin or small elements peripheral to and typically integrally formed e.g. from a single sheet of material, with a central portion on which another object may or may not be mounted, thereby to impart at least one degree of freedom of motion to the central portion and objects mounted thereupon.

Latch, latching layer, latching mechanism: This term is intended to include any device for selectively locking one or more moving elements into a fixed position. Typically, "top" and "bottom" latching layers are provided, which may be side by side and need not be one atop the other, and each latching layer includes one or many latching mechanisms which may or may not correspond in number to the number of moving elements to be latched. The term "latch pair" is a pair of latches for an

individual moving element e.g. including a top latch and a bottom latch, which may be side by side and need not be one atop the other.

Moving elements: These are intended to include any moving elements each constrained to travel alternately back and forth along an axis in response to an alternating electromagnetic force applied thereto. Moving elements are also termed herein "micro-speakers", "pixels", "micro-actuators", "membranes" (individually or collectively) and "pistons".

Spacers, also termed "space maintainers": Include any element or elements mechanically maintaining the respective positions of the electrodes and moving elements

The term "direct digital speaker" is used herein to include speakers that accept a digital signal and translate the signal into sound waves without the use of a separate digital to analog converter. Such speakers may sometime include an analog to digital converter as to allow them to translate analog signals instead or in addition to digital signals. Such speakers may include DDS (Direct Digital Speakers), DDL (Direct Digital Loudspeakers), DSR (Digital Sound Reconstruction) speakers, digital uniform loudspeaker arrays, matrix speakers, and MEMS speakers. The term "direct digital speaker" as used herein is intended to include speaker apparatus having a multiplicity of pressure-producing elements, which generate pressure either by virtue of their motion e.g. as specifically described herein or by heating and cooling the medium in which they reside, e.g. air, or by accelerating the medium in which they reside e.g. by ionizing the medium and providing a potential difference along an axis, or by operating as valves to selectively tap reservoirs of medium e.g. air, pressurized differently from the surrounding environment. The number of operating pressure producing elements (i.e. elements which are operating to generate pressure) is typically a monotonically increasing function of, e.g. proportional to, the intensity of the input signal, if analog, or to the digitally encoded intensity of the input signal, if digital.

The term "clock" used herein refers to the time duration associated with a single interval of the system clock. The term "directivity pattern" as used herein refers to the pattern of the spatial distribution of the acoustic energy generated by speaker apparatus.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the present invention are illustrated in the following drawings: Fig. IA is a simplified functional block diagram illustration of actuator apparatus constructed and operative in accordance with a preferred embodiment of the present invention.

Fig. IB is an isometric illustration of the array of moving elements of Fig. IA constructed and operative in accordance with a preferred embodiment of the present invention in which each moving element comprises a magnet and each is constrained to travel, except when latched, alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements.

Figs. 1C - IG are simplified top view illustrations of latches constructed and operative in accordance with five alternative embodiments of the present invention which can serve as alternatives to the latch specifically shown in Fig. IB.

Fig. 2 A shows the array of Fig. IB in a first, bottom extreme position responsive to an electromagnetic force applied downward.

Fig. 2B shows the array of Fig. IB in a second, top extreme position responsive to an electromagnetic force applied upward.

Fig. 2C is similar to Fig. 2B except that one of the individual moving magnets is not responding to the upward force because that individual magnet is latched into its top extreme position by a corresponding electric charge disposed above the individual moving magnet and functioning as a top latch. Figs. 3A - 3C are respective top, cross-sectional and isometric views of a skewed array of moving elements each constrained to travel alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements by a coil wrapped around the array.

Fig. 4A is an exploded view of an actuator device including an array of moving elements each constrained to travel alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements by a coil, and a latch, formed as a layer, operative to selectively latch at least one subset

of the moving elements in at least one latching position thereby to prevent the individual moving elements from responding to the electromagnetic force.

Fig. 4B is a simplified flowchart illustration of a preferred actuation method operative in accordance with a preferred embodiment of the present invention. Fig. 5 is an isometric static view of the actuator device of Fig. 4A constructed and operative in accordance with a preferred embodiment of the present invention in which the array of moving elements is formed of thin foil, each moving element being constrained by integrally formed flexures surrounding it.

Fig. 6 A is an exploded view of a portion of the actuator device of Fig. 5. Figs. 6B and 6C are a perspective view illustration and an exploded view, respectively, of an assembly of moving elements and associated flexures, latches and spacer elements constructed and operative in accordance with a preferred embodiment of the present invention which reduces leakage of air through the flexures.

Fig. 6D is a cross-sectional view of the apparatus of Figs. 6B - 6C showing three moving elements in top extreme, bottom extreme and intermediate positions respectively.

Fig. 6E is a legend for Fig. 6D.

Fig. 7A is a static partial top view illustration of the moving element layer of Figs. 5 - 6C. Fig. 7B is a cross-sectional view of the moving element layer of Figs. 5 - 6 taken along the A-A axis shown in Fig. 7A.

Fig. 7C is a perspective view of the moving element layer of Figs. 5 - 7B wherein an individual moving element is shown moving upward toward its top extreme position such that its flexures extend upward out of the plane of the thin foil. Fig. 7D is a perspective view of a moving element layer constructed and operative in accordance with an alternative embodiment of the present invention in which the disc-shaped permanent magnets of the embodiment of Figs. 5 - 7C are replaced by ring-shaped permanent magnets.

Fig. 7E is a side view illustration of the flexure-restrained central portion of an individual moving element in the embodiment of Fig. 7D.

Fig. 8A is a control diagram illustrating control of the latches and of the coil- induced electromagnetic force for a particular example in which the moving elements are arranged in groups that can each, selectively, be actuated collectively, wherein each

latch in the latching layer is associated with a permanent magnet, and wherein the poles of all of the permanent magnets in the latching layer are all identically disposed.

Fig. 8B is a flowchart illustrating a preferred method whereby a latching controller may process an incoming input signal and control moving elements' latches accordingly, in groups.

Fig. 8 C is a simplified functional block diagram illustration of a processor, such as the processor 802 of Fig. 8A, which is useful in controlling substantially any of the actuator devices with electrostatic latch mechanisms shown and described herein.

Fig. 8D is a simplified flowchart illustration of a preferred method for initializing the apparatus of Figs. 1 - 8C.

Fig. 8E is a simplified isometric view illustration of an assembled speaker system constructed and operative in accordance with a preferred embodiment of the present invention.

Fig. 8F is a simplified flowchart illustration of a preferred method of operation for generating a sound using apparatus constructed and operative in accordance with an embodiment of the present invention.

Fig. 9A is a graph summarizing certain, although typically not all, of the forces brought to bear on moving elements in accordance with a preferred embodiment of the present invention. Fig. 9B is a simplified pictorial illustration of a magnetic field gradient inducing layer constructed and operative in accordance with a preferred embodiment of the present invention.

Figs. 9C - 9D illustrate the magnetic field gradient induction function of the conductive layer of Fig. 9B. Fig. 1OA is a simplified top cross-sectional illustration of a latching layer suitable for latching moving elements partitioned into several groups characterized in that any number of moving elements may be actuated by collectively actuating selected groups from among the partitioned groups, each latch in the latching layer being associated with a permanent magnet, wherein the poles of all of the permanent magnets in the latching layer are all identically disposed.

Fig. 1 OB is a simplified electronic diagram of an alternative embodiment of the latch layer of Figs. 1 - 1OA in which each latch is individually controlled by the latching controller 50 of Fig. 8C. It is appreciated that the latches are shown to be

annular however alternatively may have any other suitable configuration as described herein. The layer of Fig. 1OB comprises a grid of vertical and horizontal wires defining junctions. A gate such as a field-effect transistor is typically provided at each junction. To open an individual gate thereby to charge the corresponding latch, voltage is provided along the corresponding vertical and horizontal wires.

Fig. HA is a timing diagram showing a preferred control scheme used by the latch controller in uni-directional speaker applications wherein an input signal representing a desired sound is received, and moving elements constructed and operative in accordance with a preferred embodiment of the present invention are controlled responsively, so as to obtain a sound pattern in which the volume in front of the speaker is greater than in other areas, each latch in the latching layer being associated with a permanent magnet, and the poles of all of the permanent magnets in the latching layer preferably all or substantially all being similarly or identically disposed. Fig. 1 IB is a schematic illustration of an example array of moving elements to which the timing diagram of Fig. 1 IA pertains.

Fig. HC is a timing diagram showing a preferred control scheme used by the latch controller in omni-directional speaker applications wherein an input signal representing a desired sound is received, and moving elements constructed and operative in accordance with a preferred embodiment of the present invention are controlled responsively, so as to obtain a sound pattern in which the volume in front of the speaker is similar to the volume in all other areas surrounding the speaker.

Figs. 12A and 12B are respectively simplified top view and cross-sectional view illustrations of the moving element layer in accordance with an alternative embodiment in which half of the permanent magnets are placed north pole upward and half north pole downward.

Fig. 13 is a simplified top view illustration similar to Fig. 1OA except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward.

Fig. 14 is a control diagram illustrating control of the latches and of the coil- induced electromagnetic force for a particular example in which the moving elements are arranged in groups that can each, selectively, be actuated collectively, similar to Fig.

8A except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward.

Fig. 15A is a timing diagram showing a preferred control scheme used by the latch controller in uni-directional speaker applications, which is similar to the timing diagram of Fig. 1 IA except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward.

Fig. 15B is a schematic illustration of an example array of moving elements to which the timing diagram of Fig. 15 A pertains.

Fig. 15C is a graph showing changes in the number of moving elements disposed in top and bottom extreme positions at different times and as a function of the frequency of the input signal received by the latching controller of Fig. 8C.

Fig. 16A illustrates a moving element layer which is an alternative to the moving element layer shown in Figs. IA and 2 A - 2C in which the layer is formed from a thin foil such that each moving element comprises a central portion and surrounding portions.

Fig. 16B is still another alternative to the moving element layer shown in Figs. IA and 2A - 2C in which a sheet of flexible material e.g. rubber capable of enabling motion i.e. there are rigid discs under the magnet, the magnet might be the rigid element but it might not be rigid enough.

Fig. 16C is an isometric view of a preferred embodiment of the moving elements and surrounding flexures depicted in Figs. 7 A - 7E or 16A in which flexures vary in thickness. Fig. 16D is an isometric illustration of a cost effective alternative to the apparatus of Fig. 16C in which flexures vary in width.

Fig. 17 is a top cross-sectional view illustration of an array of actuator elements similar to the array of Fig. 3A except that whereas in Fig. 3A, consecutive rows of individual moving elements or latches are respectively skewed so as to increase the number of actuator elements that can be packed into a given area, the rows in Fig. 17 are unskewed and typically comprise a rectangular array.

Fig. 18 is an exploded view of an alternative embodiment of an array of actuator elements in which the cross-section of each actuator element is square rather than round.

Fig. 19 is an isometric array of actuators supported within a support frame providing an active area which is the sum of the active areas of the individual actuator arrays.

Fig. 2OA is a simplified generally self-explanatory functional block diagram illustration of a preferred system for achieving a desired directivity pattern for a desired sound stream using a direct digital speaker with characteristics as indicated in the drawing, such as that shown and described herein in Figs. IA - 19.

Fig. 2OB is a simplified generally self-explanatory functional block diagram illustration of a preferred system for achieving a desired directivity pattern for a desired sound stream which is of general applicability in that it need not employ a direct digital speaker with characteristics as indicated in Fig. 2OA e.g. that shown and described herein in Figs. IA - 19 and may instead employ any suitable direct digital speaker.

Fig. 21 is a simplified flowchart illustration of per-clock operation of the moving element constraint controller 3050 of Fig. 20, in accordance with certain embodiments of the present invention.

Fig. 22 A is a simplified diagram of an omni-directional propagation pattern. Fig. 22B is a diagram of a preferred positioning of a moving element array relative to the focal point of the desired omni-directional sound propagation pattern of Fig. 22A.

Fig. 23 is a simplified pictorial illustration of speaker apparatus constructed and operative in accordance with Figs. 2OA - 22B and operative, e.g. by virtue of having been so programmed, to generate omni-directional sound which is particularly suitable for an environment in which consumers of the sound entirely surround the speaker, typically at more than one levels including a ground level and a first floor level as shown.

Fig. 24 is a diagram of a cylindrical pattern of sound directivity which it is achievable using an embodiment of the apparatus of the present invention.

Fig. 25 is a diagram showing one preferred positioning of the moving element array 3010, shown to be rectangular by way of example, relative to the cylindrical pattern of sound directivity shown in Fig. 24.

Fig. 26 is an isometric view of the moving element array of Figs. 2OA -20B, showing uni-directional sound generated by that moving array and propagating in a desired or predetermined direction α as indicated by arrows.

Fig. 27 is a pictorial illustration of a preferred application for speaker apparatus 3600 constructed and operative in accordance with the present invention, being constructed e.g. programmed to generate uni-directional sound in at least one typically user-selected direction.

Fig. 28 is a simplified pictorial illustration of a non-rectangular array of moving elements constructed and operative in accordance with a preferred embodiment of the present invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

The technical field of the invention is that of a digital transducer array of long- stroke electromechanical micro actuators constructed using fabrication materials and techniques to produce low cost devices for a wide variety of applications, such as audio speakers, biomedical dispensing applications, medical and industrial sensing systems, optical switching, light reflection for display systems and any other application that requires or can derive benefit from longer-travel actuation and/or the displacement of greater volumes of fluid e.g. air or liquid relative to the transducer size.

A preferred embodiment of the present invention seeks to provide a transducer structure, a digital control mechanism and various fabrication techniques to create transducer arrays with a number, N, of micro actuators. The array is typically constructed out of a structure of typically three primary layers which in certain embodiments would comprise of a membrane layer fabricated out of a material of particular low-fatigue properties that has typically been layered on both sides with particular polar aligned magnetic coatings and etched with a number, N, of unique "serpentine like" shapes, so as to enable portions of the membrane bidirectional linear freedom of movement (the actuator). The bidirectional linear travel of each moving section of the membrane is confined within a chamber (actuator channels) naturally formed typically by sandwiching the membrane layer between two mirror image support structures constructed out of dielectric, Silicon, Polymer or any other like insulating substrate, are typically fabricated with N precisely sized through holes equal

in number to the N serpentine etchings of the membrane and typically precisely positioned in a pattern which precisely aligns each through hole with each serpentine etching of the membrane. Further affixed to the outer surfaces of both the top and bottom layers of the support structure are, typically, conductive overhanging surfaces such as conductive rings or discs ("addressable electrodes"), which serve to attract and hold each actuator as it reaches its end of stroke typically by applying electrostatic charge.

A device constructed and operative in accordance with a preferred embodiment of the present invention is now described with reference to Figs. IB, 2A - 2C, 3A - 3C, 4A, 5, 6A, 7A - 7B, 8A - 8B, 9, 1OA, HA, 12A, 13, 14, 15A, 16A - C, 17 - 19.

Fig. IB is a conceptual overview of a small section of the device. Fig. 2 A depicts the movement of the moving elements under magnetic field. Fig. 2B depicts the movement of the same moving elements under an opposite magnetic field. Fig. 2C depicts the movement of the moving elements under a magnetic field while one electrode is charged. Figs. 3A - 3C are respective top, cross-sectional and perspective views of one preferred embodiment of the present invention.

Fig. 4A is an exploded view of a device constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 5 is a detailed illustration of a small section of the device constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 6A is an exploded view of the same small section. Fig. 7A is a pictorial illustration of a serpentine and moving elements subassembly constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 7B is an illustrative view of a single element, constructed and operative in accordance with a preferred embodiment of the present invention, in motion. Fig. 8A is a block diagram of a speaker system constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 8B is a flow diagram of the speaker system constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 9A illustrates a preferred relationship between the different forces applied to the moving elements. Fig. 1OA is a grouping view of the electrodes constructed and operative in accordance with a preferred embodiment of the present invention. Fig. HA is a timing and control chart constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 12A illustrates magnetic properties of moving elements

for an alternative addressing embodiment. Fig. 13 illustrates grouping of electrodes in an alternative addressing embodiment. Fig. 14 is a simplified block diagram of the speaker system in an alternative addressing embodiment. Fig. 15A is a timing and control chart for an alternative embodiment. Fig. 16A is a small section of the moving elements subassembly constructed and operative in accordance with a preferred embodiment of the present invention. Fig. 16B is a small section of a different embodiment of the moving elements subassembly, using a flexible substrate constructed and operative in accordance with a preferred embodiment of the present invention.

Whereas Figs. 3A - 3C above illustrate an array of elements in a honeycomb construction constructed and operative in accordance with a preferred embodiment of the present invention, Fig. 17 illustrates an array of elements in a square construction, which is constructed and operative in accordance with a preferred embodiment of the present invention: Fig. 18 is an exploded view of a small section of an embodiment using square shaped elements. Fig. 19 illustrates an apparatus using a plurality (array) of devices.

Effective addressing is typically achieved through unique patterns of interconnects between select electrodes and unique signal processing algorithms which typically effectively segments the total number of actuators in a single transducer, into N addressable actuator groups of different sizes, beginning with a group of one actuator followed by a group of double the number of actuators of its previous group, until all N actuators in the transducer have been so grouped.

To attain actuator strokes the transducer is typically encompassed with a wire coil, which, when electrical current is applied, creates an electromagnetic field across the entire transducer. The electromagnetic field causes the moving part of the membrane to move typically in a linear fashion through the actuator channels. If the current alternates its polarity, it causes the moving part of the membrane to vibrate. When electrostatic charge is applied to particular addressable electrode groups, it will typically cause all actuators in that group to lock at the end of the stroke, either on top or bottom of the support structure in accordance with the application requirement. Collectively the displacement provided by the transducer is achieved from the sum total of the N actuators that are not locked at any particular interval (super position).

The transducer construction is typically fully scalable in the number of actuators per transducer, the size of each actuator, and the length of stroke of each actuator, and

the number of addressable actuator groups. In certain embodiments, the actuator elements may be constructed by etching various shapes into a particular material, or by using layered metallic disks that have been coated with a flexible material or by using free floating actuator elements The membrane (flexure) materials may include Silicon, Beryllium Copper, Copper Tungsten alloys, Copper Titanium alloys, stainless steel or any other low fatigue material. The addressable electrodes of the support structure may be grouped in any pattern as to attain addressing . as appropriate for the transducer application. The addressable electrodes may be affixed such that contact is created with the membrane actuator or in such a manner that there is no physical contact with the membrane. The substrate material may be of any insulating material such as FR4, silicon, ceramic or any variety of plastics. In some embodiments the material may contain ferrite particles. The number of serpentine shapes etched into the membrane, or floating actuator elements and the corresponding channels of the support structure may be round, square or any other shape. The electromagnetic field may be created by winding a coil around the entire transducer, around sections of the transducer or around each actuator element or by placing one or more coils placed next to one or more actuator elements.

In certain embodiments a direct digital method is used to produce sound using an array of micro-speakers. Digital sound reconstruction typically involves the summation of discrete acoustic pulses of energy to produce sound-waves. These pulses may be based on a digital signal coming from audio electronics or digital media in which each signal bit controls a group of micro-speakers. In one preferred embodiment of the current invention, the nth bit of the incoming digital signal controls 2 ⁿ micro- speakers in the array, where the most significant bit (MSB) controls about half of the micro-speakers and the least significant bit (LSB) controls at least a single micro- speaker. When the signal for a particular bit is high, all of the speakers in the group assigned to the bit are typically activated for that sample interval. The number of speakers in the array and the pulse frequency determine the resolution of the resulting sound-wave. In a typical embodiment, the pulse frequency may be the source-sampling rate. Through the post application of an acoustic low-pass filter from the human ear or other source, the listener typically hears an acoustically smoother signal identical to the original analog waveform represented by the digital signal.

According to the sound reconstruction method described herein, the generated sound pressure is proportional to the number of operating speakers. Different frequencies are produced by varying the number of speaker pulses over time. Unlike analog speakers, individual micro-speakers typically operate in a non-linear region to maximize dynamic range while still being able to produce low frequency sounds. The net linearity of the array typically results from linearity of the acoustic wave equation and uniformity between individual speakers. The total number of non-linear components in the generated sound wave is typically inversely related to the number of micro-speakers in the device. In a preferred embodiment a digital transducer array is employed to implement true, direct digital sound reconstruction. The produced sound's dynamic range is proportional to the number of micro-speakers in the array. The maximal sound pressure is proportional to the stroke of each micro-speaker. It is therefore desirable to generate long stroke transducers and to use as many as possible. Several digital transducer array devices have been developed over the years. One worth mentioning is a CMO S -MEMS micro-speaker developed at Carnegie Mellon University. Using CMOS fabrication process, they designed an 8-bit digital speaker chip with 255 square micro-speakers, each micro-speaker 216μm on a side. The membrane is composed of a serpentine Al- SiO2 mesh coated with polymer and can be electrostatically actuated by applying a varying electrical potential between the CMOS metal stack and silicon substrate. The resulting out of plane motion is the source of pressure waves that produce sound. Each membrane has a stroke of about lOμm. Such short strokes are insufficient and the generated sound levels are too soft for a loudspeaker. Another issue is that the device requires a driving voltage of 40V. Such voltage requires complex and expensive switching electronics. Preferred embodiments of the device described herein overcome some or all of these limitations and generate much louder sound levels while eliminating the need for high switching voltages.

It is believed that the shape of each transducer has no significant effect on the acoustic performance of the speaker. Transducers may be packed in square, triangle or hexagonal grids, inter alia.

The current invention typically makes use of a combination of magnetic and electrostatic forces to allow a long stroke while avoiding the problems associated with traditional magnetic or electrostatic actuators.

The moving elements of the transducer array are typically made to conduct electricity and may be magnetized so that the magnetic poles are perpendicular to the transducer array surface. Moderate conduction is sufficient. A coil surrounds the entire transducer array or is placed next to each element and generates the actuation force. Applying alternating current or alternating current pulses to the coil creates an alternating magnetic field gradient that forces all the moving elements to move up and down at the same frequency as the alternating current. To control each moving element, two electrodes may be employed, one above and one below the moving elements.

The current applied to the coil typically drives the moving elements into close proximity with the top and bottom electrode in turn. A small electrostatic charge is applied to the moving elements. Applying an opposite charge to one of the electrodes generates an attracting force between the moving element and the electrode. When the moving element is very close to the electrode, the attracting force typically becomes larger than the force generated by the coil magnetic field and the retracting spring and the moving element is latched to the electrode. Removing the charge or some of it from the electrode typically allows the moving element to move along with all the other moving elements, under the influence of the coil magnetic field and the flexures.

In accordance with certain embodiments, the actuator array may be manufactured from 5 plates or layers: - Top electrode layer

- Top spacers (together shown as layer 402)

- Moving elements 403

- Bottom spacers

- Bottom electrode layer (together shown as layer 404)

In accordance with certain embodiments, the array is surrounded by a large coil 401. The diameter of this coil is typically much larger than that of traditional coils used in prior art magnetic actuators. The coil can be manufactured using conventional production methods. In certain embodiments the moving element is made of a conductive and magnetic material. Moderate electrical conduction is typically sufficient. The moving element may be manufactured using many types of materials, including but not limited to rubber, silicon, or metals and their alloys. If the material cannot be magnetized or a

stronger magnet is desired, a magnet may be attached to it or it may be coated with magnetic material. This coating is typically done by application, using a screen printing process or other techniques known in the art, by epoxy or another resin loaded with magnetic powder. In some embodiments, screen printing can be performed using a resin mask created through a photo-lithographic process. This layer is typically removed after curing the resin/magnetic powder matrix. In certain embodiments the epoxy or resin is cured while the device is subjected to a strong magnetic field, orienting the powder particles in the resin matrix to the desired direction. The geometry of the moving elements can vary. In yet other embodiments, part of the moving elements may be coated with the magnet and cured with a magnetic field oriented in one direction while the rest are coated later and cured in an opposite magnetic field causing the elements to move in opposite directions under the same external magnetic field. In one preferred embodiment, the moving element comprises a plate that has a serpentine shape surrounding it, typically cut out from thin foil. Alternatively, in certain embodiments it is possible to use a thick material thinned only in the flexure area or by bonding relatively thick plates to a thin layer patterned as flexures. This shape allows part of the foil to move while the serpentine shape serves as a compliant flexure. In certain other embodiments, the moving part is a cylinder or a sphere, free to move about between the top and bottom electrodes. Fig. IB, which illustrates a conceptual overview of small section of the device in accordance with certain embodiments of the invention, serves to provide a conception overview of the complete transducer array structure. In the illustrated embodiment the moving elements are pistons 101 which are typically magnetized so that one pole 102 is on the top and the other 103 at the bottom of each piston. A magnetic field generator (not shown) that typically influences the entire transducer array structure creates a magnetic field across the entire transducer array, typically causing pistons 101 to move up and down, thereby forcing the air out of the cavity 104. An electrostatic electrode typically resides on both the top 105 and bottom 106 of each cavity. The electrodes serve as latching mechanisms that attract and hold each piston as it nears its end of stroke typically preventing the piston from moving until the latch is released, while allowing the pushed air to easily pass through. In certain embodiments, the pistons 101 are made of an electrically conductive material or coated with such material. At least

one of the elements, the piston and/ or the electrostatic electrode is typically covered by an dielectric layer to avoid shorting as pull-down occurs .

Figs. 2A - 2C, taken together, illustrate the element movement according to a preferred embodiment of the present invention. In this embodiment a coil (not shown) 5 typically surrounds the entire transducer array structure, creating a magnetic field across the entire transducer array which causes any magnetic element with freedom of movement to travel according to the alternating direction of the field. This causes the pistons to move typically up and down.

In Fig. 2A the magnetic field 201 direction is downwards. The magnetic field 10 creates a force, driving the pistons 101 of the entire array downwards.

In Fig. 2B the magnetic field 202 direction has changed and is pointing upwards. The magnetic field creates a force, driving the pistons 101 of the entire array upwards.

In Fig. 2C a positive electric charge is applied to one of the top electrodes 205. 15 The positive charge typically attracts the electrons in the piston 204, causing the top of the piston 206 to be negatively charged. The opposite charges 205 and 206 create an attraction force which, when the gap is below a critical distance, typically act to pulldown the two elements together. The magnetic field 203 direction has changed again and is pointing downwards. The piston 204 is typically held in place due to magnetic 20 attraction while the rest of the pistons are free to move, and move to the bottom due to the influence of the magnetic field 203. In this particular embodiment the charge applied to the electrode is positive. Alternatively, a negative charge may be applied to the electrodes, which will induct a negative charge to accumulate in the near-side of the adjacent piston.

25 Figs. 3 A - 3 C show top, cross-sectional and perspective views of one preferred embodiment.

In certain embodiments a coil 304 wrapped around the entire transducer array generates an electromagnetic field across the entire array structure, so that when current is applied, the electromagnetic field causes the pistons 302 to move up 301 and down 30 303.

Fig. 4A shows an exploded view of the device constructed and operative in accordance with certain embodiments of the invention. As shown, the exploded view of a transducer array structure reveals that it comprises the following primary parts:

(a) A coil surrounding the entire transducer array 401 generates an electromagnetic field across the entire array structure when voltage is applied to it. A preferred embodiment for the coil is described herein with reference to Figs. 9B - 9D.

(b) In certain embodiments a top layer construction 402 may comprise a spacer layer and electrode layer. In a certain embodiment this layer may comprise a printed circuit board (herein after "PCB") layer with an array of accurately spaced cavities each typically having an electrode ring affixed at the top of each cavity.

(c) The moving elements ("pistons") 403 in the current embodiment may be comprised of a thin foil of conductive magnetized material cut or etched with many very accurate plates typically surrounded by "serpentine" shapes that serve as compliant flexures that impart the foils with a specific measure of freedom of movement.

(d) A bottom layer construction 404 may comprise a spacer layer and electrode layer. In a certain embodiment this layer may comprise a dielectric layer with an array of accurately spaced cavities each typically having an electrode ring affixed at the bottom of each cavity.

Fig. 5 shows details of a small section of a device constructed and operative in accordance with a preferred embodiment of the present invention. A cross section detailed dimensional view of the transducer array according to the illustrated embodiment shows the following structure: the moving elements ("pistons"), typically made from a thin foil 501 that has been cut or etched into precise plate and serpentine shapes having a magnetized layer on the top 502 and bottom 503, is accurately positioned so the center of each plate shape is precisely aligned with the center of each of the cavities of a top layer dielectric 504 and the bottom layer dielectric cavity 505 that collectively serve as travel guides and air ducts. At the external edges of each duct both on the top 506 and on the bottom 507 is a copper ring ("electrode") latching mechanism which, when electrostatic charge is applied, typically attracts each moving element to create contact between the moving elements ("pistons") and latches and holds each moving element ("piston") as it nears the end of each stroke, thereby preventing the moving element ("piston") from moving until the latch is released typically by terminating the electrostatic charge to the electrode. .

Fig. 6 A shows an exploded view of the same small section as shown in Fig. 5. and reveals that in this embodiment the thin foil which has been etched with precise serpentine shapes to create a moving element ("piston") with the center of each shape

affixed with a magnetized layer on the top and bottom, is centered and enclosed in the cavities of mirror image on the top 602 and bottom 603 dielectric.

Fig. 7 A shows a serpentine shape and moving elements subassembly constructed and operative in accordance with a preferred embodiment of the present invention. A top static view of the thin foil shows the moving element in this embodiment is typically constructed by etching a precise round serpentine shape that allows the center of the shape 701 freedom of movement restrained by the flexures of the shapes 703 which have been etched out of the material, thereby to form interspersing cavities 702. A cross sectional view reveals that the foil typically has polar aligned layers of magnets, affixed to both the top 704 and the bottom 705 of the tin foil moving element layer. As an alternative to this embodiment, a layer of magnets may be affixed only to one side of the thin foil.

Fig. 7B is an illustrative view of a single element in motion, showing the upward freedom of movement of certain embodiments where the magnetized center 706 of a single serpentine shape is free to extend upward while being guided and restrained by the serpentine etched flexures 707. Not shown in the illustration is the opposite (downward) movement of the serpentine shape as it travels in the opposite direction, and by doing so the flexures extend downward.

In certain embodiments the top of each shape center 708 and the bottom of each layer 709 are affixed magnetized layers that have been aligned in the same magnetic polarity.

Fig. 8A shows a block diagram of the speaker system in accordance with a preferred embodiment of the present invention. In certain embodiments the digital input signal (common protocols are I2S, I2C or SPDIF) 801 enters into a logic processor 802 which in turn translates the signal to define the latching mechanism of each grouping of moving elements. Group addressing is typically separated into two primary groups, one for latching the moving elements at the top, and one for latching the moving elements at the bottom of their strokes. Each group is typically then further separated into logical addressing groups typically starting with a group of at least one moving element, followed by another group that doubles the moving elements of the previous group, followed by a another group which again doubles the number of elements of the previous, and so on, until all moving elements of the entire array have been grouped.

The Nth group comprises 2 ^" moving elements.

In the embodiment depicted in the block diagram of Fig. 8A, the top group of one element group 803, a two element group 804 and then a four element group 805 are shown and so on, until the total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 802. The same grouping pattern is typically replicated for the bottom latching mechanisms where a one element group 807 may be followed by a two element group 808 and then a four element group 809 and so on, until the total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 802. The processor 802 may also control an alternating current flow to the coil that surrounds the entire transducer array 812, thus creating and controlling the magnetic field across the entire array. In certain embodiments a power amplifier 811 may be used to boost current to the coil.

Fig. 8B illustrates a flow diagram of the speaker system. In certain embodiments where the sampling rate of the digital input signal 813 might be different from the device natural sampling rate, the resampling module 814 may re-sample the signal, so that it matches the device's sampling-rate. Otherwise, the resampling module 814 passes the signal through as unmodified.

The scaling module 815 typically adds a bias level to the signal and scales it, assuming the incoming signal 813 resolution is M bits per sample, and the sample values X range between -2 ^(M - ^]) and 2 ^(M -')-l.

It is also assumed that in certain embodiments the speaker array has N element groups (numbered 1..N), as described in Fig. 8A.

K is defined to be: K=N-M Typically, if the input resolution is higher than the number of groups in the speaker (M>N), K is negative and the input signal is scaled down. If the input resolution is lower than the number of groups in the speaker (M<N), K is positive and the input signal is scaled up. If they are equal, the input signal is not scaled, only biased. The output Y of the scaling module 815 may be: Y=2 ^K [X+2 ^M" ]. The output Y is rounded to the nearest integer. The value of Y now ranges between 0 and 2N- 1.

The bits comprising the binary value of Y are inspected. Each bit controls a different group of moving elements. The least significant bit (bitl) controls the smallest group (group 1). The next bit (bit2) controls a group twice as big (group 2). The next bit

(bit3) controls a group twice as big as group 2 etc. The most significant bit (bitN) controls the largest group (group N). The states of all the bits comprising Y are typically inspected simultaneously by blocks 816, 823, ... 824.

The bits are handled in a similar manner. Following is a preferred algorithm for inspecting bit 1 :

Block 816 checks bitl (least significant bit) of Y. If it is high, it is compared to its previous state 817. If bitl was high previously, there is no need to change the position of the moving elements in group 1. If it was low before this, the processor waits for the magnetic field to point upwards, as indicated by reference numeral 818 and then, as indicated by reference numeral 819, the processor typically releases the bottom latching mechanism Bl, while engaging the top latching mechanism Tl, allowing the moving elements in group 1 to move from the bottom to the top of the device.

If block 816 determines that bitl of Y is low, it is compared to its previous state 820. If bitl was low previously, there is no need to change the position of the moving elements in group 1. If its previous state was high, the processor waits for the magnetic field to point downwards, as indicated by reference numeral 821 and then, as indicated by reference numeral 822, the processor releases the top latching mechanism Tl while engaging the top latching mechanism Bl, allowing the moving elements in group 1 to move from the top to the bottom of the device. Fig. 9 A shows typical relationships between the different major forces applied to moving elements. The different forces being applied to the moving elements typically work in harmony to counterbalance each other in order to achieve the desired function. Forces toward the center are shown as negative forces, while forces driving the element further away from the center (either toward the up or down latching mechanisms) are shown as positive forces.

In the present embodiment the moving element is influenced by 3 major forces: a. Magnetic force, created by the interaction of the magnetic field and the hard magnet. The direction of this force depends on the polarity of the moving element magnet, the direction of the magnetic field and the magnetic field gradient. b. Electrostatic force, typically created by applying a certain charge to the electrode and an opposite charge to the moving element. The direction of this force is such as to attract the moving element to the electrode (defined as positive in this figure).

This force increases significantly when the distance between the moving element and

the electrode becomes very small, and/or where this gap comprises material with a high dielectric constant. c. Retracting force created by the flexures, (which act as springs). The direction of this force is always towards the center of the device (defined as negative in this figure). This force is relatively small since the flexures are compliant, and is linear in nature.

The relationship between the forces shows that typically, as the moving element increasingly nears the end of its stroke, the electrostatic force (generated by the latching mechanism) increases, ultimately achieving sufficient force to attract and latch the moving element. When the latch is released, the retracting and magnetic forces are typically able to pull the moving element away from the latch toward the center, thereby inducing travel of the moving element. As the moving element travels to the center, typically, the retracting force of the flexure diminishes and ultimately is overcome, and is then controlled by the electromagnetic force and the kinetic energy of the moving element.

Fig. 1OA shows a sectional view of the grouping pattern applied in certain embodiments to the moving element ("pistons") for purposes of digital addressing, as described previously in Fig. 8. In this embodiment there is a group of one element in the center 1001 followed by a two element group 1002, followed by a four element group 1003, followed by a eight element group 1004, followed by a 16 element group 1005, and so on.

As shown in this embodiment, to the extent possible each increasing group has been arranged to extend around the previous group, however this geometrical configuration can be altered in order to accomplish different audio and /or constructive objectives. For example moving the "epicenter" to the outer circumference of the transducer array enables easier wire routing between each group and the processor 802 (refer to Figs. 8A - 8B).

Fig. HA shows a preferred timing and control chart. The time chart describes preferred logic and algorithms for generating a specific sound waveform. In the scope of this description, the timeline is divided into slots, numbered II, 12 and so on. This simple example shows a device that uses 7 moving elements divided into 3 groups. The first group comprises one moving element "Pl" and is controlled by the top latching mechanism "Tl" and the bottom latching mechanism "Bl". The second group

comprises two moving elements "P2" and "P3" which are synchronized and move together. This group is controlled by the top latching mechanism "T2" and the bottom latching mechanism "B2". The second group comprises four moving elements "P4", "P5", "P6" and "P7", which are synchronized and move together. This group is controlled by the top latching mechanism "T3" and the bottom latching mechanism "B3".

The "clock" chart at the top of the figure represents the system clock. This clock is typically generated outside the device and is transferred to the processor 802 (refer to Fig. 8) alongside the sound signal. In a typical embodiment, the sampling rate of the device is 44100 Hz. In such a case, the duration of each clock interval is 22μsec and the clock changes its state every 11 μsec.

The "signal" shown in this example is the analog waveform that the device is generating. The "value" chart shows the digital sample value of the signal at each clock interval. The "magnetic" chart shows the direction (polarity) of the magnetic field generated by the coil. The polarity changes synchronously with the system clock.

This figure shows the state of each moving element using the following display convention: An element ("Pl" .. "P7") that is latched at the top 1101 is colored in black. An element that is latched at the bottom 1102 is colored in white and an element that is moving 1103 is hatched. The digital sample value dictates how many elements may be latched to the top and how many to the bottom of the array. In this example, digital sample values of -3, - 2, -1, 0, 1, 2, 3, and 4 are possible. Each value is represented by 0, 1, 2, 3, 4, 5, 6 and 7 elements, respectively, latched to the top.

In time slice Il the digital sample value is 0. This requires 3 elements latched to the top and 4 to the bottom. The magnetic field polarity is up. The top latching mechanisms Tl and T2 are engaged and so is the bottom latching mechanism B3. At the same time, the bottom latching mechanisms Bl and B2 are disengaged and so is the top latching mechanism T3. Moving elements Pl, P2 and P3 are latched to the top while P4, P5, P6 and P7 are latched to the bottom. In time slice 13, the digital sample value changes to 1. This requires 4 elements latched to the top and 3 to the bottom. The magnetic field polarization is up. The bottom latch B3 is disengaged, releasing elements P4, P5, P6 and P7 to move freely. At the

same time, the top latching mechanism T3 is engaged. The elements move upwards under the influence of the magnetic field and are latched by the currently engaged T3.

At this point, all 7 moving elements are latched to the top. In the next slice 114, the moving elements Pl, P2 and P3 would be latched to the bottom, to ensure the device is in the desired state (4 elements at the top and 3 at the bottom). In slice 14, the polarity of the magnetic field changes and is directed downwards. The top latching mechanisms Tl and T2 disengage and release the moving elements Pl, P2 and P3. At the same time, the bottom latching mechanisms Bl and B2 are engaged and the approaching moving elements Pl, P2 and P3 are latched to the bottom position. The moving elements P4, P5, P6 and P7 are held in place by the top latching mechanism T3 and are therefore restrained from moving downwards along with the other moving elements. The state of the device at this point is: Pl, P2 and P3 are latched to the bottom and P4, P5, P6 and P7 are latched to the top. In time slices 15 to 14, the latching mechanisms are engaged and disengaged to allow the moving elements to move and change their state according to the digital sample values.

Fig. 12A shows preferred magnetic properties of moving elements for addressing an alternative embodiment. A static top view of the moving element foil shows one possible alternative embodiment to the moving elements. In this embodiment two distinct group segments of the moving elements 1201 and 1202 have been created, enabling a single transducer array to process and generate a louder signal, or alternatively two separate signals (such as the left and right audio signal of stereo). The cross section view shows that in order to accomplish the two groups of this embodiment (discernible by the separated line 1203), each distinct group segment typically has opposite magnetic polarity. In one section group 1201 the layer of magnets affixed to the moving element of the thin foil has been polarized so that North (N) is on the top side of the foil 1204 and South (S) is on the bottom side 1205; while in the second section group 1202 the layer of magnets of the thin foil moving element have been polarized so that South (S) is on the top side of the foil 1206 and North (N) is on the bottom side 1207. Fig. 13 shows grouping of electrodes in an alternative embodiment. Similar to

Fig. 1OA, Fig. 13 depicts an alternative addressing scheme for the alternative embodiment that is described in Fig. 12 A. In this case the grouping pattern applied to the moving element for purposes of digital addressing is divided into two primary group

segments, half the transducer array in one primary segment group, and the other half in another primary segment group, as described in Fig. 12 A.

In this embodiment there are two equal groups each with an equal number of moving elements beginning with two groups 1301 and 1302 of one moving element each followed by two groups 1303 and 1304 with two elements in each group followed by two groupings 1305 and 1306 of four elements in each group, followed by two grouping 1307 and 1308 of eight elements in each group, followed by two groupings 1309 and 1310 of sixteen elements in each group and so on, until all moving elements of the transducer array have been grouped and addressed. As shown in the current embodiment, to the extent possible, each increasing group has been arranged to extend around the previous group, however this geometrical configuration can be altered in order to accomplish different audio and /or constructive objectives, for example moving the "epicenters" to the primary groups to opposite sides of the outer circumference of the transducer array enables easier wire routing between each group and the processor 1402 (refer to Fig. 14). It also enables the device to operate in two modes: monophonic, where both groups are used to generate one waveform at twice the amplitude, and stereophonic, where each group generates a separate sound wave, as to allow reconstruction of a stereophonic signal.

Fig. 14 shows a block diagram of the speaker system in an alternative addressing embodiment. Fig. 14 describes addressing of the alternative embodiment shown in Figs. 12 and 13. The digital input signal (I2S, I2C or SPDIF protocols) 1401 enters a logic processor 1402 which in turn translates the signal to define the latching mechanism of each the two primary grouping of moving elements. Each addressing group is separated into two primary groups, one for top and one for bottom latching mechanisms. Each group is then further separated into logical addressing groups starting with a group of one moving element, followed by another group that doubles the moving elements of the previous group, followed by a another group of double the number of elements of the previous group, and so on, until all moving elements of the entire array have been grouped. In the embodiment depicted in the block diagram of Fig. 14, the top stroke of one primary segments of moving elements begins with a one element group 1403, and then a two element group 1404, and then a four element group 1405, and so on, until the

total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 1402.

The same grouping pattern is replicated for the down stroke where a group of one element 1407 is followed by a two element group 1408, and then a four element group 1409, and so on, until the total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 1402.

This same pattern is replicated for the second primary segment of moving elements with the top stroke group starting with a one element group 1413, and then a two element group 1414, and then a four element group 1415, and so on, until the total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 1402.

This is replicated for the down stroke of the second segment beginning with a group of one element 1417, followed by a two element group 1418, and then a four element group 1419, and so on, until the total numbers of moving elements in the transducer array assembly have been addressed to receive a control signal from the processor 1402.

The processor 1402 will also control an alternating current flow to the coil that typically surrounds the entire transducer array, including both primary segments 1412, thus creating and controlling the magnetic field across the entire array. In certain embodiments a power amplifier 1411 may be used to boost current to the coil.

Fig. 15A shows a timing and control chart for an alternative embodiment. A time chart, describing the logic and algorithms, may be used to generate a specific sound waveform in the alternative embodiment described in Figs. 12 through 14. The display conventions are similar to those used in Fig. HA, and the same signal is reproduced.

The timeline is divided into slots, numbered II, 12 and so on. This simple example shows a device that uses 14 moving elements divided into two major groups (L and R), each divided into 3 minor groups 1, 2 and 3.

The digital sample value dictates how many elements may be latched to the top and how many to the bottom of the array. In this example, digital sample values of -3, - 2, -1, 0, 1, 2, 3, and 4 are possible. Each value is represented by 0, 2, 4, 6, 8, 10, 12 and 14 elements, respectively, latched to the top.

On time slice 13, the digital sample value changes from O to 1. This requires 8 elements latched to the top and 6 to the bottom. The magnetic field polarization is up. The top latches RTl and RT2 as well as the bottom latch LB3 are disengaged, releasing elements RPl, RP2, RP3, LP4, LP5, LP6 and LP7 to move freely. The magnetic polarity of LP4, LP5, LP6 and LP7 creates an upwards force, driving these elements upwards. The magnetic polarity of RPl, RP2 and RP3 is opposite and the driving force is downwards. At the same time, the latching mechanisms opposite to the element movement are engaged to grab the approaching moving elements and latch them in place. On slice 14, the polarity of the magnetic field changes and is directed downwards. The top latches LTl and LT2 as well as the bottom latch RB3 are disengaged, releasing elements LPl, LP2, LP3, RP4, RP5, RP6 and RP7 to move freely. The magnetic polarity of RP4, RP5, RP6 and RP7 creates an upwards force, driving these elements upwards. The magnetic polarity of LPl, LP2 and LP3 is opposite and the driving force is downwards. At the same time, the latching mechanisms opposite to the element movement are engaged to grab the approaching moving elements and latch them in place.

On time slices 15 to 114, the latching mechanisms are engaged and disengaged to allow the moving elements to move and change their state according to the digital sample values.

Fig. 15C illustrates production of three different pitches (22 KHz, 11 KHz and 4.4 KHz) of sound graphs II - IV respectively. Graph I shows the system clock which, in the illustrated example is 44 KHz. In the illustrated embodiment, the speaker used to generate these pitches has 2047 moving elements. When the 22 KHz sound (half of the clock) is generated, all 2047 elements change position (from top to bottom or vice versa) at each clock. When the 11 KHz (quarter of the clock) sound is generated, half of the 2047 moving elements change position at each clock. For example, if in the first clock all 2047 moving elements are in their top position, in the second clock, 1023 of these are lowered, in the third clock the remaining 1024 elements are lowered, in the fourth clock 1023 are raised, in the fifth clock the remaining 1024 elements are raised, and so forth. When the 4.4 KHz (1/10 of the clock) sound is generated, the numbers of elements which are in their top position at each clock (1340, 1852, ...) are shown on top

of Graph IV whereas the numbers of elements which are in their bottom position at each clock (707, 195, ...) are shown on the bottom of Graph IV.

Fig. 16A shows a small section of the moving elements subassembly.

Figs. 16A and 16B provide illustrated views of the moving elements in different embodiments.

The embodiment shown in Fig. 16A is of moving elements ("Pistons") constructed from a thin foil material 1601 with a precise round serpentine shape etched into the material which enables the center of the shape 1602 freedom of movement that is restrained by the flexures of the shape. Fig. 16B shows a small section of a different embodiment of the moving elements subassembly, using a flexible substrate. This embodiment is of moving elements ("pistons") constructed from a material with sufficient elasticity, such as rubber polyethylene material 1603, which either has magnetic material deposits in specific shapes and dimensions on the top and bottom of the material surface, or the material is affixed to a magnetized disk of particular dimensions 1604, enabling freedom of movement that is restrained by the material itself.

Fig. 2C shows a small section of a different embodiment of the moving elements subassembly, using free-floating components. This embodiment is of free floating moving elements ("pistons") constructed from magnetized material with polar opposites at each end. In this particular embodiment North is on top and South on the bottom.

Fig. 3 B illustrates a top view of a complete transducer array structure in certain embodiments, based on a honeycomb design, which enables a fill factor of 48 percent of the surface area. Fig. 17 illustrates a top view of a completed transducer array structure in certain embodiments, based on a square design, which enables a fill factor of 38 percent of the surface area.

Fig. 18 shows an exploded view of a small section of an embodiment using square shaped elements. This embodiment shows a transducer array structure that utilizes square shape elements intended to increase the fill factor and allow higher sound pressure levels per transducer area. As in previous embodiments, the same structural elements are used. A coil surrounds the entire transducer array (not shown). When voltage is applied, the coil generates an electromagnetic actuation force across the entire array structure.

A top layer construction, typically comprising a dielectric layer with an array of accurately spaced cavities 1802, each having an electrode ring, is affixed at the top of each cavity, to create an electrostatic latching mechanism 1801.

The moving elements ("pistons") in this embodiment comprises a thin foil of conductive magnetized material cut or etched with many very accurate "serpentine" shapes, that imparts the foils a specific measure of freedom of movement 1803 with a magnetized top 1804 and bottom 1805. Each moving element is guided and restrained by four flexures.

A bottom layer construction, typically comprising a dielectric layer with an array of accurately spaced cavities 1806, each having an electrode ring affixed at the bottom of each cavity, creates an electrostatic latching mechanism 1807.

Fig. 19 shows an apparatus including a plurality (array) of devices. The structure shows the use of plurality in certain embodiments of array transducers 1902 as to create a device 1901 capable of generating louder sound pressure levels or use beam- forming techniques (which extend beyond the scope of this invention) to create directional sound waves.

The array may have any desired shape, and the round shapes in the description are only for illustrative purposes.

The device constructed and operative in accordance with one embodiment of the present invention and described above with reference to Figs. IB, 2A - 2C, 3 A - 3C,

4A, 5, 6A, 7A - 7B, 8A - 8B, 9A, 1OA, HA, 12A, 13, 14, 15A, 16A - C, 17 - 19 is now described both more generally, e.g. with reference to Fig. IA, and in further detail.

Alternative embodiments are also described.

Reference is now made to Fig. IA which is a simplified functional block diagram illustration of actuator apparatus for generating a physical effect, at least one attribute of which corresponds to at least one characteristic of a digital input signal sampled periodically in accordance with a clock. According to a preferred embodiment of the present invention, the apparatus of Fig. IA comprises at least one actuator device, each actuating device including an array 10 of moving elements each typically constrained to travel alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array 10 of moving elements. Each moving element is constructed and operative to be responsive to electromagnetic force.

Each moving element may therefore comprise a conductor, may be formed of a ferro

magnetic material, may comprise a permanent magnet e.g. as shown in Fig. 6C, and may comprise a current-bearing coil.

A latch 20 is operative to selectively latch at least one subset of the moving elements 10 in at least one latching position thereby to prevent the individual moving elements 10 from responding to the electromagnetic force. An electromagnetic field controller 30 is operative to receive the clock and, accordingly, to control application of the electromagnetic force by a magnetic field generator, 40, to the array of moving elements. A latch controller 50 is operative to receive the digital input signal and to control the latch accordingly. The latch controller 50, in at least one mode of latch control operation, is operative to set the number of moving elements 10 which oscillate freely responsive to the electromagnetic force applied by the magnetic field generator, e.g. coil 40 to be substantially proportional to the intensity of the sound, coded into the digital input signal it receives. Preferably, when the intensity of sound coded into the digital input signal is at a positive local maximum, all moving elements are latched into a first extreme position. When the intensity of sound coded into the digital input signal is at a negative local maximum, all moving elements are latched into a second, opposing, extreme position.

Preferably, a physical effect, e.g. sound, resembling the input signal is achieved by matching the number of moving elements in an extreme position e.g. a top position as described herein, to the digital sample value, typically after resampling and scaling as described in detail below. For example, if the digital sample value is currently 10, 10 moving elements termed herein MEl, .... MElO may be in their top positions. If the digital sample value then changes to 13, three additional moving elements termed herein MEI l, ME 12 and ME 13 may be raised to their top position to reflect this. If the next sample value is still 13, no moving elements need be put into motion to reflect this. If the digital sample value then changes to 16, 3 different moving elements (since MEI l, ME 12 and ME 13 are already in their top positions), termed herein Ml 4, Ml 5 and M 16, may be raised to their top positions to reflect this.

In some embodiments, as described in detail below, moving elements are constructed and operative to be operated collectively in groups, such as a set of groups whose number of moving elements are all sequential powers of two, such as 31 moving elements constructed to be operated in groups having 1, 2, 4, 8, 16 moving elements, respectively, each. In this case, and using the above example, when the sample value is,

say, 10, the two groups including 8 and 2 moving elements respectively are both, say, up i.e. all moving elements in them are in their top positions. When the sample value changes to 13, however, it is typically impractical to directly shift 3 moving elements from their bottom positions to their top positions since in this example, due to the binary grouping, this can only be done by raising the two groups including 1 and 2 moving elements respectively, however, the group including 2 moving elements is already raised. But the number of top pixels may be otherwise matched to the sample value, 13 : Since 13 = 8 + 4 + 1, the two groups including 4 and 1 pixels may be raised, and the group including 2 pixels may be lowered, generating a net pressure change of +3, thereby to generate a sound resembling the input signal as desired, typically after resampling and scaling.

More generally, moving elements translated toward a first extreme position such as upward generate pressure in a first direction termed herein positive pressure. Moving elements translated toward the opposite extreme position such as downward generate pressure in the opposite direction termed herein negative pressure. A certain amount of positive or negative pressure may be obtained either by translating the appropriate number of moving elements in the corresponding direction, or by translating n moving elements in the corresponding direction and others, m in number, in the opposite direction, such that the difference n-m corresponds to e.g. equals the sampled signal value, typically after re-sampling and scaling.

The moving elements are typically formed of a material which is at least moderately electrically conductive such as silicon or silicon coated by a metal such as gold.

If the moving elements comprise permanent magnets, the permanent magnets are typically magnetized during production such that the magnetic poles are co-linear to the desired axes of motion. A coil that typically surrounds the entire transducer array generates the actuation force. To control each moving element, two latch elements (typically comprising electro static latches or "electrodes") are typically used, e.g. one above and one below the moving elements. According to one embodiment, the actuator is a speaker and the array of moving elements 10 is disposed within a fluid medium. The controllers 30 and 50 are then operative to define at least one attribute of the sound to correspond to at least one characteristic of the digital input signal. The sound has at least one wavelength thereby

to define a shortest wavelength present in the sound and each moving element 10 typically defines a cross section which is perpendicular to the moving element's axis and which defines a largest dimension thereof, the largest dimension of each cross-section typically being small relative to, e.g. an order of magnitude smaller than, the shortest wavelength. Fig. IB is an isometric illustration of the array 10 of moving elements constructed and operative in accordance with a preferred embodiment of the present invention. In this embodiment, each moving element 10 comprises a magnet and each is constrained to travel, except when and if latched, alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements 10 by the magnetic field generator 40.

Figs. 1C - IG are simplified top view illustrations of latch elements 72, 73, 74, 76, and 77, any of which may, in combination with similar or dissimilar others form the electrostatic latch 20 in accordance with alternative embodiments of the present invention. At least one of the latch elements, 72, may have a perforated configuration, as shown in Fig. 1C. In Fig. ID, a latch element 73 is shown having a notched configuration as to allow concentration of electrostatic charge at the sharp portions of the latch thereby to increase the latching force applied to the corresponding moving element. In Fig. IE, at least one latch element, 74, has a configuration which includes a central area 75 which prevents air from passing so as to retard escape of air, thereby to cushion contact between the moving element 10 and the latching element itself. At least one latch element, 76, may have a ring configuration, as shown in Fig. IF and, by way of example, in Fig. IB. Latch element 77 of Fig. IG is still another alternative embodiment which is similar to latch element 74 of Fig. IE except that at least one radial groove 78 is provided so as to eliminate induced current in the latch. Fig. 2 A shows the array of Fig. IB in a first, bottom extreme position responsive to an electromagnetic force applied, by coil or other magnetic field generator 40 of Fig. IA, downward. Fig. 2B shows the array of Fig. IB in a second, top extreme position responsive to an electromagnetic force applied, by coil or other magnetic field generator 40 of Fig. IA, upward. Fig. 2C is similar to Fig. 2B except that one of the individual moving magnets, 204, is not responding to the upward force applied by magnetic field generator 40 because that individual magnet is latched into its top extreme position by a corresponding electric charge disposed above the individual moving element and

functioning as a top latch. It is appreciated that in the embodiment of Figs. IA - 2C, the latch 20 comprises an electrostatic latch, however this need not be the case.

Typically, the apparatus of Figs. 2A - 2C comprises a pair of latch elements 205 and 207 for each moving element, termed herein "top" and "bottom" latch elements for simplicity although one need not be above the other, the latch elements including one or more electrodes and a space maintainer 220 separating the electrodes. In embodiments in which the latch 20 comprises an electrostatic latch, the space maintainer 220 may be formed of an insulating material.

Each pair of latching elements is operative to selectively latch its individual moving element 10 in a selectable one of two latching positions, termed herein the first and second latching positions or, for simplicity the "top" and "bottom" latching positions, thereby to prevent the individual moving elements from responding to the electromagnetic force. If the axis along which each moving element 10 moves is regarded as comprising a first half-axis and a second co-linear half-axis, the first latching position is typically disposed within the first half-axis and the second latching position is typically disposed within the second half-axis as shown e.g. in Figs. 2A - 2C.

Figs. 3 A - 3 C are respective top, cross-sectional and isometric views of a skewed array of moving elements 10 each constrained to travel alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements 10 e.g. by a coil 40 wrapped around the array as shown. Fig. 4A is an exploded view of a layered actuator device including an array of moving elements 403 each constrained to travel alternately back and forth along a respective axis in response to an alternating electromagnetic force applied to the array of moving elements 403 by a coil 401, and a latch, formed as at least one layer, operative to selectively latch at least one subset of the moving elements 403 in at least one latching position thereby to prevent the individual moving elements 403 from responding to the electromagnetic force. Typically, the electromagnetic force is generated using a coil 401 that surrounds the array 403 as shown. The latch typically comprises a pair of layers: a top latch layer 402 and bottom latch layer 404 which, when charged, and when the moving elements are in an appropriate electromagnetic field as described herein, latch the moving elements into top and bottom extreme positions respectively. Each of the latch layers 402 and 404

typically comprises an electrode layer and spacer layer as shown in detail in Figs. 5 - 6A. The spacer layers 402 and 404 may generally be formed from any suitable dielectric material. Optionally, ferrite or ferro-magnetic particles may be added to the dielectric material to decrease undesirable interaction between the magnets in the magnet layer. 5 In Figs. 5 - 6A, both flexures and annular magnets or conductors or ferromagnets are provided, however it is appreciated that this is not intended to be limiting. Alternatively, for example, other shaped magnets may be provided, or the annular elements may be replaced by coils, and free-floating moving elements may be provided without flexures, or the moving elements may have a peripheral elastic or

10 flexible portion or be associated with a peripheral elastic or flexible member, all as shown and described in detail herein.

Fig. 4B is a simplified flowchart illustration of a preferred actuation method operative in accordance with a preferred embodiment of the present invention. In Fig. 4B, a physical effect is generated, at least one attribute of which corresponds to at least

15 one characteristic of a digital input signal sampled periodically in accordance with a system clock signal. As shown, the method typically comprises (step 450) providing at least one array of moving elements 10 each constrained to travel alternately back and forth along an axis 15 (Fig. IB) in response to an alternating electromagnetic force applied to the array of moving elements 10 e.g. by magnetic field generator 40. In step

20 460, at least one subset of the moving elements 10 is selectively latched in at least one latching position by a latch 20 thereby to prevent the individual moving elements 10 from responding to the electromagnetic force applied by magnetic field generator 40. In step 470, the system clock signal is received and, accordingly, application of the electromagnetic force to the array of moving elements is controlled. In step 480, the

25 digital input signal is received, and the latching step 460 is controlled accordingly. Typically, as described above, the latch 20 comprises a pair of layers, each layer comprising an array of electrostatic latch elements and at least one space maintainer layer separates the electrostatic latch layers and is formed of an insulating material. Typically, the latch and at least one space maintainer are manufactured using PCB

30 production technology (Fig. 4B, step 450). The array of moving elements typically comprises a magnetic layer 403 sandwiched between a pair of electrode layers spaced from the magnetic layer by a pair of dielectric spacer layers. Typically, at least one of the layers is manufactured using wafer bonding technology, layer laminating

technology, and/or PCB production technology and/or combination of these technologies (Fig. 4B, step 455).

Fig. 5 is an isometric static view of the actuator device of Fig. 4A constructed and operative in accordance with a preferred embodiment of the present invention in which the array of moving elements 10 is formed of thin foil, each moving element being constrained by integrally formed flexures 606 surrounding it. The flexures typically include foil portions 703 interspersed with cut-out portions 702. Fig. 6A is an exploded view of a portion of the actuator device of Fig. 5.

According to a preferred embodiment of the present invention, 3 flexures are provided since at least three flexures are required to define a plane. In the case of the moving elements shown and described herein, the plane defined by the flexures is typically a plane perpendicular to the desired axes of motion of the moving elements or any plane suitably selected to constrain the moving elements to travel along the desired axes. Generally, it is desired to minimize the area of the flexures so as to exploit the available area of the device for the moving elements themselves since the process of actuation is performed by the moving elements such that, from the point of view of the functionality of the device, the area of the flexures is overhead. For example, if the actuator is a speaker, the moving elements push air thereby to create sound whereas the flexures and the gaps defining them do not. Therefore, it is generally desirable that the total length of the flexures be similar to the perimeter of the moving elements (e.g. as opposed to being double the perimeter of the moving elements). Therefore, it may be desired to treat the total length of the flexures as given and consequently, the more flexures provided, the shorter each flexure which translates into higher stress under the same translation i.e. to achieve the same amplitude of motion of the moving elements.

As a result, it is believed to be preferable to provide only three flexures i.e. no more than the minimum number of flexures required to securely hold the moving element, e.g. to define a plane normal to the axis of motion of the moving elements.

Figs. 6B and 6C are isometric and exploded view illustrations, respectively, of an assembly of moving elements, latches and spacer elements constructed and operative in accordance with a preferred, low air leakage, embodiment of the present invention. Air leakage refers to air passing from the space above the moving element to the space below the moving element or vice versa.

Fig. 6D is a cross-sectional view of the apparatus of Figs. 6B - 6C showing three moving elements 10 in top extreme, bottom extreme and intermediate positions 610, 620 and 630 respectively. Fig. 6E is a legend for Fig. 6D. Typically, in the embodiment of Figs. 6B - 6E, at least one of the moving elements is configured to prevent leakage of air through the at least one flexure. As shown, at least one space maintainer 640 is disposed between the array of moving elements 10 and the latching mechanism 20, the space maintainer defining a cylinder 660 having a cross section, and wherein at least one of the moving elements 10 comprises an elongate element 670 whose cross-section is small enough to avoid the flexures and a head element 680 mounted thereupon whose cross-section is similar to the cross-section of the cylinder 660. It is appreciated that for simplicity, only a portion of flexures 606 are shown.

Fig. 7A is a static partial top view illustration of the moving element layer of Figs. 5 - 6C. Fig. 7B is a cross-sectional view of the moving element layer of Figs. 5 - 6 taken along the A-A axis shown in Fig. 7A. Fig. 7C is a perspective view of the moving element layer of Figs. 5 - 7B wherein an individual moving element is shown moving upward toward its top extreme position such that its flexures bend and extend upward out of the plane of the thin foil. As shown, in Figs. 7A - 7C, at least one of the moving elements 10 of Fig. IA has a cross section defining a periphery 706 and is restrained by at least one flexure attached to the periphery. Typically, at least one moving element 10 and its restraining typically serpentine flexures are formed from a single sheet of material. Alternatively, as shown in Fig. 16B, at least one flexure 1605 may be formed of an elastic material. It is appreciated that the flexure-based embodiment is only one possible embodiment of the present invention. In contrast, as shown e.g. in Fig. IB, each moving element may simply comprise a free floating element.

Fig. 7D is a perspective view of a moving element layer constructed and operative in accordance with an alternative embodiment of the present invention. Fig. 7E is a side view illustration of the flexure-restrained central portion 705 of an individual moving element. In the embodiment of Figs. 7D - 7E, the moving elements 10 of Fig. IA comprise typically annular permanent magnets 710 rather than the discshaped permanent magnets 502 of the embodiments of Figs. 5 - 7C. Typically, each moving element 10 has first and second opposing typically circular surfaces 711 and

712 facing first and second endpoints 713 and 714 of the moving element's axis 715 of

motion respectively and at least one permanent magnet 710 is disposed on at least one of the first and second circular surfaces 71 1 and 712. If two permanent magnets 710 are provided, then the two are aligned such that the same pole points in the same direction as shown in Fig. 7E. Fig. 8 A is a control diagram illustrating control of latch 20 by latch controller 50 of Fig. IA, and of the typically coil-induced electromagnetic force, by controller 30 of Fig. IA, for a particular example in which the moving elements 10 are arranged in groups Gl, G2, ... GN that can each, selectably, be actuated collectively, wherein each latch in the latching layer is typically associated with a permanent magnet, and wherein the poles of all of the permanent magnets in the latching layer are all identically disposed. The latch typically comprises, for each group or each moving element in each group, a top latch and a bottom latch. The top and bottom latches for group Gk (k = 1 , ..., N) are termed Tk and Bk respectively. In Fig. 8 A the two controllers are both implemented in processor 802. Fig. 8B is a flowchart illustrating a preferred method whereby latching controller 50 of Fig. IA may process an incoming input signal 801 and control latches 20 of moving elements 10 accordingly, in groups. The abbreviation "EM" indicates electromagnetic force applied upward or downward, depending on the direction of the associated arrow, to a relevant group of moving elements. In the embodiment illustrated in Fig. 8B, if at time t, the LSB of the re-scaled PCM signal is 1 (step 816), this indicates that the speaker elements in group Gl may be in the selected end-position. If (step 817) group Gl is already in the selected end-position, no further action is required, however if the group Gl is not yet in the selected end-position, the latching controller 50 waits (step 818) for the electromagnetic field to be upward and then (step 819) releases the bottom latches in set Bl and engages the top latches in set Tl. This is also the case, mutatis mutandis, for all other groups G2, ... GN.

In Fig. 8B, the notation Tk or Bk followed by an upward pointing or downward pointing arrow indicates latching or releasing (upward or downward arrow respectively) of the top or bottom (T or B respectively) latch of the k'th group of moving elements. Fig. 8C is a simplified functional block diagram illustration of a processor, such as the processor 802 of Fig. 8A, which is useful in controlling substantially any of the actuator devices with electro-static latch mechanisms shown and described herein. A single processor, in the embodiment of Fig. 8C, implements both electromagnetic field

controller 30 and latch controller 50. The electromagnetic field controller 30 typically receives the system clock 805 which is typically a square wave and generates a sine wave with the same frequency and phase, providing this to the coil 40 as an actuating signal. The DSP 810 may for example comprise a suitably programmed TI 6000 digital signal processor commercially available from Texas Instruments. The program for the DSP 810 may reside in a suitable memory chip 820 such as a flash memory. The latch controller 50, in at least one mode of latch control operation, is operative to set the number of moving elements which oscillate freely responsive to the electromagnetic force applied by the coil 40 to be substantially proportional to the intensity sound coded in the digital input signal.

The electromagnetic field controller 30 typically controls an alternating current flow to the coil 40 that typically surrounds the entire array of moving elements 10, thus creating and controlling the magnetic field across the entire array. In certain embodiments a power amplifier 811 may be used to boost current to the coil 40. The electromagnetic field controller 30 typically generates an alternating electromagnetic force whose alternation is synchronous with the system clock 805 as described in detail below with reference to Fig. 1 IA, graph I.

The latch controller 50 is operative to receive the digital input signal 801 and to control the latching mechanism 20 accordingly. Typically, each individual moving element 10 performs at most one transition per clock i.e. during one given clock, each moving element may move from its bottom position to its top position, or move from its top position to its bottom position, or remain at one of either of those two positions. A preferred mode of operation of the latch controller 50 is described below with reference to Fig. 1 IA. According to a preferred embodiment of the present invention, retention of moving elements 10 in their appropriate end positions is affected by the latching controller 50.

Preferably, the latching controller 50 operates on the moving elements in groups, termed herein "controlled groups". All moving elements in any given group of moving elements are selectably either latched into their top positions, or into their bottom positions, or are unlatched. Preferably, the "controlled groups" form a sequence Gl, G2, ... and the number of speaker elements in each controlled group Gk is an integer, such as 2, to the power of (k -1), thereby allowing any desired number of speaker elements to be operated upon (latched upward, downward or not at all) since any given number can

be expressed as a sum of powers of, for example, two or ten or another suitable integer. If the total number of speaker elements is selected to be one less than an integral power (N) of 2 such as 2047, it is possible to partition the total population of speaker elements into an integral number of controlled groups namely N. For example, if there are 2047 speaker elements, the number of controlled groups in the sequence Gl, G2, ... is 11.

In this embodiment, since any individual value of the re-scaled PCM signal can be represented as a sum of integral powers of 2, a suitable number of speaker elements can always be placed in the selected end-position by collectively bringing all members of suitable controlled groups into that end-position. For example, if at time t the value of the re-scaled PCM signal is 100, then since 100 = 64 + 32 + 4, groups G3, G6 and G7 together include exactly 100 speaker elements and therefore, at the time t, all members of these three groups are collectively brought to the selected end position such as the "up" or "top" position and, at the same time, all members of all groups other than these three groups are collectively brought to the un-selected end position such as the "down" or "bottom" position. It is appreciated that each moving element has bottom and top latches, each typically generated by selectively applying suitable local electrostatic forces, associated therewith to latch it into its "down" and "up" positions respectively. The set of bottom and top latches of the speaker elements in group Gk are termed Bk and Tk latches respectively. Fig. 8D is a simplified flowchart illustration of a preferred method for initializing the apparatus of Figs. IA - 8C. According to the method of Fig. 8D, the array of moving elements 10 is put into initial motion including bringing each moving element 10 in the array of moving elements into at least one latching position. As described herein, both top and bottom latching positions are typically provided for each moving element 10 in which case the step of bringing each moving element in the array into at least one latching position typically comprises bringing a first subset of the moving elements in the array into their top latching positions and a second subset, comprising all remaining elements in the array, into their bottom latching positions. The first and second subsets are preferably selected such that when the moving elements in the first and second subsets are in their top and bottom latching positions respectively, the total pressure produced by fluid such as air displaced by the moving elements 10 in the first subset is equal in magnitude and opposite in direction to the total pressure produced by fluid such as air displaced by the moving elements in the second subset.

The moving elements 10 typically bear a charge having a predetermined polarity and each of the moving elements defines an individual natural resonance frequency which tends to differ slightly from that of other moving elements due to production tolerances, thereby to define a natural resonance frequency range, such as 42 - 46 KHz, for the array of moving elements. As described herein, typically, first and second electrostatic latching elements are provided which are operative to latch the moving elements 10 into the top and bottom latching positions respectively and the step of putting the array of moving elements into motion comprises:

Step 850: Charge the first (top or bottom) electrostatic latch of each moving element included in the first subset with a polarity opposite to the pole, on the moving element, facing that latch. The first and second subsets may each comprise 50% of the total number of moving elements.

Step 855: Charge the second (bottom or top) electrostatic latch of each moving element included in the second subset with a polarity opposite to the pole, on the moving element, facing that latch.

Step 860: As described above, the moving elements are designed to have a certain natural resonance frequency, f _r . Design tools may include computer aided modeling tools such as finite elements analysis (FEA) software. In step 860, fcuc, the frequency of the system clock, which determines the timing of the alternation of the electromagnetic field in which the moving elements are disposed, is set to the natural resonance frequency of the moving element in the array which has the lowest natural resonance frequency, referred to as f _m j _n and typically determined experimentally or by computer-aided modeling.

Steps 865 - 870: The system clock frequency may then be monotonically increased, from an initial value of f _m i _n to subsequent frequency values separated by δf until the system clock frequency has reached the natural resonance frequency of the moving element in the array which has the highest natural resonance frequency, referred to as f _max and typically determined experimentally or by computer-aided modeling. It is appreciated however that alternatively the system clock frequency might be monotonically decreased, from f _max to f _m i _n , or might be varied non-monotonically.

It is appreciated that when a moving element 10 is excited at its natural resonance frequency, f _r , the moving element increases its amplitude with every cycle, until reaching a certain maximal amplitude termed hereinafter A _max . Typically, the

duration δt required for the moving element to reach A _max is recorded during set-up and the magnetic force applied during the initialization sequence is selected to be such that A _max is twice as large as the gap the moving element needs to travel from its idle state to either the top or bottom latch. The Q factor or quality factor is a known factor which compares the time constant for decay of an oscillating physical system's amplitude to its oscillation period. Equivalently, it compares the frequency at which a system oscillates to the rate at which it dissipates its energy. A higher Q indicates a lower rate of energy dissipation relative to the oscillation frequency. Preferably, the Q factor of the moving elements is determined either computationally or experimentally. The Q factor as determined describes how far removed the frequency fc _LK needs to be from f _r (two possible values, one below f _r and one above f _r ) before the amplitude drops to 50% of A _max . The difference between the two possible values is δf.

As a result of the above steps, a sequence of electromagnetic forces of alternating polarities is now applied to the array of moving elements. The time interval between consecutive applications of force of the same polarity varies over time due to changes induced in the system clock, thereby to define a changing frequency level for the sequence. This results in an increase, at any time t, of the amplitude of oscillation of all moving elements whose individual natural resonance frequency is sufficiently similar to the frequency level at time t. The frequency level varies sufficiently slowly (i.e. only after a suitable interval δt, which may or may not be equal in all iterations) to enable the set S, of all moving elements whose natural resonance frequency is similar to the current frequency level, to be latched before the electromagnetic field alternation frequency level becomes so dissimilar to their natural resonance frequency as to cease increasing the amplitude of oscillation of the set S of moving elements. The extent of variation of the frequency level corresponds to the natural resonance frequency range. Typically, at the end of the initiation sequence (step 872), the system clock fc _LK is set to the predefined system frequency, typically being the average or median natural resonance frequency of the moving elements in the array, i.e. 44KHz. One method for determining the range of the natural resonance frequencies of the moving elements is to examine the array of moving elements using a vibrometer and excite the array at different frequencies.

Fig. 8E is a simplified isometric view illustration of an assembled speaker system constructed and operative in accordance with a preferred embodiment of the present invention. Mounted on a PCB 2100 is the array of actuator elements including moving elements 10 (not shown) sandwiched between latching elements 20. The array is surrounded by coil 40. Control lines 2110 are shown over which the latch control signals generated by latch controller 50 (not shown) in processor 802 travel to the latch elements 20. Amplifier 811 amplifies signals provided by the magnetic field generation controller 30 (not shown) in processor 802 to the coil 40. A connector 2120 connects the apparatus of Fig. 8E to a digital sound source. For simplicity, conventional components such as power supply components are not shown.

A preferred method of operation for generating a sound using apparatus constructed and operative in accordance with an embodiment of the present invention is now described based on Fig. 8F. The method of Fig. 8F is preferably based on the sound's representation in the time domain, typically a PCM (pulse-code modulation) representation.

Resampler 814 of Fig. 8F: Unless the sampling rate of the PCM happens to be the same as the system clock, the PCM is resampled to bring its sampling rate up to or down to the system clock frequency (top row in Fig. 1 IA) of the apparatus of Fig. IA. Generally, any suitable sampling rate may be employed. Specifically, the system of the present invention generates sound waves having at least two different frequencies, one of which is the desired frequency as determined by the input signal and the other of which is an artifact. The artifact frequency is the clock frequency i.e. the sampling rate of the system. Therefore, preferably, the system sampling rate is selected to be outside of the human hearing range i.e. at least 20 KHz. Nyquist sampling theory teaches that the system clock must be selected to be at least double that of the highest frequency the speaker is designed to produce.

Sealer 815: The PCM word length is typically 8, 16 or 24 bits. 8 bit PCM representations are unsigned, with amplitude values varying over time from 0 to 255, and 16 and 24 bit PCM representations are signed, with amplitude values varying over time from -32768 to 32767 and -8388608 to 8388607 respectively. The speaker of Figs. 1 - 2C typically employs an unsigned PCM signal and therefore, if the PCM signal is signed e.g. if the PCM word length is 16 or 24 bits, a suitable bias is added to obtain a corresponding unsigned signal. If the PCM word length is 16 bits, a bias of 32768

amplitude units is added to obtain a new range of 0 - 65535 amplitude units. If the PCM word length is 24 bits, a bias of 8388608 amplitude units is added to obtain a new range of 0 - 16777215 amplitude units.

The PCM signal is then further re-scaled as necessary such that its range, in amplitude units, is equal to the number of speaker elements in the apparatus of Figs. 1 - 2C. For example, if the number of speaker elements is 2047, and the PCM signal is an 8 bit signal, the signal is multiplied by a factor of 2048/256 = 8. Or, if the number of speaker elements is 2047, and the PCM signal is a 16 bit signal, the signal is multiplied by a factor of 2048/65536 = 1/32. Sound is then generated to represent the re-scaled PCM signal by actuating a suitable number of speaker elements in accordance with the current value of the re- scaled PCM signal. It is appreciated that the speaker elements have two possible end- states, termed herein the "down" and "up" end-states respectively, and illustrated schematically in Figs. 2A and 2B respectively. An individual one of these end-states is selected and the number of speaker elements in that end-state at any given time matches the current value of the re-scaled PCM signal, the remaining speaker elements at the same time being in the opposite end-state. For example, if there are 2047 speaker elements, the selected end-state is "up" and the value of the re-scaled PCM signal at time t is 100, the number of speaker elements in the "up" and "down" end states at time t are 100 and 1947 respectively. According to certain embodiments of the invention, there is no importance to the particular speaker elements selected to be in the "up" state as long as their total number corresponds to the current value of the re-scaled PCM signal.

The following loop is then performed M times each time a sample is generated by sealer 815. M is the number of actuator elements in the apparatus of Fig. IA. i is the index of the current loop. V _t is used to designate the current sample value exiting sealer 815 (for which M iterations of the loop are being performed). Generally, the number of moving elements to be latched into their top positions, is exactly equal to the value of V _t and all remaining moving elements are to be latched into their bottom positions. Therefore, while i is still smaller than V _t , the i'th moving element or pixel, termed in Fig. 8F "Pi" is latched to its top position. This is done by checking (Fig. 8F, step 840) whether, when moving element i was processed in the previous loop (t-1), it was in its top latching position or in its bottom latching position. If the former was the case,

nothing needs to be done and the method jumps to incrementation step 842. If the latter was the case, element i is marked as an element which needs to be latched into its top position (step 839). To latch all remaining moving elements into their bottom positions, do the following for all moving elements whose index exceeds V ₁ : check (step 838) which are already in their bottom positions; these moving elements need no further treatment. All others are marked (step 841) as elements which need to be latched into their bottom positions. Once all M elements have been marked or not marked as above, perform the following:

Verify that the magnetic field points upward, or wait for this (step 843), and, for the V ₁ or less pixels which are to be raised, discharge the bottom latches and charge the top latches (step 844). Next, wait for the magnetic field to point downward (step 845), and, for the (M- V _t ) or less pixels which are to be lowered, discharge the top latches and charge the bottom latches (step 846). At this point, the flow waits for the next sample to be produced by sealer 815 and then begins the M iterations of the loop just described for that sample.

It is appreciated that steps preceding step 843 are preferably executed during the half clock cycle in which the magnetic field polarity is downwards. Step 844 is preferably executed at the moment the magnetic field changes its polarity from downwards to upwards. Similarly, step 846 is preferably executed at the moment the magnetic field changes polarity again from upwards to downwards. It is also appreciated that in order for the device to remain synchronized with the digitized input signal, steps 814 - 846 are all preferably executed in less than one clock cycle.

Fig. 9A is a graph summarizing the various forces brought to bear on moving elements 10 in accordance with a preferred embodiment of the present invention. Fig. 9B is a simplified pictorial illustration of a magnetic field gradient inducing layer constructed and operative in accordance with a preferred embodiment of the present invention and comprising at least one winding conductive element 2600 embedded in a dielectric substrate 2605 and typically configured to wind between an array of channels 2610. Typically, there are no channels 2610 along the perimeter of the conductive layer of Fig. 9B so that the gradient induced within channels adjacent the perimeter is substantially the same as the gradient induced in channels adjacent the center of the conductive layer.

If the layer of Fig. 9B is separate from the spacer layers described above, then the channels in the layer of Fig. 9B are disposed opposite and as a continuation of the channels in the spacer layers described in detail above. The cross-sectional dimensions, e.g. diameters, of channels 2610 may be different than the diameters of the channels in the spacer layer. Alternatively, the layer of Fig. 9B may serve both as a spacer layer and as a magnetic field inducing layer in which case the channels 2610 of Fig. 9B are exactly the spacer layer channels described hereinabove. It is appreciated that, for simplicity, the electrodes forming part of the spacer layer are not shown in Fig. 9B.

Figs. 9C and 9D illustrate the magnetic field gradient induction function of the conductive layer of Fig. 9B. In Fig. 9C, the current flowing through the winding element 2600 is indicated by arrows 2620. The direction of the resulting magnetic field is indicated by X's 2630 and encircled dots 2640 in Fig. 9C, indicating locations at which the resulting magnetic field points into and out of the page, respectively.

Fig. 1OA is a simplified top cross-sectional illustration of a latching layer included in latch 20 of Fig. IA in accordance with a preferred embodiment of the present invention. The latching layer of Fig. 1OA is suitable for latching moving elements partitioned into several groups Gl, G2, ... whose latches are electrically interconnected as shown so as to allow collective actuation of the latches. This embodiment is typically characterized in that any number of moving elements may be actuated by collectively charging the latches of selected groups from among the partitioned groups, each latch in the latching layer typically being associated with a permanent magnet, wherein the poles of all of the permanent magnets in the latching layer are all identically disposed. Each group Gk may comprise 2 to the power of (k-1) moving elements. The groups of moving elements may spiral out from the center of the array of moving elements, smallest groups being closest to the center as shown.

Fig. 1OB is a simplified electronic diagram of an alternative embodiment of the latching layer of Fig. 1OA in which each latch is individually rather than collectively controlled (i.e. charged) by the latching controller 50 of Fig. IA. It is appreciated that the latches are shown to be annular, however alternatively they may have any other suitable configuration e.g. as described herein. The layer of Fig. 1OB comprises a grid of vertical and horizontal wires defining junctions. A gate such as a bi-polar field-effect transistor is typically provided at each junction. To open an individual gate thereby to

charge the corresponding latch, suitable voltages are provided along the corresponding vertical and horizontal wires.

Fig. HA is a timing diagram showing a preferred charging control scheme which may be used by the latch controller 50 in Fig. IA in uni-directional speaker applications wherein an input signal representing a desired sound is received, and moving elements 10 constructed and operative in accordance with a preferred embodiment of the present invention are controlled responsively, by appropriate charging of their respective latches, so as to obtain a sound pattern in which the volume in front of the speaker is greater than in other areas, each latch in the latching layer being associated with a permanent magnet, and the poles of all of the permanent magnets in the latching layer all being identically disposed. Fig. HB is a schematic illustration of an example array of moving elements 10 to which the timing diagram of Fig. HA pertains.

A preferred mode of operation of the latch controller 50 is now described with reference to Figs. 11A - B. For clarity, the preferred mode of operation is described merely by way of example with reference to a speaker comprising 7 pixels numbered Pl, P2, ...P7 as shown in Fig. 1 IB. Further according to the example used to explain the preferred mode of operation of latch controller 50, the 7 pixels are actuated in three groups comprising 1, 2 and 4 pixels respectively. Generally, the latch controller 50 uses various decision parameters, as described in detail herein, to determine how to control each individual moving element in each time interval. Speakers constructed and operative in accordance with a preferred embodiment of the present invention are typically operative to reproduce a sound which is represented by the analog signal of Graph II and is then digitized and supplied to a speaker of the present invention. The values of the digital signal are shown in Fig. 1 IA, graph III.

Graph IV shows the alternation of the electromagnetic force applied to the moving elements 10 by the coil or other magnetic field generator 40. Graph V is the signal provided by latching controller 50 to the top latch of an individual moving element, Pl seen in Fig. 1 IB, forming, on its own, a first group Gl of moving elements consisting only of Pl. Graph VI is the signal provided by latching controller 50 to the bottom latch of Pl. The states of Pl, due to the operation of the latches associated therewith, are shown in Graph VII, in which black indicates the top extreme position in

which the top latch engages Pl, white indicates the bottom extreme position in which the bottom latch engages Pl, and hatching indicates intermediate positions.

Graph VIII is the signal provided by latching controller 50 to the top latch/es of each of, or both of, moving elements P2 and P3 seen in Fig. 1 IB, which together form a second group GII of moving elements. Graph IX is the signal provided by latching controller 50 to the bottom latch/es of GII. The states of P2 and P3, due to the operation of the latches associated therewith, are shown in Graphs X and XI respectively, in which black indicates the top extreme position in which the top latch engages the relevant moving element, white indicates the bottom extreme position in which the bottom latch engages the relevant moving element, and hatching indicates intermediate positions of the relevant moving element.

Graph XII is the signal provided by latching controller 50 to the top latch/es of each of, or all of, moving elements P4 - P7 seen in Fig. HB, which together form a third group GUI of moving elements. Graph XIII is the signal provided by latching controller 50 to the bottom latch/es of GUI. The states of P4 - P7, due to the operation of the latches associated therewith, are shown in Graphs XIV - XVII respectively, in which black indicates the top extreme position in which the top latch engages the relevant moving element, white indicates the bottom extreme position in which the bottom latch engages the relevant moving element, and hatching indicates intermediate positions of the relevant moving element.

Graph XVIII schematically illustrates the moving elements Pl - P7 of Fig. 1 IB in their various positions, as a function of time.

For example, in interval 15, the clock is high (graph I), the digitized sample value is 2 (graph III), which indicates that 5 elements need to be in their top positions and 2 elements in their bottom positions as shown in interval 15 of Graph XVIII. Since latch actuation in this embodiment is collective, this is achieved by selecting groups Gl and G3 which together have 5 elements (1 + 4) to be in their top positions whereas the two moving elements in G2 will be in their bottom positions. The magnetic field points upward in interval 15 as shown in Graph IV. In interval 14, the moving element in Gl was in its bottom position as shown in Graph XVIII and therefore needs to be raised. To do so, control signal Bl is lowered (graph VI) and control signal Tl is raised (graph V). As a result, the moving element of Gl assumes its top position as shown in graph VII.

In interval 14, the moving elements in G2 are already in their bottom positions as shown

in Graph XVIII and therefore the top control signal T2 remains low as seen in graph VIII, the bottom control signal B2 remains high as seen in graph IX and consequently, as shown in Graphs X and XI respectively, the two moving elements (P2 and P3) in G2, remain in their bottom extreme positions. As for group G3, in interval 14, the moving elements in G3 are already in their top positions as shown in Graph XVIII and therefore the top control signal T3 remains high as seen in graph XII, the bottom control signal B3 remains low as seen in graph XIII and consequently, as shown in Graphs XIV - XVII respectively, the four moving elements (P4 - P7) in G3, remain in their top extreme positions. Preferably, when the input signal in graph II is at a positive local maximum, all moving elements are in their top position. When the input signal is at a negative local maximum, all moving elements are in their bottom position.

Fig. HC is a timing diagram showing a preferred control scheme used by the latch controller 50 in omni-directional speaker applications wherein an input signal representing a desired sound is received, and moving elements constructed and operative in accordance with a preferred embodiment of the present invention are controlled responsively, so as to obtain a sound pattern in which the loudness of the sound in an area located at a certain distance in front of the speaker is similar to the loudness in all other areas surrounding the speaker at the same distance from the speaker.

As shown, the step of selectively latching comprises latching specific moving elements at a time determined by the distance of the specific moving elements from the center of the array (e.g. as indicated by r in the circular array of Fig. HB). Typically, when it is desired to latch a particular subset of moving elements, typically corresponding in number to the intensity of a desired sound, the moving elements are latched not simultaneously but rather sequentially, wherein moving elements closest to the center are latched first, followed by those moving elements disposed, typically in layers, concentrically outward from the center. Typically, the moving elements in each layer are actuated simultaneously. Typically, the temporal distance δt between the moment at which a particular moving element is latched and between the moment at which the first, central, moving element or elements was or were latched is r/c where c is the speed of sound.

It is appreciated that the moving elements in graph X of Fig 11C are shown to comprise flexible peripheral portions, however this is merely by way of example and is not intended to be limiting.

Figs. 12A and 12B are respectively simplified top view and cross-sectional view illustrations of the moving element layer in accordance with a preferred embodiment of the present invention in which half of the permanent magnets are placed north pole upward and half north pole downward. A particular advantage of this embodiment is that moving elements can be raised both when the electromagnetic field points upward and when it points downward rather than waiting for the field to point upward before lifting a moving element and waiting for the field to point downward before lowering a moving element. Although the illustrated embodiment shows the two subsets separated from one another, this not need be the case. The two subsets may be interleaved with one another.

Fig. 13 is a simplified top view illustration similar to Fig. 1OA except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward. Whereas in the embodiment of Fig. 1OA, there was one group each of size 1, 2, 4, ... (which may be arranged sequentially around the center as shown in Fig. 1OA although this need not be the case) in the embodiment of Fig. 13, there are two groups of each size, thereby generating two sequences of groups of size 1, 2, 4, .... In the illustrated embodiment the groups in the first sequence are termed GlL, G2L, G3L, .... and the groups in the first sequence are termed GlR, G2R, G3R, .... Each of these sequences is arranged within one semicircle, such as the left and right semicircles as shown. The arrangement of the groups within its semicircle need not be in order of size of the group extending concentrically outward as shown and can be any desired arrangement, however, preferably, both groups are arranged mutually symmetrically within their individual semicircle. It is appreciated that by using suitable coil designs, the same effect can be achieved using permanent magnets that are all polarized in the same direction while the coil generates magnetic fields having a certain polarization across half of the moving elements and having an opposite polarization across the other half.

A particular feature of the embodiments of Figs. 1OA and 13 is that latch elements corresponding to certain moving elements are electrically interconnected

thereby to form groups of moving elements which can be collectively latched or released by collectively charging or discharging, respectively, their electrically interconnected latches.

Fig. 14 is a control diagram illustrating control of the latches and of the coil- induced electromagnetic force for a particular example in which the moving elements are arranged in groups that can each, selectively, be actuated collectively, similar to Fig. 8A except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward as shown in Fig. 13, whereas in Fig. 8A, the poles of all of the permanent magnets in the latching layer are all identically disposed. As shown in Fig. 14, latching signals are provided to all of groups GlL, G2L, G3L, .... and GlR, G2R, G3R. The top latching signals for these groups are indicated as LTl, LT2, LT3, .... and RTl, RT2, RT3 respectively. The bottom latching signals for these groups are indicated as LBl, LB2, LB3, .... and RBl, RB2, RB3. Fig. 15A is a timing diagram showing a preferred control scheme used by the latch controller 50 in uni-directional speaker applications, which is similar to the timing diagram of Fig. HA except that half of the permanent magnets in the latching layer are disposed north pole upward and the remaining half of the permanent magnets in the latching layer are disposed north pole downward as shown in Fig. 13 whereas in Fig. 1 IA the poles of all of the permanent magnets in the latching layer are all identically disposed. Fig. 15B is a schematic illustration of an example array of moving elements to which the timing diagram of Fig. 15A pertains.

As described above, a particular advantage of the embodiment of Figs. 13 - 15A as opposed to the embodiment of Figs. 8 A, 1OA and HA is that moving elements can be raised both when the electromagnetic field points upward and when it points downward rather than waiting for the field to point upward before lifting a moving element and waiting for the field to point downward before lowering a moving element. It is appreciated that no elements move in 50% of the time slots in Fig. 1 IA which may introduce distortion of sound and is relatively inefficient. In contrast, elements move in 100% of the time slots in Fig. 15A (other than slots in which no motion is required since the digital signal value is unchanged) thereby preventing distortion and enhancing efficiency.

For example, in interval 15, the digitized signal value changes from 1 to 2 as shown in graph II of Figs. 1 IA and 15 A. Consequently, moving element Pl in Fig. 1 IA needs to be raised i.e. released from its current, bottom extreme position and latched into its top extreme position, however whereas in 15, control signal Bl is lowered and control signal Tl is raised, in interval 16 nothing happens. In Fig. 15 A, in contrast, where moving elements LPl (and RPl) need to be raised, control signal LBl is lowered and control signal LTl is raised in interval 15, and immediately afterward, in interval 16, the RBl control signal is lowered and the RTl signal is raised, resulting in upward motion of RPl without the delay incurred in Fig. 1 IA. Generally in the embodiment of Figs. 13 - 15 A, since half of the magnets (say, the left half) point north up and the remaining (right) half point north down, when it is desired to move elements 10 upward, this can always be done without delay. If the magnetic field points up, the moving elements in the left half of the array can be moved upward before those in the right half, whereas if by chance the magnetic field is found to be pointing down, the moving elements in the right half of the array can be moved upward before those in the left half.

Fig. 15C is a graph showing changes in the number of moving elements disposed in top and bottom extreme positions at different times and as a function of the frequency of the input signal received by the latching controller 50 of Fig. IA. Fig. 16A is an isometric view illustration of a moving element layer which is an alternative to the moving element layer shown in Figs. IA and 2 A - 2C in which the layer is formed from a thin foil such that each moving element comprises a central portion and surrounding portions.

Fig. 16B is an isometric view illustration of still another alternative to the moving element layer shown in Figs. IA and 2A - 2C in which the flexure structure at the periphery of each moving element comprises a sheet of flexible material e.g. rubber. The central area of each moving element comprises a magnet which may or may not be mounted on a rigid disc.

Fig. 16C is an isometric view of a preferred embodiment of the moving elements and surrounding flexures depicted in Figs. 7A - 7E or 16A in which flexures vary in thickness. In Fig. 16C, for simplicity, the component which causes the moving element 1620 to be affected by the magnetic field, which component may preferably comprise a magnet or alternatively, a ferro-magnet, conductive material or coil, is not shown. As

shown, the moving element 1620 comprises serpentine peripheral flexures 1630 having portions of varying thicknesses connecting a central portion 1640 of the moving element to a sheet 1650 interconnecting all or many moving elements. For example, the portions of varying thicknesses may include thicker portions 1660 and thinner portions 1670 respectively as shown. For example, if the diameter of the central portion 1640 of each moving element is 300 microns and the sheet is silicon, then under certain conditions, portions 1670 may be 50 microns thick whereas portions 1660 may be 100 microns thick. More generally, thicknesses are computed as a function of materials to provide application-specific flexibility and strength levels, e.g. using FEA (finite element analysis) tools.

Fig. 16D is an isometric illustration of a cost effective alternative to the apparatus of Fig. 16C in which flexures vary in width. As in Fig. 16C, for simplicity, the component which causes the moving element 1720 to be affected by the magnetic field, which component may preferably comprise a magnet or alternatively, a ferro- magnet, conductive material or coil, is not shown. As shown, the moving element 1720 comprises serpentine peripheral flexures 1730 having portions of varying widths connecting a central portion 1740 of the moving element to a sheet 1750 interconnecting all or many moving elements. For example, the portions of varying widths may include wider portions 1760 and narrower portions 1770 respectively as shown. For example, if the diameter of the central portion 1740 of each moving element is 300 microns and the sheet is silicon, then under certain conditions, portions 1770 may be 20 microns wide whereas portions 1760 may be 60 microns wide. More generally, widths are computed as a function of materials to provide application-specific flexibility and strength levels, e.g. using FEA (finite element analysis) tools. It is appreciated that the embodiments of Figs. 16C and 16D may be suitably combined, e.g. to provide flexures with varying thicknesses and varying widths, and/or varied, e.g. to provide flexures whose widths and/or thicknesses vary either continuously or discontinuously as shown, and either regularly as shown or irregularly.

In the above description, "thickness" is the dimension of the flexure in the direction of motion of the moving element whereas "width" is the dimension of the flexure in the direction perpendicular to the direction of motion of the moving element. A particular advantage of the embodiments of Figs. 16C and 16D is that in flexures of varying cross-sections, e.g. varying thicknesses or widths, the stress is not

concentrated at the roots 1680 or 1780 of the flexures and is instead distributed over all the thin and/or narrow portions of the flexures. Also, generally, the stress on the flexures as a result of bending thereof is a steep function of the thickness, typically a cubic function thereof, and is also a function of the width, typically a linear function thereof. It is believed to be impractical, at least for certain materials such as silicon and at least for certain applications employing large displacement of the moving elements, e.g. public address speakers, to select flexure dimensions which are uniformly thin enough or narrow enough to provide sufficiently low stress so as to prevent breaking, and simultaneously stiff enough to allow natural resonance frequency at a desirable range e.g. 44KHz. For this reason as well, it is believed to be advantageous to use flexures of varying thicknesses and/or widths e.g. as illustrated in Figs. 16C - 16D.

Fig. 17 is a top cross-sectional view illustration of an array of actuator elements similar to the array of Fig. 3A except that whereas in Fig. 3A, consecutive rows of individual moving elements or latches are respectively skewed so as to increase the number of actuator elements that can be packed into a given area, the rows in Fig. 17 are unskewed and typically comprise a rectangular array in which rows are mutually aligned.

Fig. 18 is an exploded view of an alternative embodiment of an array of actuator elements, including a layer 1810 of moving elements sandwiched between a top latching layer 1820 and a bottom latching layer 1830. The apparatus of Fig. 18 is characterized in that the cross-section of each actuator element is square rather than round. Each actuator element could also have any other cross-sectional shape such as a hexagon or triangle.

Fig. 19 is an isometric array of actuators supported within a support frame providing an active area which is the sum of the active areas of the individual actuator arrays. In other words, in Fig. 19, instead of a single one actuating device, a plurality of actuating devices is provided. The devices need not be identical and can each have different characteristics such as but not limited to different clock frequencies, different actuator element sizes and different displacements. The devices may or may not share components such as but not limited to coils 40 and/or magnetic field controllers 30 and/or latch controller 50.

The term "active area" refers to the sum of cross-sectional areas of all actuator elements in each array. It is appreciated that generally, the range of sound volume (or,

for a general actuator other than a speaker, the gain) which can be produced by a speaker constructed and operative in accordance with a preferred embodiment of the present invention is often limited by the active area. Furthermore, the resolution of sound volume which can be produced is proportional to the number of actuator elements provided, which again is often limited by the active area. Typically, there is a practical limit to the size of each actuator array e.g. if each actuator array resides on a wafer.

If the speaker is to serve as a headphone, only a relatively small range of sound volume need be provided. Home speakers typically require an intermediate sound volume range whereas public address speakers typically have a large sound volume range, e.g. their maximal volume may be 120 dB. Speaker applications also differ in the amount of physical space available for the speaker. Finally, the resolution of sound volume for a particular application is determined by the desired sound quality, e.g. cell phones typically do not require high sound quality, however space is limited.

According to certain embodiments of the present invention, layers of magnets on the moving elements may be magnetized so as to be polarized in directions other than the direction of movement of the element to achieve a maximum force along the electromagnetic field gradient aligned with the desired element moving direction.

Referring again to Figs. 12A- 15B inter alia, it is appreciated that if the coil used is of a design that utilizes conductors carrying current on both sides of the elements, and the magnets are all polarized in the same direction, then the elements on one side of each conductor would move in opposite directions when current flows in the coil.

A particular feature of a preferred embodiment of the present invention is that the stroke of motion performed by the moving elements is relatively long because the field applied thereto is magnetic hence decays at a rate which is inversely proportional to the distance between the moving elements and the current producing the magnetic field. In contrast, an electrostatic field decays at a rate which is inversely proportional to the square of the distance between the moving elements and the electric charge producing the electrostatic field. As a result of the long stroke achieved by the moving elements, the velocity achieved thereby is increased hence the loudness that can be achieved increases because the air pressure generated by the high velocity motion of the moving elements is increased.

It is appreciated that the embodiments specifically illustrated herein are not intended to be limiting e.g. in the sense that the moving elements need not all be the

same size, the groups of moving elements, or individual moving elements if actuated individually, need not operate at the same resonance nor with the same clock, and the moving elements need not have the same amplitude of displacement.

The speaker devices shown and described herein are typically operative to generate a sound whose intensity corresponds to intensity values coded into an input digital signal. Any suitable protocol may be employed to generate the input digital signal such as but not limited to PCM or PWM (SACD) protocols. Alternatively or in addition the device may support compressed digital protocols such as ADPCM, MP3,

AAC, or AC3 in which case a decoder typically coverts the compressed signal into an uncompressed form such as PCM.

Design of digital loudspeakers in accordance with any of the embodiments shown and described herein may be facilitated by application-specific computer modeling and simulations. Loudness computations may be performed conventionally, e.g. using fluid dynamic finite-element computer modeling and empiric experimentation.

Generally, as more speaker elements (moving elements) are provided, the dynamic range (difference between the loudest and softest volumes that can be produced) becomes wider, the distortion (the less the sound resembles the input signal) becomes smaller and the frequency range becomes wider. On the other hand, if less speaker elements are provided, the apparatus is smaller and less costly.

Generally, if the moving elements have large diameters, the ratio between active and inactive areas (the fill factor) improves, and there is less stress on the flexures if any, assuming that the vibration displacement remains the same, which translates into longer life expectancy for the equipment. On the other hand, if the moving elements have small diameters, more elements are provided per unit area, and due to the lesser mass, less current is required in the coil or other electromagnetic force generator, translating into lower power requirements.

Generally, if the vibration displacement of the moving elements is large, more volume is produced by an array of a given size, whereas if the same quantity is small, there is less stress on the flexures, if any, and the power requirements are lower.

Generally, if the sample rate is high, the highest producible frequency is high and the audible noise is reduced. On the other hand, if the sample rate is low, accelerations, forces, stress on flexures if any and power requirements are lower.

Three examples of application-specific speakers are now described.

Example 1 : It may be desired to manufacture a mobile phone speaker which is very small, is low cost, is loud enough to be heard ringing in the next room, but has only modest sound quality. The desired small size and cost suggest a speaker with relatively small area, such as up to 300mm ² . If a relatively high target maximal loudness such as

9OdB SPL is desired, this suggests large displacement. Acceptable distortion levels

(10%) and dynamic range (6OdB) in mobile phone speakers dictate a minimal array size of 1000 elements (computed using: M=10^ ^60/20 '). Therefore, a suitable speaker may comprise 1023 moving elements partitioned into 10 binary groups, each occupying an area of about 0.3mm ² . The cell size would therefore be about 550μm x 550μm.

For practical reasons, the largest moving element that fits this space may have a diameter of 450μm. Reasonable displacement for such a moving element may be about lOOμm PTP (peak to peak) which enables the target loudness to be achieved. The sample rate may be low, e.g. 32 KHz, since mobile phones sound is limited by the cellular channel to 4KHz.

Example 2: It may be desired to manufacture high fidelity headphones having very high sound quality (highest possible) and very low noise, and which are additionally small enough to be worn comfortably, and finally, cost-effective to the extent possible. To achieve high sound quality, wide dynamic range (at least 96dB), wide frequency range (20Hz-20KHz) and very low distortion (<0.1%) may be used. The minimal number of elements may be, given these assumptions, 63000. So, for example, the speaker may have 65535 elements divided into 16 binary groups. Maximal loudness can be kept low (8OdB) so as to allow displacements of about 30μm PTP. The smallest moving element capable of such displacements is about 150μm in diameter. Such an element may occupy a cell of 200μm x 200μm or 0.04mm ² such that 65535 elements fit into an area of 2621mm ² e.g. 52mm x 52mm. The sample rate is typically at least twice the highest frequency the speaker is meant to produce, or 40KHz. The closest standard sample rate is 44.1KHz. Example 3: It may be desired to manufacture a public address speaker, e.g. for a dance club, which is very loud, has a wide frequency range, extends to very low frequencies, and has low distortion. Therefore, PA speakers typically have many large moving elements. 600μm moving elements may be used, which are capable of

displacements of 150μm PTP. Such elements occupy cells of 750μm x 750μm or 0.5625mm ² . Due to the low frequency requirement, a minimum of 262143 moving elements, partitioned into 18 binary groups, may be used. The size of the speaker may be about 40cm x 40cm. This speaker typically reaches maximal loudness levels of 12OdB SPL and extends down to 15Hz.

Reference is now made to Figs. 2OA - 2OB which is are simplified generally self-explanatory functional block diagram illustration of preferred systems for achieving a desired directivity pattern for a desired sound stream using a direct digital speaker such as any of those shown herein in Figs. IA- 19 or such as a conventional direct digital speaker which may, for example comprise that shown and described in U.S. Patent 6,403,995 to David Thomas, assigned to Texas Instruments and issued 11 June 2002, or in Diamond Brett M., et al, "Digital sound reconstruction using array of CMOS-MEMS micro-speakers", Transducers '03, The 12 ^th International Conference on Solid State Sensors, Actuators and Microsystems, Boston, June 8 - 12, 2003. If the direct digital speaker of Fig. IA is used to achieve a desired directivity pattern for a desired sound stream, then typically, blocks 3020, 3030 and 3040 in Fig. 2OA comprise blocks 20, 30 and 40 of Fig. IA respectively and block 3050 comprises latch controller 50 of Fig. IA, programmed to implement the per-clock operation of block 3050 e.g. as shown and described herein with reference to Fig. 21. Fig. 21 is a simplified flowchart illustration of per-clock operation of the moving element constraint controller 3050 of Figs. 2OA -20B, in accordance with certain embodiments of the present invention.

Step 3100 determines how many moving elements should move during the current clock. Typically, and as described in detail above with reference to Figs. 1 - 19, the number of moving elements which are to move during a given clock is generally proportional to the intensity of the input signal during that clock, suitably normalized e.g. as described above with reference to resampler 814 and sealer 815 of Fig. 8B.

Step 3200 determines which moving elements should move during the current clock, using, in some embodiments, a suitable moving element selection LUT which is typically loaded into the memory of the constraint controller 3050 of Figs. 2OA - 2OB during factory set-up. Each such LUT is typically built for a specific moving element array taking into account, inter alia, the array size and whether or not the array is

skewed. Each directivity pattern which it is desired to achieve typically requires its own LUT.

Step 3300 determines the amount of delay with which to operate each of the moving elements of moving element array 3010 or 3012 of Figs. 2OA - 2OB. Step 3200 is now described in detail. A preferred method for performing step

3200 is now described. Step 3200 typically employs a LUT (look up table) which has cells which correspond one-to-one to the pressure producing elements in the array. For example, if the array comprises a rectangle of 100 x 200 pressure producing elements then the LUT may have 100 x 200 cells. Each cell holds a uniquely appearing integer between 1 and the total number of pressure producing elements such as 20000 in the illustrated example. Therefore, the LUT assigns an ordinal number to each pressure producing element in the array. Associated in memory with the LUT is a integer parameter P which stores an indication of the number of pressure producing elements currently in a first operative configuration from among two operative configurations, characterized in that transition of the pressure producing elements therebetween produces pressure in the medium, such as air, in which the apparatus of the invention is disposed. In some embodiments, pressure in opposite directions is obtained when the elements move from the first configuration to the second, as opposed to when the elements move from the second configuration to the first. In other embodiments, pressure is obtained as long as the elements are in the first configuration, and no pressure is obtained when the elements are in the second configuration.

Typically, P is initialized during set-up as described below, and is then assigned a current value in each clock by step 3100. In the immediately following step 3200 in the same clock, P pressure producing elements are brought to their first operative configuration and N - P pressure producing elements are brought to their second operative configuration where N is the number of pressure producing elements in the array. The P elements selected to be in their first operative configuration are those whose ordinal number as determined by the LUT is smaller than P. The N-P elements selected to be in their second operative configuration are those whose ordinal number as determined by the LUT is greater than or equal to P.

One of these configurations, say the first, is typically arbitrarily considered the "positive" configuration whereas the other configuration, say the second, is then considered the "negative" configuration. Alternatively, in some applications there may

be a physical reason to select a specific one of the configurations to be the positive configuration. The pressure generated when a pressure producing element moves from the second configuration to this first configuration is termed "positive pressure" whereas the pressure generated when a pressure producing element moves from the second configuration to this first configuration is termed "positive pressure". The pressure generated by a single transition from one configuration to the other is termed herein a pressure "pulse".

During set-up, the parameter P is typically given an initial value equal to half of the number of pressure producing elements in the array such as 10000 in the present example. The array is then initialized such that each pressure producing element whose ordinal number as determined by the LUT is less than P is brought to its first configuration and the remaining pressure producing elements are brought to their second configuration.

A suitable LUT (look up table), which has cells which correspond one-to-one to the N pressure producing elements in the array, storing integers from 1 to N, may be generated as follows:

A criterion for LUT quality is first determined, which may be application- specific. One suitable criterion for LUT quality is now described.

A list is prepared of all possible subsets of consecutive integers ranging between 1 and N. In the present example, the first subset, termed hereinafter S2i, includes 2 integers: 1 and 2; the second subset, S2 ₂ includes the integers 2 and 3, and so on for all subsets containing two integers. The last two-element subset, S2 ₁₉₉₉₉ , contains the integers 19999 and 20000. The list also includes all possible three element subsets, namely, to continue the example, S3i (which includes integers 1, 2, 3), S3 ₂ (which includes integers 2, 3, 4), ... S3 ₁₉₉₉₈ (which includes integers 19998, 19999, 20000). The list also includes all 4 element subsets, 5 element subsets and so on and so forth. The last subset, S20000i contains all 20000 elements. In general, a subset containing K integers, starting at i is labeled SKj. It is appreciated that for a LUT containing N cells, the number of possible subsets M equals M=(N- l)*N/2. For each subset SKj, a set of coordinates is defined (Xj, Yj), (Xi ₊₁ ,Yj ₊ i), ...(Xj _+K-

₁ , Yj _+K -i ) such that the coordinates represent the position of the pressure-producing elements whose ordinal numbers are i, i+1, ... i+k-1 according to the current LUT.

For each subset SKj, a propagation angle θKj is computed e.g. using analytic or numeric computation methods, typically using suitable computer simulation applications such as Matlab, MatCAD or Mathematica. The sound waves' propagation angles are computed for K coherent sound sources, disposed at positions (Xj, Yj), (Xj ₊ i,Yi ₊ i), ... (Xi _+K- i,Yi+κ-i), all producing sinusoidal waves at the same phase and at a frequency equal to the system sampling rate, e.g. 44100Hz.

A "propagation angle of a subset" is defined as follows: Each subset corresponds to a subset of pressure producing elements. A reference axis is defined passing through the center of mass of the array of pressure producing elements and perpendicular to its main surface. The intensity of sound generated by the subset of pressure producing elements approaches a maximum as one retreats from the array of pressure producing elements along the reference axis. Therefore, a maximal intensity for the subset may be defined by measuring the intensity at a location L which is on the reference axis and sufficiently distant from the array so as to ensure that the differences between the distance of location L and each of the pressure producing elements in the subset are sufficiently, e.g. an order of a magnitude, smaller than the wavelength λ associated with the system clock. At least one reference plane is defined which includes the reference axis. It is appreciated that an infinite number of such reference planes exists. For cylindrical propagation applications in which a focal axis is defined, select a reference plane which includes the focal axis. It is appreciated that a LUT constructed on this basis would typically also be suitable for omnidirectional applications. For propagation applications in which a focal point is defined as described herein, select a reference plane which includes the focal point. If more than one such reference plane exists, select two such reference planes which are mutually perpendicular. The propagation angle of the subset, termed herein θKj, is defined for each reference plane selected for that subset, as follows: Define an imaginary circle within the reference plane whose center is at the point of intersection between the reference axis and the main surface of the array and whose radius is the distance between L and the main surface of the array. Select two locations on the circumference of the circle on both sides of the reference axis respectively, in which the sound intensity generated by the subset of pressure producing elements is half of the maximal intensity measured at L. The angle defined between two radii connecting the center of the circle to these two locations respectively is termed the propagation angle of the subset for that reference

plane. If the subset has two perpendicular reference planes as described above, simple or weighted average of the two propagation angles may be computed to obtain a single propagation angle θK, for the subset. If the directivity pattern across a certain reference plane, e.g. a vertical plane, is more important than that across the other, perpendicular reference plane, greater weight is assigned to the more important plane. For example, in certain applications the most important consideration may be to prevent unwanted noise from reaching locations on different floors in which case a vertical reference plane would be more heavily weighted than the horizontal reference plane.

An example of a suitable criterion for the "best-ness" of a specific LUT is: LUTscor _e = 1/ [(average of all 9K ₁ ) x (standard deviation of all θK,)]

To determine the most suitable LUT, one may use a computer simulation to test and score all possible permutations i.e. all possible N-cell LUTs, and selecting the best one thereof.

It is appreciated that the number of LUTs, each containing N cells, is N! (N factorial). If N is sufficiently large, it becomes impractical to test and evaluate all possible LUTs i.e. all possible permutations of integers into LUT cells. If such is the case, a smaller number of LUT permutations may be selected, e.g. randomly, and the best one thereof is selected.

It is appreciated that alternatively, step 3200 may be performed without resort to a fixed LUT stored during set-up. Instead, the set of P _t -P _t-1 pressure producing elements to be activated may be selected by selecting the best subset of P _t -P _t -i elements from among the set of pressure producing elements which are currently in the second operative configuration. This may be done by estimating the propagation angle θ for each possible subset of P _t -P _t-1 elements and selecting that subset which best matches the desired propagation pattern.

P _t refers to the current value of P whereas P _t- i refers to the value of P in the previous system clock.

Furthermore, it is appreciated that in those applications in which the directivity pattern is not important, any set of pressure producing elements may be employed to achieve a temporal pressure pattern dictated by the input signal.

Step 3300, in which the amount of delay with which to operate each of the moving elements of moving element array 3010 or 3012 of Figs. 2OA - 2OB, is computed, determines the directionality of the sound generated by the speaker.

Preferred methods and formulae for optionally positioning the moving element array as a function of desired directionality of propagation, if possible, and for computing delays also as a function of desired directionality of propagation, are now described, for three example propagation patterns termed herein omni-directional, cylindrical and uni- directional. It is appreciated that the three propagation patterns discussed particularly herewithin are discussed merely by way of example.

Figs. 22A - 22B, taken together, describe a simplified example of a solution for performing step 3300 when it is desired to achieve omni-directional sound i.e. sound which propagates outward through three dimensional space from a given point location termed herein the "focal point" of the omni-directional sound. Specifically, Fig. 22 A is a simplified diagram of an omni-directional propagation pattern having a focal point 3400 and Fig. 22B is a diagram of a preferred positioning of a moving element array relative to the focal point of the desired omni-directional sound propagation pattern of Fig. 22 A. In the illustrated embodiment, the array of moving elements referenced generally 3010 or 3012 in Figs. 2OA - 2OB comprises, merely by way of example, a typically non- skewed array 3410 of 14 x 21 moving elements. As shown, the array of moving elements is preferably although not necessarily positioned, as illustrated in Fig. 22B, such that its geometric center (located between the 7 ^l and 8 ^th rows, at the 11 ^th column) of the array of moving elements coincides with the focal point 3400 of the omni- directional pattern, located at the center of the concentric circles representing the omnidirectional pattern as shown in Fig. 22 A. The center of the array may also be positioned at the projection of the focal point of the omni-directional pattern onto the plane of the moving element array 3410. It is appreciated that the array need not be positioned as illustrated and may instead be positioned at any suitable location, such as a fixed location independent of the particular propagation pattern currently selected by a user.

It is appreciated that the array need not be of the specified dimensions or shape. In fact preferred embodiments of direct digital speakers are comprised of thousands to hundreds of thousands of pressure-producing elements. The shape of the array may change according to application and/ or use. It is also appreciated that the focal point referred to herein need not be positioned on the main surface defined by the array of pressure-producing elements. Changing the distance between the focal point and the main surface of the array of the pressure-producing elements changes the directionality pattern of the device. E.g.

placing the focal point on the surface (zero distance) would produce true omnidirectional directivity pattern where sounds intensity remain essentially equal regardless of the angle in which the sound propagates. Placing the focal point at a certain distance, d behind the surface of pressure -producing elements defines a projection cone (in the case of a round array) or a projection pyramid (in the case of a square or rectangular array) that is characterized by a head angle narrower than 180 degrees. Placing the focal point at an infinite distance behind the main surface of the pressure producing elements (given that the sound produced by the pressure producing elements is produced in front of the main surface) typically defines a projection cone or a projection pyramid that is very narrow and would produce a true unidirectional directivity pattern. Typically, the sound intensity throughout the projection cone or projection pyramid remains essentially equal while the intensity outside the cone or pyramid is significantly lower. It is appreciated that d may be either 0 or infinity in certain applications. In certain applications, d may be determined as a function of a user control. Fig. 23 is a simplified pictorial illustration of speaker apparatus constructed and operative in accordance with Figs. 2OA - 22B and operative, e.g. by virtue of having been so programmed, to generate omni-directional sound which is particularly suitable for the environment illustrated in Fig. 23 in which consumers of the sound entirely surround the speaker, typically at more than one levels including a ground level and a first floor level as shown.

For applications in which a pre-determined and fixed focal point of omnidirectional sound propagation is known, e.g. in a conventional planetarium, circus arena or circular auditorium, the array of moving elements provided in accordance with certain embodiments of the present invention is preferably although not necessarily positioned such that the array's center coincides with the desired focal point of the desired omni-directional propagation pattern as described above with reference to Figs. 22A - 22B. It is appreciated, however, that although it is preferable to position the array of moving elements such that its center is disposed as close as possible to a currently desired e.g. user-selected focal point of a currently desired e.g. user selected omni-directional propagation pattern, nonetheless, embodiments of the invention shown and described herein allow omni-directional propagation from a wide variety of foci to be achieved using an array of moving elements which may be stationary and need not be centered at the focal point of the omni-directional directivity pattern.

Referring back to Fig. 22A, each circle shown represents half a phase and has a radius r which is computed using the following formula: r = (N d λ / 2 + N ² λ /4 ) ^{0 5} where: N = the serial number of the circle, counting outward from the center and starting from 1, d = the distance of the plane of the non-skewed array from the focal point of the omni-directional sound λ = c T, where c = the speed of sound through the medium in which the speaker is operating, typically air, and T = the period of the system clock of Figs. 2OA or 2OB (not shown).

It is appreciated that specific delay values for the moving elements in array 3410, suitable for achieving the omni-directional pattern of Fig. 22 A, may be determined as follows:

(a) Any moving element which coincides with a circle whose serial number is N is assigned a delay value of N T /2.

(b) Any moving element which does not coincide with a circle, and instead falls between a pair of circles whose serial numbers are N and N+l is assigned a delay value by interpolating e.g. linearly between the following two values: NT/2 and (N+l)T/2.

Alternatively, a suitable formula for determining delays is described in detail below.

Fig. 24 is a diagram of a cylindrical pattern of sound directivity which it is achievable using an embodiment of the apparatus of the present invention. As shown, at each point along the sound propagates, in the plane, omni-directionally and identically from each point along a given focal axis 3510. Fig. 25 is a diagram showing one preferred positioning of the moving element array 3010, shown to be rectangular by way of example, relative to the cylindrical pattern of sound directivity shown in Fig. 24. If possible, the moving element array is preferably disposed symmetrically about the focal axis such that, as shown, its sides are respectively perpendicular to or parallel to the focal axis, or, less preferably and as illustrated, the moving element array is preferably disposed symmetrically about the projection of the focal axis upon the plane defined by the array. It is appreciated that the particular positionings described herein and illustrated need not be provided and

alternatively, the moving element array may be disposed and oriented in any suitable, application-dictated location.

Fig. 26 is an isometric view of the moving element array 3010 or 3012 of Figs. 2OA - 2OB, showing uni-directional sound generated by that moving array and propagating in a desired or predetermined direction α as indicated by arrows.

Fig. 27 is a pictorial illustration of a preferred application for speaker apparatus 3600 constructed and operative in accordance with the present invention, being constructed e.g. programmed to generate uni-directional sound in at least one typically user-selected direction. In the embodiment of Fig. 27, either or both of two uni- directional sounds streams 3610 and 3620 are generated to serve listeners located at positions 3630 and 3640 respectively.

If the array of moving elements is a rectangle 3650 having first and second internally parallel and mutually perpendicular pairs of sides, then the array is typically oriented such that the projection of the desired direction of propagation onto the plane of the array, which may be vertical as shown in Fig. 27, is parallel to one of the pairs of sides and hence perpendicular to the other pair of sides as shown in Fig. 26. In this case, the delay of each moving element in the array may be the product of cos alpha (where alpha is the angle of propagation as shown in Fig. 22) and the distance x of that moving element from a selected one of the pair of perpendicular sides. As suggested by Fig. 27, it may be desired to produce two sound streams to be perceived exclusively and simultaneously by listeners positioned in two respective azimuthal positions relative to the speaker, is now described. According to this embodiment, some of the moving elements in the array are devoted to producing the first sound stream, and the remaining elements in the array are devoted to producing the second sound. The delays used for each moving element are determined as described above in the uni-directional case. It is appreciated that more generally, any suitable number of sound streams rather than just two sound streams may be produced.

The uni-directional embodiment illustrated in Figs. 26 or 27 has a wide variety of applications, such as but not limited to (a) entertainment content providers, such as television, computer, music player or radio, including a programmable directional speaker operative to send uni-directional sound exclusively to one or more user- selected directions. Different content, such as a plurality of language

versions corresponding to a single visual content item, can be sent simultaneously to a plurality of user-selected directions, thereby to enable a group of friends or family members to share a viewing experience but to simultaneously and exclusively receive individualized audio content, e.g. to each in his own language, corresponding to the viewing experience; and (b) Sound-producing toys including a sensor operative to monitor the azimuthal and elevational position of the child relative to the toy and a directional speaker operative to send uni-directional sound exclusively toward the child in a direction which corresponds to the child's current azimuthal and elevational position relative to the toy.

Generally, a suitable formula for determining a suitable amount of delay for each moving element, for omni-directional sound propagation, is as follows: delay = [(d ² +r ² ) ^{0 5} -d]/c where r = distance between the projection of the focal point onto the moving elements array plane and a given moving element, d = the distance of the plane of the array of the moving elements from the focal point of the omni-directional sound c = the speed of sound through the medium in which the speaker is operating, typically air.

For cylindrical sound propagation, the same formula may be employed, however d is now defined as the distance of the plane of the array of moving elements from the focal axis which is typically parallel thereto

For unidirectional sound propagation, as described above, the formula employed may be delay = x cosα where: x = the distance from the moving elements array edge plane and a given moving element and α = the angle between direction and moving elements array plane.

It is appreciated that the embodiments shown and described herein generate sound propagation patterns which well approximate desired patterns such as omni-

directional, cylindrical, and uni-, bi- or even multi-directional patterns. However, at least due to the finite size of the array of moving elements, the actual sound propagation pattern is never exactly identical to the theoretically desired propagation pattern. Generally, the theoretically desired propagation pattern is better achieved at locations which are close to the moving element array, than at locations which are further from the moving element array.

It is also appreciated that the larger the array (both in terms of number of pressure-producing elements and in terms of dimensions), the more closely the desired propagation pattern is achieved. A particular feature of certain embodiments of the present invention is that a single speaker including one or more pressure-producing element arrays which arrays may be fixed, can be programmed to generate a plurality of directivity patterns differing in parameterization or even in shape.

It is appreciated that a multi-unidirectional propagation pattern may be provided, in which the user can, if desired, select the number of and/or direction and/or other characteristics of more than one uni-directional beams. The uni-directional embodiment is described herein and generalization of the uni-directional embodiment described herein to a multi-unidirectional embodiment may be achieved using techniques known in the art such as techniques used to define the direction, number of, and/or other characteristics of beam/s produced by multi-beam phased array applications e.g. RADAR beams. More generally, it is appreciated that a combination of propagation patterns may be provided, in which the user can, if desired, select the number of and/or direction and/or other characteristics of more than one component propagation patterns each of which may comprise any suitable pattern such as but not limited to a uni- directional pattern, omni-directional pattern, cylindrical pattern, or any combination thereof. Several propagation patterns are described herein and combination thereof may be achieved using techniques known in the art such as techniques used to define the direction, number of, and/or other characteristics of beam/s produced by multi-beam phased array applications e.g. RADAR beams It is appreciated that the array of moving elements need not be planar as illustrated and that alternatively the teachings of the present invention may be appropriately modified to accommodate a non-planar array of moving elements.

Fig. 28 is a simplified pictorial illustration of a non-rectangular array of moving elements. According to one embodiment of the present invention, delays for moving elements in non-rectangular arrays may be computed by circumscribing the non- rectangular array in a rectangular array and proceeding to compute delays as described herein, for the circumscribing rectangular array. Each moving element in the non- rectangular array is assigned a delay value which equals the delay value computed by this process, i.e. according to its position in the (imaginary) circumscribing rectangular array.

If the array of moving elements is not rectangular, the following rules may be employed to position the array although alternatively the invention may accommodate an array of moving elements positioned arbitrarily: i. If omni-directional propagation is desired and the designer is entirely free to position the array, the array may be positioned such that the center of mass of the non- rectangular array coincides with the focal point of the omni-directional propagation. Preferably and more generally, the array may be positioned such that the center of mass of the non-rectangular array is as close as possible to the focal point of the omnidirectional propagation ii. If cylindrical propagation is desired and the designer is entirely free to position the array, the array may be positioned such that an axis of mass 3700 of the non-rectangular array 3710, partitioning the array into two sub-arrays 3720 and 3730 of equal area as shown in Fig. 28 is disposed along the focal axis of the cylindrical propagation. If several axes of mass exist, the longest such axis is typically selected.

Preferably and more generally, the array may be positioned such that an axis of mass, preferably the longest available, of the non-rectangular array is disposed parallel to the focal axis of the cylindrical propagation. iii. If uni-directional propagation is desired and the designer is free to position the array, the array may be positioned such that the desired propagation direction is close to perpendicular to the main surface of the array.

The scope of the present invention includes but is not limited to a method for controlling direct digital speaker apparatus receiving a digital input signal and generating sound accordingly, the method comprising providing an array of pressure- producing elements, and computing a timing pattern determining if and when each pressure-producing element is operative to produce pressure pulses so as to achieve a

desired directivity pattern. The array is then operated in accordance with the timing pattern in order to achieve sound having the desired directivity pattern.

Optionally, the providing and computing steps are performed a plurality of times thereby to obtain a corresponding plurality of arrays and a corresponding plurality of timing patterns defining a corresponding plurality of directivity patterns. The method then also comprises the step of operating the plurality of arrays simultaneously in accordance with the corresponding plurality of timing patterns respectively thereby to obtain a single directivity pattern comprising a combination of the directivity patterns corresponding to the plurality of timing patterns. The plurality of arrays may in fact comprise portions of a single larger array. So, for example, a single array of pressure producing elements such as any of those shown and described herein may be partitioned into regions, e.g. quarters, and the pressure producing elements in each region may be operated in accordance with its own particular timing pattern or delay pattern. For example, this allows a pattern of several, say four, different unidirectional beams to be achieved. Alternatively, to give another example, this allows, say, omnidirectional background sound to be superimposed on one or more different foreground sound streams each respectively having its own, say, uni-directional, cylindrical or omnidirectional propagation pattern. It is appreciated that in multi-directional embodiments, each said unidirectional beam may produce a different digital input signal, e.g. the left and right channels of a stereophonic signal.

It is appreciated that the electromagnetic field controller 30 is preferably designed to ensure that the alternating current flowing through the coil maintains appropriate magnetic field strength at all times and under all conditions so as to allow sufficient proximity between the moving elements 10 and the electrostatic latches 20 to enable latching, while preventing the moving elements 10 from moving too fast and damaging themselves or the latches 20 as a result of impact.

With specific reference to the Figures, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the preferred embodiments of the present invention only, and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the invention. In this regard, no attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention. The description taken with the drawings

makes apparent to those skilled in the art how the several forms of the invention may be embodied in practice.

Features of the present invention which are described in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, features of the invention which are described for brevity in the context of a single embodiment may be provided separately or in any suitable subcombination. For example, moving elements may be free floating, or may be mounted on filament-like flexures or may have a surrounding portion formed of a flexible material. Independently of this, the apparatus may or may not be configured to reduce air leakage therethrough as described above. Independently of all this, the moving element may for example comprise a conductor, coil, ring- or disc-shaped permanent magnet, or ring- or discshaped ferromagnet and the magnets, if provided, may or may not be arranged such that the poles of some e.g. 50% thereof are oppositely disposed to the poles of the remaining e.g. 50% of the magnets. Independently of all this, the latch shape may, in cross-section, be solid, annular, perforated with or without a large central portion, or notched or have any other suitable configuration. Independently of all this, control of latches may be individual or by groups or any combination thereof. Independently of all this, there may be one or more arrays of actuator elements which each may or may not be skewed and the cross-section of each actuator element may be circular, square, triangular, hexagonal or any other suitable shape.

The present invention has been described with a certain degree of particularity, but those versed in the art will readily appreciate that various alterations and modifications may be carried out to include the scope of the following Claims: