Air pulse generating element and sound producing device with virtual valve

This application claims the benefit of U.S. provisional application No. 62/969,014, filed on Jan. 31, 2020, U.S. provisional application No. 63/029,074, filed on May 22, 2020, and U.S. provisional application No. 63/033,713, filed on Jun. 2, 2020, which are incorporated herein by reference.

1. Field of the Invention

The present application relates to an air pulse generating element and a sound producing device, and more particularly, to an air pulse generating element and a sound producing device capable of enhancing sound pressure level.

2. Description of the Prior Art

Conventionally, to cover a full range of human audible frequencies, e.g., from 20 Hz to 20 kHz, tweeter(s), mid-range speaker (s) and woofer(s) have to be incorporated within a conventional speaker set. All these devices would occupy large space of the conventional speaker set and will also raise its production cost. Hence, one of the design challenges for the conventional speaker is the impossibility to use a single sound producing device to cover the full range of human audible frequency.

Another design challenge for producing high-fidelity sound by the conventional speaker is its enclosure. The speaker enclosure is often used to contain the back-radiating wave of the produced sound to avoid cancelation of the front radiating wave in certain frequencies where the corresponding wavelengths of the sound are significantly larger than the speaker dimensions. The speaker enclosure can also be used to help improve, or reshape, the low-frequency response. Therefore, with properly selected speaker driver and enclosure parameters, the combined enclosure-driver resonance peaking can be leveraged to boost the output of sound around the resonance frequency and therefore improve the performance of resulting speaker.

However, when the speaker enclosure's internal volume is not sufficiently large, the ratio of the air volume movement to the volume of the enclosure rises and the pressure fluctuation inside the enclosure also rises, which causes nonlinearity or distortion of membrane movement near the min-max peaks of its movement range. To avoid these problems, the conventional speaker enclosures are expected to contain sufficiently large volumes. For most high-fidelity speakers, the internal enclosure volumes are therefore generally vastly larger than the physical volumes of their speaker drivers.

Recently, micro-speakers are expected to be contained within devices such as smartphones, tablet notebooks, smartwatches, smartglasses, etc. In any of such compact devices, the volume of speaker is usually made extremely small because of the size constraint of its host devices. But, if the conventional speaker has a reduced size, the quality/fidelity of the sound produced by the conventional speaker is drastically degraded because of the above issues.

Therefore, how to overcome the design challenges is an important objective in the field.

It is therefore a primary objective of the present application to provide an air pulse generating element and a sound producing device, to improve over disadvantages of the prior art.

An embodiment of the present application discloses an air pulse generating element. The air pulse generating element comprises a membrane, actuated according to a driving signal; a plate, wherein a chamber is formed between the plate and the membrane; wherein during a first phase of the driving signal, the membrane is actuated to change a volume within the chamber to generate an air pulse; wherein during a second phase of the driving signal, the membrane is actuated such that a gap is temporarily formed; wherein the temporarily formed gap is configured to provide a temporary air shunt to accelerate an air balancing process between a first air pressure at a first side of the membrane and a second air pressure at a second side of the membrane.

An embodiment of the present application further discloses a sound producing device. The sound producing device comprises a substrate; and at least an air pulse generating element, disposed on the substrate, wherein one of the at least an air pulse generating element comprises a membrane; a plate, wherein a chamber is formed between the plate and the membrane; wherein during a first phase, the membrane is actuated to change a volume within the chamber to generate an air pulse; wherein during a second phase, the membrane is actuated such that a gap is temporarily formed; wherein the temporarily formed gap is configured to provide a temporary air shunt to accelerate an air balancing process between a first air pressure at a first side of the membrane and a second air pressure at a second side of the membrane.

These and other objectives of the present invention will no doubt become obvious to those of ordinary skill in the art after reading the following detailed description of the preferred embodiment that is illustrated in the various figures and drawings.

FIG. 1 is a schematic diagram of a cross sectional view of an air pulse generating element according to an embodiment of the present application.

FIG. 2 is a schematic diagram of a top view of a membrane according to an embodiment of the present application.

FIG. 3 is a schematic diagram of a perspective view of a membrane when the membrane is actuated according to an embodiment of the present application.

FIG. 4 illustrates two lines representing membrane displacement versus driving voltage applied on an actuator.

FIG. 5 illustrates multiple samples of relative position pairs on different sides of slit.

FIG. 6 is a schematic diagram of a sectional view of an actuator according to an embodiment of the present application.

FIG. 7 is a schematic diagram of a driving signal according to an embodiment of the present application.

FIG. 8 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 9 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 10 is a schematic diagram of a plurality of driving signals according to an embodiment of the present application.

FIG. 11 is a schematic diagram of electrical pulses according to an embodiment of the present application.

FIG. 12 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 13 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 14 is a schematic diagram of a plurality of driving signals according to an embodiment of the present application.

FIG. 15 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 16 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 17 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 18 is a schematic diagram of a sound producing device according to an embodiment of the present application.

FIG. 19 illustrates pressure profiles of with and without gap opening.

FIG. 20 illustrates waveforms of a driving signal versus membrane displacement.

FIG. 21 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 22 is a schematic diagram of control signals for the driving circuit in FIG. 21 .

FIG. 23 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 24 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 25 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 26 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 27 is a schematic diagram of control signals for the driving circuit in FIG. 24 .

FIG. 28 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 29 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 30 is a schematic diagram of a driving circuit according to an embodiment of the present application.

FIG. 31 is a schematic diagram of a sound producing apparatus according to an embodiment of the present application.

U.S. Pat. No. 10,425,732 filed by Applicant provides a sound producing device comprising a plurality of air pulse generating elements which is capable of producing a plurality of PAM (pulse-amplitude modulation) air pulses at an ultrasonic pulse rate, higher than an maximum human audible frequency.

U.S. Pat. No. 10,771,893 filed by Applicant provides a SEAM (single ended amplitude modulation) driving signal for a sound producing device capable of producing single-ended PAM air pulses at ultrasonic pulse rate, in order to further enhance the sound pressure level performance and low audio frequency response. The SEAM driving signal comprises a plurality of electrical pulses, where the plurality of electrical pulses has the same polarity compared to (or with respect to) a certain voltage. For SEAM driving signal, each electrical pulse cycle comprises a PAM phase and an RST (reset) phase, which will be illustrated later on. The SEAM driving signal may be a PAM signal within the PAM phase and return to a reset voltage within the RST phase.

U.S. application Ser. No. 16/802,569 filed by Applicant provides a sound producing device which produces air pulses via chamber compression/expansion excited by membrane movement and the air pulses are propagated via through holes or openings formed either on the membrane or on a plate of the sound producing device, in order to achieve significant air pressure with small size/dimension of the sound producing device.

In application Ser. No. 16/802,569, each air pulse cycle comprises a pulse-generating time segment and a pulse-isolation time segment, and the compressed/expanded chamber air pressure (generated within the pulse-generating time segment) would be balanced with/to an ambient air pressure only through the opening(s), which requires long pressure balancing time, a time for the device to balance the chamber air pressure with/to the ambient air pressure.

The long pressure balancing time causes the pressure variation within the pulse-generating time segment to be canceled by the pressure variation within the pulse-isolation time segment. Thereby, a net pressure generated over the air pulse generating cycle would be small, which degrades the sound pressure level (SPL) performance in the human audible frequency band. In addition, when the pulse cycle ends before the pressure balancing time ends, significant residual air pressure may remain within the chamber at the end of the air pulse cycle, which leads to inter-pulse interference and causes degradation of sound quality. On the other hand, lengthen the pressure balancing time of RST phase will reduce the duty factor of PAM phase, leading to lower effective SPL over the pulse cycle.

To overcome the disadvantages stated in the above, the present application provides an air pulse generating element and a sound producing device capable of reducing the pressure balancing time by accelerating the speed of airflow, during the RST phase, between the volume inside the chamber and the ambient outside of the chamber and therefore speeding up the process of pressure balancing, reducing the time required to achieve balance, minimizing residual pressure at the end of RST phase, and therefore enhancing the net pressure to further enhance SPL performance while reducing inter pulse-cycle modulation.

FIG. 1 is a schematic diagram of a cross sectional view of an air pulse generating element 10 according to an embodiment of the present application. The air pulse generating element 10 comprises a membrane 12 , an actuator 14 , and a plate 16 . A chamber 13 is formed between the membrane 12 , the plate 16 and the spacer 104 . The actuator 14 is disposed on the membrane 12 . Openings/orifices 160 , regarded as pressure ejection orifices (PEOs), are formed on the plate 16 . The orifices 160 may be, but not limited thereto, formed by an etching process, and are permanently opened.

Similar to U.S. Pat. Nos. 10,425,732 and 10,771,893, or application Ser. No. 16/802,569, the air pulse generating element 10 is able to generate a plurality of air pulses. The air pulses generated by the air pulse generating element 10 would have non-zero offset in terms of SPL, where the non-zero offset is a deviation from a zero SPL, representing a pressure value of an ambient pressure outside the air pulse generating element 10 . Also, the plurality of air pulses generated by the air pulse generating element 10 may be aperiodic over a plurality of pulse cycles.

Similar to application Ser. No. 16/802,569, the membrane 12 is actuated to cause membrane displacement. The membrane displacement of the membrane 12 may compress or expand a chamber volume of the chamber 13 , to change a chamber pressure of the chamber 13 within an air pulse cycle T CY,A , such that an ultrasonic air pressure pulse is generated and propagated to the ambient outside of element 10 via the orifices 160 . Details of generating air pulse via chamber compression/expansion through orifices/openings may be referred to Ser. No. 16/802,569, which is not narrated herein for brevity.

Different from application Ser. No. 16/802,569, gap(s) is temporarily formed on the membrane 12 during the RST phase. The temporarily formed gap(s), during the RST phase, is configured to provide a temporary shunt to accelerate an air pressure balancing process between two sides of the membrane 12 .

In the description and claims of the present application, “slit” or “slit opening” shall refer to the opening widthwise along the cuts through membrane 12 , being parallel to the X-Y plane. “Gap” or “gap opening” shall refer to the opening along the Z direction, the direction of membrane displacement, between membrane portions, perpendicular to the X-Y plane.

In an embodiment, in order to form the gap(s), slit(s) may be formed on the membrane, which is to create another airflow passage way (in addition to the ones through the orifices 160 ), especially for accelerating the balancing of the chamber air pressure and/versus an ambient air pressure during RST phase of a pulse generating cycle. By driving the actuator 14 appropriately, gap opening may be temporarily formed to create a shunt between the 2 sides of membrane 12 , connecting the pressure/volume above the membrane 12 in the charmer 13 with the pressure/volume below the membrane 12 in the chamber 17 , to speed up the pressure balancing process, reduce the pressure balancing time and minimize the residual pressure at the end of the pulse cycle.

For example, FIG. 2 is a schematic diagram of a top view of the membrane 12 according to an embodiment of the present application. As shown in FIG. 2 , slits 11 a , 11 b , 11 c and 11 d are formed on the membrane 12 . The slits 11 a , 11 b , 11 c and 11 d partition the membrane 12 into membrane portions 12 a , 12 b , 12 c and 12 d . That is, the membrane portion 12 a is between the slits 11 d and 11 a , the membrane portion 12 b is between the slits 11 a and 11 b , and so on and so forth. The actuator 14 comprises actuating portions 14 a , 14 b , 14 c and 14 d disposed on the membrane portions 12 a , 12 b , 12 c and 12 d , respectively.

Take the slit 11 a as an example, the slit 11 a is formed between the membrane portions 12 a and 12 b . In other words, the membrane portion 12 a and the actuating portion 14 a are at a first side S 1 of the slit 11 a , and the membrane portion 12 b and the actuating portion 14 b are at a second side S 2 of the slit 11 a . When the membrane portions 12 a and 12 b are sufficiently actuated (or the actuating portions 14 a and 14 b are sufficiently driven), a first displacement Uz_a of the membrane portion 12 a along the Z direction by the first side S 1 is significantly larger than a second displacement Uz_b of the membrane portion 12 b along the Z direction by the second side S 2 , such that a gap 110 a is (temporarily) formed. The Z direction is parallel to the Z axis and perpendicular the XY plane, and the membrane 12 is parallel to the XY plane when the membrane 12 is not actuated (or when the membrane 12 is at a neutral state). The size of the gap 110 a is determined by a membrane displacement difference ΔUz=|Uz_a−Uz_b| between two sides of the slit 11 a . The larger is the difference ΔUz, the wider will be the gap 110 a.

A status of the slit openings being formed can be visualized by FIG. 3 . FIG. 3 is a schematic diagram of a 3D perspective view of the membrane 12 when the membrane 12 , including portions 12 a and 12 b , is sufficiently actuated. As can be seen from FIG. 3 , the gap 110 a is clearly formed/opened. The gap 110 a will be formed/opened when the membrane displacement difference ΔUz is significantly large, i.e., larger than a thickness of the membrane 12 . In the present application, the membrane portions 12 a and 12 b being sufficiently actuated and the actuating portions 14 a and 14 b being sufficiently driven indicate that the resulting displacement difference ΔUz is larger than the membrane thickness and the gap 110 a is formed/opened, so that the air pressure in compression chamber 13 above membrane 12 may be balanced with/to the air pressure below membrane 12 , such as in the chamber 17 of FIG. 1 , by connecting these two volumes to each other directly via the gap 110 a.

The width of slit, in the X-Y plane, is generally around 1 μm. The shape/pattern of the slit (e.g., 11 a ) is not limited. For example, in the embodiment illustrated in FIG. 2 , the slit 11 a extends toward a central portion 120 of the membrane 12 , e.g., from a corner 15 a . The slit 11 a may comprise a straight segment close to the membrane corner 15 a and a curved segment close to the central portion 120 of the membrane 12 . The slit 11 a may form as a hook pattern, and the slits 11 a - 11 d forms as a vortex pattern. A curvature of the slit 11 a may increase as the slit 11 a extending from the corner 15 a toward the central portion 120 . Specifically, a first radius of curvature at a first point A on the slit 11 a is smaller than a second radius of curvature at a second point B on the slit 11 a , where the point A is farther away from the corner 15 a compared to the point B, i.e., and a first length along the slit 11 a between the point A and the corner 15 a is larger than a second length along the slit 11 a between the point B and the corner 15 a.

Rationale of forming the gap openings is described below. Refer to points C and D illustrated in FIG. 2 . The point C is located on the membrane portion 12 a by the S 1 side of slit 11 a , and the point D is located on the membrane portion 12 b by the S 2 side of slit 11 a . The point D is opposite to the point C. The displacement Uz_a of the membrane portion 12 a at the point C is driven by the actuating portion 14 a and the displacement Uz_b of the membrane portion 12 b at the point D is driven by the actuating portion 14 b . A distance DC from the point C to an anchor edge of the membrane portion 12 a is longer than a distance DD from the point D to an anchor edge of the membrane portion 12 b . Since less distance implies stronger stiffness, deformation at the point D would be less than deformation of the point C, even applying the same driving force. In addition, the arrow DC overlaps with the region of the actuating portion while the arrow DD does not, which implies that the driving force applied at the point C is stronger than which at the point D. Combining those factors, the displacement Uz_a of the membrane portion 12 a at the point C, where driving force strength is stronger while stiffness is lower, would be larger than the displacement Uz_b of the membrane portion 12 b at the point D.

The FEM (Finite Element Method) simulation results of an embodiment of FIG. 3 is shown in FIG. 4 , where lines 401 - 402 representing the membrane displacement versus the applied voltage driving the actuator 14 . The line 401 represents an average of the displacement Uz_a by the side S 1 along the slit 11 a , and the line 402 represents an average of the displacement Uz_b by the side S 2 along the slit 11 a . According to the results shown in FIG. 4 , it represents that the slope of the line 401 is significantly greater/steeper than which of the line 402 , and thus, a gap (e.g., 110 a ) would be developed (i.e., opened or formed) when the driving voltage applied to the actuator 14 (including the actuating portions 14 a and 14 b ) is sufficiently large, where the applied voltage being sufficiently large means that the membrane displacement difference is larger than the thickness of the membrane 12 .

Furthermore, the membrane portion 12 a may have a first edge by the side S 1 along the slit 11 a , and the membrane portion 12 b may have a second edge by the side S 2 along the slit 11 a . The second edge may be far away from the actuating portion 14 b compared to the first edge and the actuating portion 14 a . For example, a first distance between the first edge and the actuating portion 14 a is less than a second distance between the second edge and the actuating portion 14 b . In this case, a first slope corresponding to a first edge of the membrane portion 12 a , representing a first relationship between a first displacement of the membrane portion 12 a along the first edge versus the applied voltage, is steeper than a second slope corresponding to a second edge of the membrane portion 12 a , representing a second relationship between a second displacement of the membrane portion 12 b along the second edge versus the applied voltage.

In another perspective, FIG. 5 illustrates 6 samples of relative position pairs 501 - 506 of the point C on the membrane portion 12 a and the point D on the membrane portion 12 b , corresponding to six progressively higher actuator driving voltage levels V 1 -V 6 , as labeled on the horizontal axis of FIG. 5 . Vertical axis represents displacements of the membrane portions 12 a (at the point C) and 12 b (at the point D).

As illustrated in FIG. 5 , the point C is significantly below the point D when the voltage V 1 (e.g., V 1 =1V) is applied on the actuator 14 . Due to the slope difference of the lines 401 and 402 , the membrane portion 12 a would be deformed more than the membrane portion 12 b , as the driving voltage applied to the actuator 14 rises. When the driving voltage applied to the actuator 14 is up to V 6 (e.g., V 6 =36V), a gap 507 is created, where the gap 110 a will be “fully opened”. When gap 110 a is fully opened, as illustrated in FIG. 3 , the volumes on the two sides of membrane 102 transition from being separated to being connected. In other words, an air shunt is formed temporarily between the pressures in the two volumes on the two side of the membrane 102 by the formation of gap 110 a (and 110 b , 100 c , 110 d , etc.).

In other words, the membrane portion 12 a is partially below the membrane portion 12 b when the applied voltage is at the voltage levels V 1 or V 2 . The membrane portion 12 a is substantially aligned to the membrane portion 12 b when the applied voltage is around the voltage level V 3 . The membrane portion 12 a is partially above the membrane portion 12 b when the applied voltage is at the voltage levels V 4 or V 5 . When the applied voltage is V 6 , greater than a certain voltage level (between V 5 and V 6 ), the membrane portion 12 a is completely above the membrane portion 12 b.

Note that, the numerical values shown in FIG. 5 are for illustrative purpose, practical applied voltage may be adjusted according to practical circumstance.

The slit width should be sufficiently small, e.g., 1 um in practice. The slits 11 a - 11 d form narrow channels over the thickness of the membrane, which may be 5-7 um in practice. Airflow through narrow channels can be highly damped due to viscous forces/resistance along the walls of the airflow pathways, known as boundary layer effect within field of fluid mechanics. In comparison, the gap opening 110 a , as illustrated by 507 of FIG. 5 , are between the corners/edges of two plates, which represents minimum surface parallel to the direction of the airflow and therefore minimize the viscous force exerted on the airflow. So, when the membrane 12 is below a gap actuating voltage (e.g., when the driving voltage applied on the actuator 14 is the voltages V 1 -V 5 ) and no gap in Z-direction (e.g., 110 a or 507 ) is formed, the air flows through the slits 11 a - 11 d will be much smaller compared to the airflow through PEO 160 in plate 16 , or much smaller than the airflow through the air shunt, which means that the slits 11 a - 11 d may be considered substantially sealed relative to the airflow through PEO 160 in plate 16 during the period of the pulse generation cycle.

In other words, the slit width is sufficiently small such that, the airflow/leakage through the slit when the applied voltage is between V 1 and V 5 is negligible compared to (e.g., less than 1% of) the airflow through the air shunt when the gap 110 a / 507 is formed.

In addition, supposed that V 1 ′ denotes a minimum voltage level corresponding to which the membrane portion 12 a is partially below the membrane portion 12 b and V 5 ′ denotes a maximum voltage level corresponding to which the membrane portion 12 a is partially above the membrane portion 12 b . A difference between V 1 ′ and V 5 ′ may be determined by the thickness of the membrane 12 . The difference between V 1 ′ and V 5 ′ may be larger when the membrane 12 is thicker. In other word, the thickness of the membrane 12 may determine a range of the voltage swing SW during the period of the pulse generation cycle in which the slit is considered/substantially sealed, i.e., the airflow/leakage through the slit is negligible compared to the airflow through the PEO 160 or through the air shunt.

The type of driving force applied on the membrane 12 (produced by the actuator 14 ) is not limited. For example, the membrane 12 may be driven by piezoelectric force, electrostatic force, electromagnetic force, or electrothermal force, which is not limited thereto. For example, the actuator 14 may be a NED (Nanoscopic Electrostatic Drive) actuator.

Take piezoelectric actuator as an example, the actuator 14 may comprise a first (top) electrode 14 T, an actuating material 14 P and a second (bottom) electrode 14 B, as FIG. 6 illustrates. The actuating material 14 P may comprise PZT (lead zirconate titanate) material (preferably with high ε 31 ). The actuator 14 may receive a SEAM driving signal DS.

FIG. 7 illustrates a schematic diagram of a waveform of the driving signal DS corresponding to an input audio signal AUD. The driving signal DS may be applied to the top electrode 14 T while a DC bias voltage V B is applied to the bottom electrode 14 B, and voltages V 14T and V 14B show in FIG. 7 represent voltages on the top electrode 14 T and the bottom electrode 14 B, respectively. The driving signal DS comprises a plurality of electrical pulses EP. The electrical pulse EP has an electrical pulse cycle T CY,E . The electrical pulse cycle T CY,E comprises a PAM (pulse-amplitude modulation) phase, corresponding to time period T PAM , and a RST (reset) phase, corresponding to time period T RST . In reality, the air pulse cycle T CY,A and the electrical pulse cycle T CY,E would have the same length, or more or less have the same length. Both of them may be abbreviated as the pulse cycle T c y. In an embodiment, the voltage V 14B may represent a (bias) voltage applied on the bottom electrode ( 14 B).

As the SEAM driving signal, all of the plurality of electrical pulses EP would have the same polarity with respect to a voltage V RST , or be unipolar with respect to the voltage V RST . In the embodiment illustrated in FIG. 7 , the electrical pulses EP are all negative (i.e., have negative polarity) with respect to (or compared to) the voltage V RST . That is, except for the transition periods, the voltage level of DS will be below V RST during the PAM phase, and equal to V RST during RST phase.

A voltage swing SW within the PAM phase T PAM may be determined by an instantaneous sample TS of the input audio signal AUD. The driving signal DS swings over the voltage swing SW within the PAM phase T PAM resulting in PAM phase DS voltage levels in the range of V 1 ˜V 5 of FIG. 5 , corresponding to the status where no gaps are formed across the slit openings and therefore the “gap” are considered “non-opened” within the PAM phase T PAM .

By driving membrane with proper voltage so that gaps, such as 110 a of FIG. 3 , are formed temporarily during the RST phase along slits 11 a ˜ 11 d to let the airflow pass through membrane 12 while no such gaps appears during the PAM phase, operations of the slits 11 a - 11 d and their associated temporary gaps 110 a - 110 d , are similar to the open-and-close movement of valves. Note that, no physical/actual valve is constructed and no mechanical contact occurred across the surfaces of slits 11 a ˜ 11 b in the air pulse generating element 10 , however, the air pulse generating element 10 retains functionality of valve(s) by utilizing the different slopes of displacement vs. voltage illustrated in FIG. 4 , and the interplay between the differential displacement and the thickness of the membrane as illustrated in FIG. 5 , to cause gap openings to be formed temporarily during the RST phase (the “opened” state) and disappear during the PAM phase (the “non-opened” state). Thereby, complexity and drawbacks of realizing physical valves, such as stiction, with the MEMS sound producing device is sophisticatedly bypassed.

Due to the presence of valve functionality in the absence of physical valve, the mechanism of the temporarily forming the Z-direction gap opening, such as 110 a in FIG. 3 , is also referred to as “virtual valve” in the present application. Note that, a critical factor that enables “virtual valve” to exhibit the functionality of “physical valve” is the brevity time of the valve needs to appear to be “non-opened”. Despite the width of the slit being 1 μm or less, given long enough time, air will flow through the slit and the pressure in chamber 13 would not be maintained. So, although it is critical to minimize the slit width, in the X-Y plane, it is also crucial to consider the length of time in order to keep the net airflow during the PAM phase under acceptable range. If the length of PAM phase is too long, significant air may flow through slits 11 a ˜ 11 d even without forming gap such as 110 a and the pressure within chamber 13 may fall too much and the output SPL may drop significantly. Therefore the pattern, length and width of slits, the durations of PAM phase T PAM and the duration of RST phase T RST need to be addressed as a whole in order to optimize the performance from the “virtual valve” operation discussed above.

Furthermore, referring back to FIG. 1 , the air pulse generating element 10 may optionally comprises ridges 102 , side walls 106 and a back cover 108 . The ridges 102 are disposed on the plate 16 . A back chamber 17 is formed between the membrane 12 , the sidewalls 106 and the back cover 108 . Optionally, a vent hole 109 may be formed within the back cover 108 . The vent hole 109 may be a hole or a tube and may form a bass reflex port in similar manner as conventional speaker back enclosure, causing the 180° off phase back radiation to experience additional 180° phase shift at the port resonance and become in-phase with the front radiation, resulting in an increase of SPL at the resonance frequency of the back chamber volume and the optional opening (vent hole 109 ).

FIG. 19 illustrates curves 201 and 201 * representing pressure profiles P versus time t. The vertical axis P is the difference between pressure inside the volume of chamber 13 and the ambient pressure outside of device 10 and 0 means no difference from ambient pressure. Dashed line 301 corresponds to driving signal DS, where the leading edge of dashed rectangle corresponds to the start of PAM phase of one particular pulse cycle. For line 301 , vertical axis corresponds to voltage level (DS−V RST ) and, therefore, 0 corresponds to the voltage level of V RST of FIG. 7 . The curve 201 * represents pressure profile within the chamber 13 when V RST corresponds to V 5 of FIG. 5 , i.e. no gap opening during the RST phase. The curve 201 represents pressure profile within the chamber 13 when V RST corresponds to V 6 of FIG. 5 , i.e. one or more gap opening, such as 110 a , is formed during the RST phase. The leading portion of curves 201 * and 201 overlap with each other, because DS signal level is the same in both cases, leading to the same membrane behavior during the PAM phase. The trailing portion of the curve 201 *, roughly corresponding to the RST phase where DS V 5 , has a long tail because, at DS V 5 , no gap will be formed over membrane 12 and the pressure inside chamber 13 , resulting from membrane movement during PAM phase, needs to be neutralized with the ambient above plate 16 , mainly by the airflow through the PEO 160 . In comparison, the trailing portion of the curve 201 , roughly corresponding to RST phase where DS V 6 , has a lower peak pressure, a faster decay time and a smaller final residual, because temporary shunts are formed to connect the volumes on the two sides of membrane 12 with each other through the gaps, such as illustrated by 3D perspective view 110 a or 2D cross sectional view 507 , and these shunts allow the air to quickly exchange between the two connected volumes and therefore accelerates the speed to achieve pressure balance between the two temporarily connected volumes. Note that, instead of waiting for majority of the airflow through the listener-facing side PEO of device 10 (opening 160 on plate 16 ) as in the case of curve 201 *, majority of the pressure balancing/neutralizing airflow occurs via air exchange between chamber 13 and the volume 17 on the back-facing side of device 10 in the case of curve 201 .

Supposed that the air pulse cycle under the case without forming gap opening is T CY,A *. A net air pressure ΔP CY,A * produced within the air pulse cycle T CY,A *, is evaluated by an integration of the curve 201 * over a time interval [0, T CY,A *], which is an area encompassed by the curve 201 * within the time interval [0, T CY,A *], equivalently, a summation of an area 203 and an area 204 *. Due to the areas 203 and 204 * are comparable, the net air pressure ΔP CY,A * is small, and the resulting SPL would be small.

Another drawback of no gap opening formed is the amount of residual pressure. As can be seen from FIG. 19 , a residual pressure 205 * is left at an end of the air pulse cycle T CY,A *, which would cause inter-pulse interference and degrade sound quality.

The small net air pressure and residual pressure problems can be solved by the forming gap opening during the RST phase by proper membrane design and corresponding driving signal such as illustrated by FIGS. 2, 3 and 5 . Supposed that the air pulse cycle, under the case that gap opening(s) is formed during the RST phase, is T CY,A . A net air pressure ΔP CY,A produced within the air pulse cycle T CY,A is a summation of the area 203 and an area 204 . Due to the gap opening(s) being formed during the RST phase, the area 204 will be significantly smaller relative to area 204 *, and thereby the net air pressure ΔP CY,A and the SPL would be significantly enhanced.

Furthermore, the residual pressure may be reduced from 205 * to 205 , which is zero or close to zero. In other words, it can be considered that there is barely or insignificant residual pressure at the end of the air pulse cycle T CY,A . Or, at the end of the air pulse cycle T CY,A , the chamber air pressure may be almost equal to the ambient air pressure.

Moreover, given the far faster decay of curve 201 relative to curve 201 *, the air pulse cycle may be reduced from T CY,A * to T CY,A , while still reducing the residual pressure from 205 * to 205 , thereby the increase duty factor from T PG /T CY,A * to T PG /T CY,A , where T PG represents the pulse-generating time segment, and thus improve the output SPL significantly.

In addition, an efficiency of generating air pressure may be evaluated by a ratio r=(ΔP CY /Δ|P CY |). ΔP CY may be defined as ΔP CY =∫P(t) dt and Δ|P CY | may be defined as Δ|P CY |=∫|P(t)| dt, where the integration operation ∫·dt is performed over one air pulse cycle. According to finite-element simulations, for the case without gap opening (represented by the curve 201 *), the ratio r may be 1˜3%, while the case where the gap opening is formed (represented by the curve 201 ), the ratio r increases to 30˜40%, an improvement of roughly 20 times.

In this regard, the air pulse generating element 10 may generate a plurality of net air pulses, and the plurality of net air pulses would have the same polarity relative to the ambient pressure outside the air pulse generating element 10 .

In order to optimize membrane displacement, it is desirable for the duration of the driving signal PAM pulse width to be approximately equal to or slightly larger than half the cycle time of membrane resonance frequency. Therefore, the relationship between the driving signal duty factor 1/n, membrane resonant frequency f R , and pulse rate f pulse may be described by equation:

1 2 · f R ≈ 1 n · f P ⁢ u ⁢ l ⁢ s ⁢ e ⁢ ⁢ or ⁢ ⁢ f R f P ⁢ u ⁢ l ⁢ s ⁢ e ≈ n 2 .

Supposed the ratio

f R f P ⁢ u ⁢ l ⁢ s ⁢ e

can be expressed as

f R f P ⁢ u ⁢ l ⁢ s ⁢ e = k · n ,

where k=0.42˜0.55. When this condition is satisfied, the membrane can reach the maximum displacement while the duty factor approximate maximum and therefore help maximize the resulting output SPL. In an embodiment of duty factor=1/3, i.e. n=3, the ratio f R /∫ pulse shall be within 1.25 to 1.65.

To elaborate, FIG. 20 illustrates schematic diagram of waveforms of the driving signal DS versus membrane displacements with respect to different k. In FIG. 20 , a dot line represents a waveform of the driving signal DS; a dashed-double-dot line, a bold-solid line and a dot-dashed line represent waveforms of membrane displacement with k>>0.6, k≈0.5 and k<<0.4, respectively.

If the ratio (f R /f pulse ) is too low, e.g., f R /f pulse <<0.4×n, the membrane 12 would be too slow to react to the driving signal and the membrane 12 may not have reached its steady state (DC) displacement by the end of the PAM phase, let along the overshooting, as the dot-dash line (k<0.4) in FIG. 20 illustrated. On the other hand, if the ratio (f R /f pulse ) is too high, e.g., f R /f pulse >>0.6n, the membrane movement would have already passed its overshooting peak and would be in a ringing pattern and may have even fallen below the steady state displacement of membrane 12 , as the dashed-double-dot line (k>0.6) in FIG. 20 illustrated. In comparison, when f R /f pulse ≈0.5 n, the peaking of the membrane displacement, comprising both the DC displacement and the overshooting, is aligned with the end of the PAM phase, as illustrated by the bold-solid line in FIG. 20 , and therefore maximize the pressure generated during the PAM phase.

FIG. 8 is a schematic diagram of a sound producing device (SPD) 80 according to an embodiment of the present application. The SPD 80 comprises a substrate 81 and a plurality of air pulse generating elements 82 . The air pulse generating elements 82 are disposed on the substrate 81 . Material of substrate 81 is not limited. The air pulse generating elements 82 may be realized by the air pulse generating element 10 , as an example. In fact, the air pulse generating element 82 may be realized by elements comprising membrane with slit(s) such that the gap opening(s) can be temporarily formed, i.e., be realized by elements comprising the “virtual valve” mechanism, which is within the scope of the present application.

Furthermore, the concept of “pulse interleaving”, disclosed in U.S. Pat. No. 10,536,770 may be exploited to achieve a higher overall pulse rate. When the “pulse interleaving” is operated at an overall pulse rate PR (in terms of pulses per second), the plurality of air pulse generating elements can be subdivided into several groups (e.g., N groups), each group of air pulse generating elements may generate air pulses at a pulse rate PR/N, and air pulses generated by different groups of air pulse generating elements are spatially-and-temporarily interleaved. Details of the “pulse interleaving” concept/mechanism may be referred to U.S. Pat. No. 10,536,770, which would not be narrated herein for brevity.

Note that, as can be seen from FIG. 4 and FIG. 5 , actuating the membrane such that the gap opening is formed over the slit lines during every pulse cycle T c y would require high voltage swing between the electrodes 14 T and 14 B. With the capacitance existing between the electrodes within the actuator (e.g., between the electrodes 14 T and 14 B), the amount of charge required to charge and discharge the actuator electrodes of the air pulse generating element 82 would be high.

On top of the large amount of charge required, the fast rising and fast falling PAM driving waveform exacerbate the situation, causing sharp spikes of switching currents during transitions between PAM and RST phases. Strong EMI (Electro Magnetic Interference) would be induced. Moreover, since the sheet resistance of the bottom electrode has significant resistance (e.g., the typical sheet resistance of the bottom electrode of PZT may be 0.5Ω/□ in practice, where □ denotes a symbol representing “square”), the high current would easily induce significant voltage drop, causing the driving voltage to be dissipated along the signal passage and would have decayed badly before reaching the electrode of the piezo actuator layer 14 P. As an illustration, consider a bottom electrode wiring trace consists of a conductor of a 1:2 rectangle, or 2 squares lengthwise, which will have a trace resistance of 2×0.5 Ω/□=1Ω. When a switching pulse current of 30 A flows through this conductor, the voltage drop would be 30V. Although this will not actually occur because the driving voltage itself is only roughly 30V, but it is clear that the rising and falling edge will be dramatically slowed down and the driving waveform as seen by the piezoelectric material layer 14 P would have deteriorated badly. This example shows the importance to minimize the bottom electrode wiring trace resistance in order to minimize voltage drop due to the high sheet resistance metal layer typically used as the bottom electrode in the piezoelectric actuator.

In order to overcome the above issues, a novel interleaving and cell arrangement methodology (as well as driving circuit) is proposed in the present application, where the air pulse generating elements, such as the 82 's of FIG. 8 , are subdivided into N spatially interleaved subsets, and these N subsets of elements are driven by N sets of temporarily interleaved driving waveform of 1/N duty factor, such that the current from the bottom electrodes of one subset of elements travels to a neighboring subset of elements directly, without needing to leave the device 80 from a pad, and reenter device 80 from another pad.

In the following description, the term air pulse generating element and air pulse generating cell will be used interchangeably.

FIG. 9 illustrates a schematic diagram of an SPD 90 according to an embodiment of the present application. The SPD 90 comprises air pulse generating elements 92 _ 1 - 92 _ 3 . As shown in the left portion of FIG. 9 , the air pulse generating elements 92 _ 1 , 92 _ 2 , 92 _ 3 are arranged in a 3×M array, where the elements 92 _ 1 occupy the 1 st rows of the array, the elements 92 _ 2 occupy the 2 nd row, the elements 92 _ 3 occupy the 3 rd row, and M herein denotes a number of elements within each row. The SPD 90 may comprise a substrate (not shown in FIG. 9 for brevity) on which the air pulse generating elements 92 _ 1 - 92 _ 3 is disposed. One/each of the air pulse generating elements 92 _ 1 - 92 _ 3 may comprise an actuator with a first electrode (e.g., top electrode) and a second electrode (e.g., bottom electrode). The actuator within each of the (air pulse generating) elements 92 _ 1 - 92 _ 3 may have similar or same cross sectional structure of the actuator 14 illustrated in FIG. 6 . In an embodiment, the bottom electrodes of the air pulse generating elements 92 _ 1 - 92 _ 3 may be electrically connected via the common metal layer, such as Platinum (Pt) alloy, which makes up the common bottom electrode 14 b.

Construct of the air pulse generating elements 92 _ 1 - 92 _ 3 is not limited. The air pulse generating elements 92 _ 1 - 92 _ 3 may be realized by the air pulse generating element 10 illustrated in FIG. 1 , or other kind of (piezo actuated) air pulse generating element, which may be able to generate a plurality of air pulses, where the plurality of air pulses may produce a non-zero offset and be aperiodic over a plurality of pulse cycles. In an embodiment, gaps may be also temporarily formed on the membranes within the air pulse generating elements 92 _ 1 - 92 _ 3 during the RST phase.

In addition, actuators within the air pulse generating elements 92 _ 1 - 92 _ 3 may have same/similar structure as the actuator 14 , which are reproduced in the right portion of FIG. 9 , in a cross sectional perspective. Top electrodes of the M air pulse generating elements 92 _ 1 are connected together to form one driving electrode 92 _ 1 T. Likewise for 92 _ 2 T and 92 _ 3 T. For sake of convenience, in the right portion of FIG. 9 , top electrodes of the elements 92 _ 1 , 92 _ 2 , 92 _ 3 are denoted as 92 _ 1 T, 92 _ 2 T, 92 _ 3 T, respectively.

Electrodes 92 _ 1 T- 92 _ 3 T are driven by driving signals DS 1 -DS 3 , respectively. That is, the electrodes 92 _ 1 T- 92 _ 3 T receives the driving signals DS 1 -DS 3 , respectively. The driving signals DS 1 -DS 3 may be temporally interleaved SEAM driving signal. That is, all of the electrical pulses within the driving signals DS 1 -DS 3 have the same polarity with respect to a certain voltage, e.g., V RST . In addition, electrical pulses EP_ 1 within the driving signal DS 1 , electrical pulses EP_ 2 within the driving signal DS 2 and electrical pulses EP_ 3 within the driving signal DS 3 are temporally and mutually interleaved, which means that, ignoring the transition time/edge of the electrical pulse, signal portions of the pulses EP_ 1 , EP_ 2 , EP_ 3 (out of the voltage V RST ) would not mutually overlap temporally. In other words, without considering the pulse transition time/edge of the electrical pulses, PAM phases of the driving signal DS 1 , PAM phases of the driving signal DS 2 and PAM phases of the driving signal DS 3 would not overlap one another. In signal waveform aspect, the portion of the pulses EP_ 1 out of the voltage V RST would neither overlap with the portion of the pulse EP_ 2 out of the voltage V RST temporally, nor with the portion of the pulse EP_ 3 out of the voltage V RST temporally, and vice versa.

Generally speaking, the pulse generating elements/cells 92 _ 1 - 92 _ 3 of device 90 may be (spatially) subdivided into n subdivisions/groups, where n=3 in the example shown in FIG. 9 . The bottom electrodes of all pulse generating elements of device 90 may be connected together and wired to a DC bias voltage source to receive the voltage V B via the metal layer used to construct the bottom electrode of the actuator, such as 14 B in FIG. 6 . Cells within each of the n subdivisions/groups may be arranged as a cell-rows (or simply a row), where n=3 and m≥4 in FIG. 9 , and cell-row means a row of cells. The cell-rows are disposed by and placed parallel to each other, in order to maximize the contact area between cell-rows and therefore minimize the square count of bottom electrodes between cell-rows.

The driving signals DS 1 -DS 3 may be SEAM waveform having duty factor of 1/n, where n may be an integer. In the embodiment shown in FIG. 9 , n=3. The timing between driving signals DS 1 -DS 3 is to be arranged, such that the trailing edge of DS 1 's PAM phase aligns with the leading edge of DS 2 's PAM phase, the trailing edge of DS 2 's PAM phase aligns with the leading edge of DS 3 's PAM phase, and the trailing edge of DS 3 's PAM phase aligns with the leading edge of DS 1 's PAM phase. In this arrangement, at the transition from the PAM phase of DS 1 to the PAM phase of DS 2 , the voltage signal swing applied to electrode 92 _ 1 T and 92 _ 2 T will be time aligned but polarity inverted and the current flow of top electrode 92 _ 1 T will be the opposite of the current flow of top electrode 92 _ 2 T.

By arranging cell-rows 92 _ 1 , 92 _ 2 and 92 _ 3 in spatially interleaved manner as illustrated in FIG. 9 , the broad interface between cell-rows helps lowers the inter cell-row bottom metal layer square-count significantly and therefore reduces the inter cell-row resistance between the bottom electrodes of 92 _ 1 , 92 _ 2 and 92 _ 3 . When a pair of opposite current flows occur between top electrodes 92 _ 1 T and 92 _ 2 T during the transition from the PAM phase of DS 2 to the PAM phase of DS 3 , the corresponding current flows at the bottom electrodes of cell-rows 92 _ 1 and 92 _ 2 will flow toward each other via the broad interface between these cell-rows 92 _ 1 and 92 _ 2 along paths of the least resistance. Due to the minimum distance traveled by these current pulses, both the voltage drop (I·R) and the power consumed (I 2 ·R) will be minimized, which lead to both higher fidelity of driving signals DS 1 ˜DS 3 on electrodes 92 _ 1 T- 92 _ 3 T and lower power consumption (and heat generation) on the metal layer of the bottom electrode.

Similar discussion applies to the transitions between the PAM phases of DS 2 and DS 3 and between the PAM phases of DS 3 and DS 1 .

From the perspective at the chip level, by time-align the transition edges among PAM phases between 92 _ 1 T- 92 _ 3 T, the bottom electrode current pulses caused by these PAM phase transitions will largely neutralize one another, within the metal layer forming the bottom electrodes, leading to greatly reduced net current flow on the chip-level V B electrode, which is the pad connecting the metal layer of the bottom electrodes to the outside. In summary, from the perspective chip- 90 , the current flow caused by SEAM switching waveform self-annihilates one another between neighboring cell-rows, within the bottom electrode metal layer of device 90 , and the net current flow for the device 90 chip-level bottom electrode will be largely independent of the individual SEAM pulse switching waveform applied to top electrodes 92 _ 1 T- 92 _ 3 T. Instead, the current flow in and out of chip-level V B electrode would be determined mainly by the envelope of the 3 interleaved SEAM driving signals DS 1 ˜DS 3 , i.e., the input audio signal AUD (see FIG. 1 ). Thereby, power consumption would be significantly reduced.

From the signal waveform aspect, FIG. 10 illustrates the driving signals DS 1 -DS 3 , where the plot in the bottom of FIG. 10 illustrates a superposition of the driving signals DS 1 -DS 3 , and the rest plots in FIG. 10 illustrates the driving signals DS 1 -DS 3 individually. As FIG. 10 shows, the driving signal DS 1 comprises an electrical pulse EP 1 , the driving signal DS 2 comprises an electrical pulse EP 2 , and the driving signal DS 3 comprises electrical pulses EP 3 and EP 4 . The pulse EP 1 -EP 4 are all negative (or have negative polarity) with respect to the voltage V RST .

The electrical pulse EP 1 has a leading/falling transition edge TE 6 and a trailing/rising transition edge TE 1 . The electrical pulse EP 2 has a leading/falling transition edge TE 2 and a trailing/rising transition edge TE 5 . The electrical pulse EP 3 has a leading/falling transition edge TE 3 . The electrical pulse EP 4 has a trailing/rising transition edge TE 4 .

As shown in FIG. 10 , the transition edge TE 1 coincides with the transition edge TE 2 temporally, and the transition polarities of the transition edges TE 1 and TE 2 are opposite to each other. In addition, during the coincident transition of the transition edge TE 1 and the transition edge TE 2 , the switching pulse current would flow between the bottom electrodes of the air pulse generating elements 92 _ 1 - 92 _ 2 via the common metal layer.

For the sake of clarity, the transition edges TE 1 ˜TE 6 are widened and illustrated in FIG. 11 , where the transition edge TE 1 is corresponding a transition period/time TT 1 and the transition edge TE 2 is corresponding a transition period/time TT 2 , and so on. In an embodiment, the transition period/time corresponding to the transition edge TE 1 and the transition period/time corresponding to the transition edge TE 2 may partially, or completely, overlap with each other, which means that there exist a certain time te such that the time te is within both the transition time TT 1 and the transition time TT 2 . Preferably, the transition time TT 1 and the transition time TT 2 is completely overlapped. It means that, the time at which the transition edge TE 1 starts rising coincides with the time at which the transition edge TE 2 start falling (denoted as t 1 in FIG. 11 ), and the time at which the transition edge TE 1 stop rising coincides with the time at which the transition edge TE 2 stop falling (denoted as t 2 in FIG. 11 ). In the present application, either “the transition edges TE 1 and TE 2 being partially overlapped” or “the transition edges TE 1 and TE 2 being completely overlapped” is referred to the transition edges TE 1 and TE 2 coinciding with each other.

Similarly, referring back to FIG. 10 , the transition edge TE 3 coincides with the transition edge TE 5 temporally, and the transition edge TE 4 coincides with the transition edge TE 6 .

Rationale of exploiting coinciding transition edges with opposite transition polarities is iterated below. Take the transition times TT 1 and TT 2 in FIG. 11 as example, within the transition time TT 1 , it can be regarded that a current would flow into the top electrode, through the actuating material and to the bottom electrode of the (air pulse generating) element 92 _ 1 ; within the transition time TT 2 , it can be regarded that a current would flow from the bottom electrode, through the actuating material and out of the top electrode of the (air pulse generating) element 92 _ 2 . Given the fact that the transition times TT 1 and TT 2 coincide with other (and the bottom electrodes of the air pulse generating elements 92 _ 1 , 92 _ 2 are electrically connected), it can be regarded that a switching current, the current produced due to pulse transition, flows into the top electrode of the element 92 _ 1 , through the actuating material of the element 92 _ 1 and to bottom electrodes of elements 92 _ 1 , 92 _ 2 , and flows through the actuating material of the element 92 _ 2 and out of the top electrode of the air pulse generating element 92 _ 2 . Note that, during the transition time TT 1 /TT 2 , the switching current flows directly between elements 92 _ 1 and 92 _ 2 and would not flow to external path. Therefore, power consumption and unnecessary voltage drop (due to resistance of the bottom electrode material) are significantly reduced.

Note that, the SPD 90 comprises three arrays of air pulse generating elements in FIG. 9 , which is not limited thereto. As long as 1) the SPD comprises a plurality of (more than one) air pulse generating elements; and 2) two driving electrical pulses for two different air pulse generating elements have transition edges coinciding with each other and opposite transition polarities, the requirements of the present application is satisfied, which would be within the scope of the instant application. In other words, the two driving electrical pulses in item 2 stated in the above are consecutive to each other. Preferably, the two air pulse generating elements (air pulse generating element) neighbor each other, or a distance between the two air pulse generating elements is less than a length/width of the air pulse generating element itself.

In this regard, the SPD of the present application may comprise only two air pulse generating elements, as illustrated in FIG. 12 . An SPD A 0 is illustrated in a left portion of FIG. 12 . The SPD A 0 is similar to the SPD 90 . Different from the SPD 90 , the SPD A 0 comprises only two air pulse generating elements 92 _ 1 - 92 _ 2 , driven by the driving signals DS 1 -DS 2 . Waveforms of the driving signals DS 1 -DS 2 are illustrated in a right portion of FIG. 12 . As shown in FIG. 12 (similar to FIG. 10 ), the transition edges TE 1 and TE 2 coincide with each other and the transition polarities thereof (of the transition edges TE 1 and TE 2 ) are opposite to each other.

Similarly, FIG. 13 illustrates a schematic diagram of an SPD A 0 ′ according to an embodiment of the present application. Different from the SPD A 0 comprising two air pulse generating elements ( 92 _ 1 - 92 _ 2 ), the SPD A 0 ′ comprises three air pulse generating elements 92 _ 1 - 92 _ 3 , driven by the driving signal DS 1 -DS 3 , respectively.

FIGS. 10-12 illustrate a microscopic view of the waveforms of the driving signals DS 1 -DS 3 , demonstrating the transition edges thereof. In another perspective, for the SPD 90 , FIG. 14 illustrates a macroscopic view of waveforms of the driving signals DS 1 -DS 3 . In an embodiment, the labels “ 1 ”, “ 2 ” and “ 3 ” in FIG. 14 may represent the electrical pulses EP_ 1 within the driving signal DS 1 , the electrical pulses EP_ 2 within the driving signal DS 2 , and the electrical pulses EP_ 3 within the driving signal DS 3 , respectively.

As can be seen from FIG. 14 , the driving signals DS 1 , DS 2 , DS 3 are all SEAM signals, which means that the electrical pulses EP_ 1 , EP_ 2 , EP_ 3 (within the driving signals DS 1 , DS 2 , DS 3 , respectively) have the same polarity (e.g. negative in the embodiment FIG. 13 ) with respect to the voltage V RST , and are amplitude modulated according to the input audio signal AUD. Further, the driving signals DS 1 , DS 2 , DS 3 are mutually interleaved, with the rising/trailing transition edges of the electrical pulses EP_ 3 coincide with the falling/leading transition edges of the electrical pulses EP_ 1 at a time denoted as T 3=>1 , the rising/trailing transition edges of the electrical pulses EP_ 1 coincide with the falling/leading transition edges of the electrical pulses EP_ 2 at a time denoted as T 1=>2 , and the rising/trailing transition edges of the electrical pulses EP_ 2 coincide with the falling/leading transition edges of the electrical pulses EP_ 3 at a time denoted as T 2=>3 .

In FIG. 9 , one driving signal drives one cell-row of air pulse generating elements, which is not limited thereto. In the present application, one driving signal may drive multiple cell-rows of air pulse generating elements. For example, FIG. 15 is a schematic diagram of an SPD B 0 according to an embodiment of the present application. In FIG. 15 , the elements 92 _ 1 arranged as the 1 st and the 2 nd rows are driven by the driving signal DS 1 , the elements 92 _ 2 arranged as the 3 rd and the 4 th rows are driven by the driving signal DS 2 , and the elements 92 _ 3 arranged as the 5 th and the 6 th rows are driven by the driving signals DS 3 .

Note that, in FIG. 15 , a distance between the element 92 _ 1 and the element 92 _ 3 is longer than a distance between the element 92 _ 1 and the element 92 _ 2 and/or a distance between the element 92 _ 2 and the element 92 _ 3 . Longer distance brings higher resistance. In other words, at the transition time T 3=>1 , the current flowing between the bottom electrodes of the elements 92 _ 1 and the bottom electrodes of the elements 92 _ 3 would result in more voltage drop, compared to the one (voltage drop) caused at the transition time T 1=>2 by the current between (bottom electrodes of) the elements 92 _ 1 and 92 _ 2 , or the one caused at the transition time T 2=>3 by the current between the elements 92 _ 2 and 92 _ 3 .

To alleviate the extra resistance (or voltage drop) between the elements 92 _ 1 and 92 _ 3 brought by long distance between non-adjacent cells/elements, conducting lines may be included. FIG. 16 is a schematic diagram of an SPD B 0 ′ according to an embodiment of the present application. The SPD B 0 ′ further comprises conducting lines B 01 and B 02 . The conducting lines B 01 and B 02 may be made of conductive material (e.g., metal). The conducting line B 01 is disposed by/above the 1 st row of the array and electrically connected to bottom electrodes of the 1 st row of the air pulse generating elements 92 _ 1 ; the conducting line B 02 is disposed by/below the 6 th row of the array and electrically connected to bottom electrodes of the 6 th row of the air pulse generating elements 92 _ 3 . The conducting lines B 01 and B 02 may receive the voltage V B . Thus, in the SPD B 0 ′ shown in FIG. 16 , the resistance between the elements 92 _ 1 and 92 _ 3 would be close to the one (i.e., resistance) between the elements 92 _ 1 and 92 _ 2 or the one between the elements 92 _ 2 and 92 _ 3 .

Arrangement of the elements 92 _ 1 , 92 _ 2 and 92 _ 3 is not limited. For example, FIG. 17 is a schematic diagram of an SPD C 0 according to an embodiment of the present application. Different from the SPD B 0 , the arrangement of the elements 92 _ 1 , 92 _ 2 and 92 _ 3 within the SPD C 0 are arranged symmetrically, with respect to a symmetry axis Sym. The elements 92 _ 1 may be arranged in the 1 st row and the 6 th row of the array, the elements 92 _ 2 may be arranged in the 2 nd row and the 5 th row of the array, and the elements 92 _ 3 may be arranged in the 3 rd row and the 4 th row of the array, where the symmetry axis Sym lies between the 3 rd row and the 4 th row of the array.

The benefit of the arrangement of the SPD C 0 in FIG. 17 is to annihilate the severe EMI by its symmetry. Annihilating EMI by symmetry is elaborated below. During the transition time T 1=>2 (or equivalently, TT 1 in FIG. 11 ), a first pulse current would flow/rush into the top electrodes of the elements 92 _ 1 in the 1 st and 6 th rows. Meanwhile, during the transition time T 1=>2 (or equivalently, TT 2 in FIG. 11 ), a second pulse current, with amplitude similar/close to the first pulse current, would flow/rush out of the top electrodes of the elements 92 _ 2 in the 2 nd and 5 th rows. Thereby, a first EM (Electro Magnetic) radiation, generated between the elements 92 _ 1 in the 1 St row and the elements 92 _ 2 in the 2 nd row, would be 180 degree out of phase with a second EM radiation, generated between the elements 92 _ 2 in the 5 th row and the elements 92 _ 1 in the 6 th row. Due to the EMI interference being destructive, the first and the second EM radiations would auto-annihilate each other and result in a minimal net EM radiation. Same rationale also applies to the transition times T 2=>3 and T 3=>1 .

Furthermore, in an embodiment, the bottom electrodes of some of the elements 92 _ 1 , 92 _ 2 and 92 _ 3 , e.g., the bottom electrodes of the rightmost elements 92 _ 1 , 92 _ 2 and 92 _ 3 , may also receive the voltage V B .

FIG. 18 is a schematic diagram of an SPD C 0 ′ according to an embodiment of the present application. The SPD C 0 ′ is similar to the SPD C 0 . Different from the SPD C 0 , the SPD C 0 ′ further comprises conducting lines C 01 , C 02 and C 03 . Similarly, the conducting lines C 01 , C 02 and C 03 are made of conductive material (such as metal), receive the voltage V B , and play similar roles as the conducting lines B 01 and B 02 . Function of the conducting line C 01 /C 02 is to provide the voltage V B to the elements 92 _ 1 . Function of the conducting line C 03 is to provide the voltage V B to the elements 92 _ 3 . The relationship between (and the rationale behind having) the conducting lines C 01 /C 02 and C 03 of FIG. 18 is the same as the relationship between (and the rationale behind having) the conducting lines B 01 and B 02 of FIG. 16 .

Note that, referring back to FIG. 14 , the label “ 1 / 2 / 3 ” in FIG. 14 may denote the electrical pulses EP_ 1 /EP_ 2 /EP_ 3 , which is not limited thereto. The label “ 1 ” in FIG. 14 may refer to the electrical pulses EP_ 2 within the driving signal DS 2 or the electrical pulses EP_ 3 with the driving signal DS 3 as well. Similarly, the label “ 2 ” in FIG. 14 may refer to the electrical pulses EP_ 1 or the electrical pulses EP_ 3 , and/or the label “ 3 ” in FIG. 14 may refer to the electrical pulses EP_ 1 or the electrical pulses EP_ 2 , which depends on practical requirements.

In addition, topology of driving circuit generating the driving signals is not limited. FIG. 21 is a schematic diagram of a driving circuit E 0 according to an embodiment of the present application. The driving circuit E 0 may be configured to drive SPD (e.g., A 0 shown in FIG. 12 ) comprising the air pulse generating elements 92 _ 1 , 92 _ 2 . Note that, in SPD A 0 , the element 92 _ 1 may be disposed adjacent to the element 92 _ 2 , bottom electrodes of the elements 92 _ 1 and 92 _ 2 may be (electrically) connected and receives the voltage V B , and the top electrodes 92 _ 1 T and 92 _ 2 T receives the driving signal DS 1 and DS 2 , respectively.

As FIG. 21 shows, the driving circuit E 0 comprises nodes N 1 , N 2 , swapping modules E 01 , E 02 and switches S 1 -S 4 . The swapping module E 01 /E 02 respectively comprises a charge-exchanging switch CXSW 1 /CXSW 2 , an inductor L 1 /L 2 and a diode D 1 /D 2 connected in series. The swapping modules E 01 , E 02 are coupled between the nodes N 1 , N 2 . The nodes N 1 , N 2 and the swapping modules E 01 , E 02 are connected as a loop structure, which means that the node N 1 /N 2 can be regarded as a beginning of the loop structure as well as an end of the loop structure.

The node N 1 is coupled to the top electrode 92 _ 1 T of the air pulse generating element 92 _ 1 , configured to deliver the driving signal DS 1 to the air pulse generating element 92 _ 1 (i.e., to the top electrode 92 _ 1 T). Similarly, the node N 2 is coupled to the top electrode 92 _ 2 T of the air pulse generating element 92 _ 2 , configured to deliver the driving signal DS 2 to the air pulse generating element 92 _ 2 (i.e., to the top electrode 92 _ 2 T). Note that, supposed that the air pulse generating element 92 _ 1 / 92 _ 2 is PZT actuated, a capacitance C PZT,1 /C PZT,2 (not shown in FIG. 21 ) may exist between the node N 1 /N 2 and the bottom electrode(s).

The switches S 1 -S 2 are coupled to the node N 1 , and the switches S 3 -S 4 are coupled to the node N 2 . Besides coupled to the nodes N 1 and N 2 , a terminal of the switch S 1 and a terminal of the switch S 3 receive the voltage V RST . A terminal of the switch S 2 and a terminal of the switch S 4 receive a signal V AUD,1 and a signal V AUD,2 , respectively. V AUD,1 and V AUD,2 are two instances of samples of the input audio signal AUD, corresponding to the driving signals DS 1 and DS 2 , respectively.

The switches S 1 and S 3 are controlled by reset control signals RST 1 and RST 2 , respectively. The switches S 2 and S 4 are controlled by PAM control signals PAM 1 and PAM 2 , respectively. The charge-exchanging switches CXSW 1 and CXSW 2 are controlled by charge-exchanging control signals CX 1 and CX 2 , respectively.

The driving signal DS 1 /DS 2 may be both SEAM signals, which means that the driving signal DS 1 /DS 2 may be unipolar with respect to the voltage V RST (as taught in U.S. Pat. No. 10,771,893). In an embodiment, the voltage V RST may be sufficiently high such that the gap(s) would be formed on the membrane(s) within the air pulse generating element(s), especially when the air pulse generating element 92 _ 1 / 92 _ 2 shares structure similar to (or the same as) the air pulse generating element 10 . In this regard, the voltage level of V RST may be higher than voltage levels of V AUD,1 and V AUD,2 .

Timing diagrams of control signals RST 1 , RST 2 , PAM 1 , PAM 2 , CX 1 and CX 2 are shown in FIG. 22 . The driving signals DS 1 and DS 2 are also illustrated in bottom of FIG. 22 . Instead of sampling V AUD at the start of the PAM phases and hold the voltage level throughout the PAM phases, in this embodiment, driving signals DS 1 and DS 2 will track the audio signal during the PAM phases. For illustration purpose, in the following description, the switch is conducting when the corresponding control signal is in high state (logic “1”) and is non-conducting, or cut-off, when the corresponding control signal is in low state (logic “0”).

CND 1 /CND 2 represents time intervals during which the charge-exchanging switch CXSW 1 /CXSW 2 is conducting while the switches S 1 ˜S 4 are non-conducting and charge-exchanging switch CXSW 2 /CXSW 1 is non-conducting.

During a period T 1 , before a time interval CND 1 , switches S 2 , S 3 are conducting and switches S 1 , S 4 , CXSW 1 , CXSW 2 are cut-off, such that the driving signal DS 1 at the node N 1 would track the voltage level V AUD (labeled as V AUD,1 to differentiate from that of CND 2 , which is labeled as V AUD,2 ) and the driving signal DS 2 at the node N 2 would be the voltage level V RST . Note that, at a beginning of the conduction interval CND 1 , a voltage level at the node N 2 (i.e., V RST ) is higher than a voltage level at the node N 1 (i.e., V AUD,1 ).

After the start of the time interval CND 1 , when the charge-exchanging switch CXSW 1 is turned ON and start conducting and the switches S 1 -S 4 , CXSW 2 are all cutoff, current flow in inductor L 1 can be described by

I L ⁢ 1 , t i = ∫ t = t 0 t i ⁢ V N ⁢ 2 - V f - V N ⁢ 1 L ⁢ 1 ⁢ dt ,

where t 0 corresponds to the start of time interval CND 1 , V N1 , V N2 are voltage levels at nodes N 1 , N 2 and V f is the forward conducting voltage of diode D 1 . The current flowing through L 1 will keep increasing until V N2 =V N1 +V f where I L1 reaches its peak and starts decreasing until the current reaches 0 and the diode D 1 is turned off. Note that, during the course of I L1 decreasing, V N1 will swing toward V RST while V N2 will swing toward V AUD . In other words, during the interval CND 1 , assuming minimum energy losses due to I L1 ·V f and R L1 ·I L1 2 , by virtue of the inertia of inductor current I L1 , electrical charge stored in C PZT,2 of node N 2 will be transferred to C PZT,1 of node N 1 , resulting in the voltage levels of node N 1 and node N 2 being swapped effectively. Similarly, during the interval CND 2 , assuming minimum energy losses due to I L2 ·V f and R L2 ·I L2 2 , by virtue of the inertia of inductor current I L2 , voltage levels at node N 1 and node N 2 will swapped effectively.

After the conduction interval CND 1 (CND 2 ) and during a period T 2 (T 1 ), the switches S 1 , S 4 (S 2 , S 3 ) are conducted and the switches S 2 , S 3 (S 1 , S 4 ) and CXSW 1 , CXSW 2 are cutoff, the driving signal DS 1 (DS 2 ) at the node N 1 (N 2 ) would be near the voltage level V RST and the driving signal DS 2 (DS 1 ) at the node N 2 (N 1 ) would be near the voltage level V AUD,2 (V AUD,1 ).

Furthermore, during the conduction interval CND 1 (CND 2 ) toward the end of the conduction interval CND 1 (CND 1 ), the voltage level at the node N 2 (N 1 ) may overshoot or undershoot the voltage level V AUD,2 (V AUD,1 ) and, likewise, the voltage level at the node N 1 (N 2 ) may overshoot or undershoot the voltage level V RST . Once the period T 2 (T 1 ) starts, and the switches S 1 , S 4 (S 2 , S 3 ) are conducting, the voltage level at the node N 1 (N 2 ) will be pulled to V RST and the voltage level at the node N 2 (N 1 ) will be pulled to V AUD,2 (V AUD,1 ).

Note that, the period T 1 represents the PAM phase of the driving signal DS 1 , and the period T 2 represents the PAM phase of the driving signal DS 2 . Furthermore, from the waveform illustrated in the bottom of FIG. 22 , a rising edge of the driving signal DS 1 and a falling edge of the driving signal DS 2 are formed/coincided within the conduction interval CND 1 . Similarly, a falling edge of the driving signal DS 1 and a rising edge of the driving signal DS 2 are formed/coincided within the conduction interval CND 2 , by using (the same rationale of) the swapping module E 02 . In this case, the temporal interleaved electrical pulses EP_ 1 (within DS 1 ) and electrical pulses EP_ 2 (within DS 2 ) are generated, where transition edges of electrical pulses EP_ 1 and EP_ 2 coincide with each other and have inverse polarities.

Specifically, referring back to the swapping module E 01 in FIG. 21 and the interval CND 1 in FIG. 22 as an example, at the beginning of CND 1 , due to a voltage difference between the nodes N 1 and N 2 , the inductor L 1 would conduct a current (denoted as I 1 ) from the node N 2 to the node N 1 , and also play a role to resist current variation of the current I 1 . Besides, the inductor L 1 and the capacitance C PZT,1 /C PZT,2 may be viewed as form a LC oscillator. Nevertheless, the diode D 1 included in the swapping module E 01 functions as a rectifying component to allow the current flowing only from the node N 2 to the node N 1 and to block/prevent current I 1 flowing from the node N 1 back to the node N 2 . As a result, the current I 1 from the node N 2 to the node N 1 would be drawn, which is equivalent to the amount of electrons (or charges) is transferred/removed from the capacitance C PZT,1 at the node N 1 to the capacitance C PZT,2 at the node N 2 , during/after the conduction interval CND 1 .

By the same rationale, during/after the conduction interval CND 2 , the amount of electronics/charges would be (more or less) transferred back to the capacitance C PZT,1 at the node N 1 , via the loop structure formed by the nodes N 1 , N 2 and the swapping modules E 01 , E 02 . Furthermore, the conduction interval CND 1 corresponding to transferring charges from the node N 1 to N 2 and the conduction interval CND 2 corresponding to transferring charges from N 2 back to the node N 1 may take place within one (electrical) pulse cycle T CY,E .

Thereby, by interleaving DS 1 /DS 2 and by time align the transition PAM/RST phase transition edges of DS 1 /DS 2 , during the transitions (i.e., the rising/falling edges) of between the PAM phase and the RST phase of driving signals DS 1 /DS 2 , the energy stored in the piezo-speaker loading capacitance would be mostly recycled/reused, such that power consumption due to the PAM switching activities is significantly reduced. Therefore, the power consumed would be mainly determined by the envelop variation of the input audio signal AUD and largely independent of the PAM switching activities.

Note that, in the embodiment shown in FIG. 21 , diodes are used to perform the rectifying function, which is not limited thereto. Rectifying switches controlled by rectifying (control) signals may replace diodes in performing the rectifying function. For example, FIG. 23 is a schematic diagram of a driving circuit F 0 according to an embodiment of the present application. The driving circuit F 0 comprises swapping modules F 01 and F 02 . The swapping module F 01 /F 02 comprises a rectifying switch RTSW 1 /RTSW 2 . The rectifying switch RTSW 1 /RTSW 2 is controlled by a rectifying signal RT 1 /RT 2 . In the present application, diodes and rectifying switches are regarded as rectifying components.

The rectifying signal RT 1 /RT 2 controls the rectifying switch RTSW 1 /RTSW 2 to be conducted within (but not out of) the conduction interval CDN 1 /CDN 2 , as shown in right portion of FIG. 23 , to rectify the current flowing therethrough. In one embodiment zero crossing detection (ZCD) circuit (not shown in FIG. 23 ) may be included to generate the rectifying signal RT 1 /RT 2 for the rectifying switch RTSW 1 /RTSW 2 , such that the rectifying switch RTSW 1 /RTSW 2 will go from conducting to cutoff when the current flow through the inductor L 1 /L 2 approaches zero, and replacing the rectifying function of diodes D 1 /D 2 as a result.

Furthermore, ΔR in right portion of FIG. 23 represents a time gap between rising edges of CX n and RT n , ΔR is known as dead-time in fields such as gate driver for class-D amplifier, and ΔF represents a time difference between falling edges of CX n and RT n , where n=1 or 2 herein. ΔR (with ΔR≥0) and ΔF (with ΔF≥0) may be designed according to practical situation. Generally, ΔR corresponds to the time required to fully turn OFF S 1 ˜S 4 and ΔF may be determined by a ZCD circuit.

Note that, the internal components of the swapping modules F 01 and F 02 are identical. The only difference between F 01 and F 02 is the control timing of those two switches. It is therefore possible to merge two swapping modules F 01 and F 02 into one, F 0 X as in FIG. 24 , in which the swapping module F 0 X is treated as F 01 during time interval DN 1 and the swapping module F 0 X is treated as F 02 during time interval DN 2 . By such substitution of control signals, the swapping module F 0 X will not only be able to draw current from the node N 2 to the node N 1 , but also be able to draw current from the node N 1 to the node N 2 . That is, using swapping module with rectifying switch brings flexibility of drawing current in two opposite directions while reducing the number of inductors required from two to one.

Furthermore, it is known that two switches (controlled by signals, denoted as, e.g., CSG 1 and CSG z ) connected in series can be combined into one switch controlled by a control signal CSG=CSG 1 & CSG 2 , where “&” represent a logic AND operation. In this regard, the switches CXSW and RTSW in FIG. 24 can be combined as one switch (denoted as SW), and a swapping module F 0 X′ is formed and shown in FIG. 25 . The control signal for the switch SW within the swapping module F 0 X′ in FIG. 25 can be obtained by performing CX & RT, where CX and RT are control signals for the switches CXSW and RTSW, respectively, within the swapping module F 0 X shown in FIG. 24 . Note that, the switch SW also plays a role of rectifying component, and the control signal thereof may be dependent on a result of the ZCD circuit.

In another aspect, the voltage level at the node N 1 (equivalently, the driving signal DS 1 ) and the voltage level at the node N 2 (equivalently, the driving signal DS 2 ) are substantially swapped, at the end of interval CND 1 . For example, in the case of the circuit E 0 shown in FIG. 21 , before the conduction interval CND 1 , the voltage level at node N 2 is approximately V RST , higher than the voltage level at node N 1 , which is approximately at V AUD,1 ; after the conduction interval CND 1 ends, the voltage level of node N 1 is increased up toward V RST and the voltage level of node N 2 is decreased down toward V AUD,1 , close to V AUD,2 . In this case, the voltage levels at node N 1 and node N 2 are considered to be “substantially swapped” after the conduction interval CND 1 .

In addition, a general situation would also be considered as “voltage level of node N 1 and voltage level of node N 2 is substantially swapped.” That is, before the conduction interval CND 1 , the voltage level of node N 1 is smaller than a certain threshold and the voltage level of node N 2 is larger than the certain threshold; after the conduction interval CND 1 ends, the voltage level of node N 1 is increased up to be larger than the certain threshold and the voltage level of node N 2 is decreased down to be smaller than the certain threshold. In this case, the voltage levels of N 1 and N 2 are considered to be “substantially swapped” after the conduction interval CND 1 . In an embodiment, the certain threshold may refer to the voltage level obtained by performing pure charge-sharing operation between the capacitance C PZT,1 at the node N 1 and the capacitance C PZT,2 at the node N 2 .

FIG. 26 is a schematic diagram of a driving circuit G 0 according to an embodiment of the present application. The driving circuit G 0 may be configured to drive SPD comprising the air pulse generating elements 92 _ 1 - 92 _ 3 , e.g., the SPDs 90 , A 0 ′, B 0 , B 0 ′, C 0 , C 0 ′, in which the driving signal DS 1 -DS 3 are required, where the element(s) 92 _ 1 is disposed adjacent to the element(s) 92 _ 2 and the element(s) 92 _ 2 is disposed adjacent to the element(s) 92 _ 3 . Similarly, the bottom electrodes of the elements 92 _ 1 - 92 _ 3 may be (electrically and physically) connected and receive the voltage V B , and the top electrodes 92 _ 1 T- 92 _ 3 T receives the driving signals DS 1 -DS 3 , respectively.

In addition to components within the driving circuit E 0 , the driving circuit G 0 further comprises a node N 3 and a swapping module E 03 , where the swapping module E 03 comprises a charge-exchanging switch CXSW 3 , an inductor L 3 and a diode D 3 connected in series, similar to the swapping module E 01 /E 02 . A capacitance C PZT,3 (not shown in FIG. 24 ) may exist between the node N 3 and the bottom electrode of the element 92 _ 3 . In the driving circuit G 0 shown in FIG. 26 , the swapping module E 02 is coupled between the nodes N 2 and N 3 , and the swapping module E 03 is coupled between the nodes N 3 and N 1 . Similar to the driving circuit E 0 , the nodes N 1 -N 3 and the swapping module E 01 -E 3 as shown in FIG. 26 also forms a loop structure, which means that any of the nodes N 1 -N 3 may be regarded as a beginning as well as an end terminal of the loop structure.

Timing diagram of control signals for the driving circuit G 0 are shown in FIG. 27 . The swapping modules E 01 -E 03 , coupled between the nodes N 1 -N 3 , are conducted within the conduction interval CND 1 -CND 3 , respectively, and the swapping modules E 01 -E 03 can be regarded as iteratively conducted. Similar to the operation of the driving circuit E 0 , within the conduction interval CND 1 (corresponding to the transition time T 1=>2 shown in FIG. 14 ), an amount of electrons/charges stored in the capacitance/capacitor C PZT,1 at the node N 1 is transferred to the capacitance/capacitor C PZT,2 at the node N 2 . Within the conduction interval CND 2 (corresponding to the transition time T 2=>3 shown in FIG. 14 ), an amount of charges in the capacitance/capacitor C PZT,2 at the node N 2 is transferred to the capacitance/capacitor C PZT,3 at the node N 3 . Within the conduction interval CND 3 (corresponding to the transition time T 3=>1 shown in FIG. 14 ), an amount of charges in the capacitance/capacitor C PZT,3 at the node N 3 is transferred back to the capacitance/capacitor C PZT,1 at the node N 1 . In addition, via the loop structure formed within the driving circuit G 0 , charges may be transferred from the capacitance/capacitor C PZT,1 at the node N 1 , relayed through the capacitance/capacitor C PZT,2 at node N 2 , the capacitance/capacitor C PZT,3 at node N 3 and transferred back to the capacitor C PZT,1 at the node N 1 , within a time of the electrical pulse cycle T CY,E (shown in FIG. 7 and FIG. 14 ), where the driving signals DS 1 -DS 3 would all have the electrical pulse cycle T CY,E . Therefore, energy stored in the piezo speaker loading capacitance C PZT,1 , C PZT,2 and C PZT,3 would be recycled/reused and power consumed by SPD is mainly determined by the envelop of the input audio signal AUD.

In practice, approximately 60˜88% energy recycling is achievable. Such energy recycling is particularly important for generating SEAM driving signals of the present invention where the (reset) voltage V RST may be required to be sufficiently high (i.e. significantly higher than a threshold) to create sufficient large Z-direction gaps (e.g., 110 a ) on the membrane in order to accelerate the pressure balancing of FIG. 19 , reducing the area of 204 * to the area of 204 .

Similar to the discussion of FIG. 23 , the diodes in the driving circuit G 0 may be replaced by the rectifying switches. FIG. 28 is a schematic diagram of a driving circuit G 0 ′ according to an embodiment of the present application. Details of the driving circuit G 0 ′ may be referred to paragraphs stated in the above, which is not narrated herein for brevity.

Moreover, since at any moment only 1 instance of the swapping modules G 01 -G 03 is active, so it is possible to share the inductors, as illustrated by a driving circuit G 1 in FIG. 29 . Different from FIG. 28 , the inductors L 1 -L 3 in FIG. 28 are replaced by a shared inductor LS within the driving circuit G 1 in FIG. 29 , where the shared inductor LS is shared by swapping modules G 11 -G 13 within the driving circuit G 1 . In addition, the two terminals of the shared inductor LS are labeled as La and Lb and the wirings within G 11 -G 13 indicates the connection to the shared inductor LS.

Further note that, the switches RSTW 1 and CXSW 2 actually have the same wiring, so that they can be merged physically, but need to be controlled with control signal appropriate for function expected of their “role”. Similarly, the switches RSTW 2 and CXSW 3 can also be merged physically.

In other words, the switches RSTW 1 and CXSW 2 in the driving circuit G 1 may be merged as a switch, denoted as SW 2 b , connecting the node N 2 and the terminal Lb; the switches RSTW 2 and CXSW 3 in the driving circuit G 1 may be merged as a switch, denoted as SW 3 a , connecting the node N 3 and the terminal La. The driving circuit G 1 may be manipulated/simplified as a driving circuit G 2 , which is illustrated in FIG. 30 . In the driving circuit G 2 , the switch SW 1 a (connecting the node N 1 and the terminal La) is equivalent to the switch CXSW 1 of the driving circuit G 1 , and the switch SW 1 b (connecting the node N 1 and the terminal Lb) is equivalent to the switch RTSW 1 of the driving circuit G 1 . In this case, during the conduction interval CND 1 , the switches SW 1 a and SW 2 b are conducted; during the conduction interval CND 2 , the switches SW 2 b and SW 3 a are conducted; during the conduction interval CND 3 , the switches SW 1 b and SW 3 a are conducted.

FIG. 31 is a schematic diagram of a sound producing apparatus J 0 according to an embodiment of the present application. The sound producing apparatus J 0 comprises a sound producing device J 01 and a driving circuit J 02 . The SPD J 01 and the driving circuit J 02 may be properly chosen from the SPDs and the driving circuits stated in the above, depending on practical requirements.

Notably, the embodiments stated in the above are utilized for illustrating the concept of the present application. Those skilled in the art may make modifications and alterations accordingly, and not limited herein. For example, the air pulse generating elements driven by the driving circuits are not limited to be PZT actuated. Other actuating material may be exploited. Nevertheless, the driving circuits of the present application are especially suitable for actuator with capacitance formed between top and bottom electrodes. In addition, an order of the charge-exchanging switch, the inductor and the rectifying component within the swapping module is not limited. As long as they are connected in series within the swapping module, the requirement of the present application is satisfied, which is also within the scope of the present application.

In summary, the present application utilizes the temporarily opened slit openings to reduce the pressure balancing time, so as to enhance SPL performance; the present application arranges the air pulse generating elements in neighbor and drives the neighboring air pulse generating elements by electrical pulses with coinciding falling and rising transition edges, so as to reduce power consumption and EMI; the present application utilizes swapping modules to exchange electronic charges between the nodes, so as to save energy.

Those skilled in the art will readily observe that numerous modifications and alterations of the device and method may be made while retaining the teachings of the invention. Accordingly, the above disclosure should be construed as limited only by the metes and bounds of the appended claims.