GB2474848A - Planar loudspeaker - Google Patents

Abstract

The loudspeaker comprises a substantially planar, flexible diaphragm 2a and means for driving the diaphragm. The drive means comprises a linear conductor 1 a carried on the diaphragm and a magnet assembly 3. Vibration damping means 6, 8, 9 is arranged around the periphery of the diaphragm. The diaphragm is tensioned such that the speed of sound within the diaphragm is lower than the speed of sound in air (typically > 1 %).

Description

Loudspeaker

Technical Field of the Invention

The present invention relates to a loudspeaker. In particular, the present invention relates to loudspeakers that utilise a planar, flexible diaphragm.

Backqound Ideally a High Fidelity loudspeaker should cover the full audio range from 20Hz (or lower) to 20KHz (or higher) without audible resonances or distortion. For good stereo imaging the point the sound comes from should not appear to move with frequency. A fundamental problem with achieving this is that to give an even distribution of sound into the listening environment, the physical width of the radiating surface must be small at high frequencies, but in order to move enough air to reproduce bass frequencies at reasonable volumes it must be much larger.

There have been many approaches to this problem, but they mainly fall into two groups.

One approach is to have two or more drive units each sized for a smaller range of frequencies and an electrical cross-over to switch between them at some point in the frequency range. A problem with this approach is that for a significant proportion of the audio range, you have more than one driver producing sound and that a driver working at the top of its range has a different sound dispersion pattern from one at the bottom of its range. High-order crossovers and multiple drive units are often used to ameliorate this problem, but this results in an expensive solution, phase anomalies and a (difficult to drive) reactive load for the amplifier. Unless the effective acoustic centres of the drivers are coincident in all three axes, there will also be blurring of the stereo image. Another approach is to have a driver with a flexible radiating surface and to drive it over a small area. At low frequencies the whole surface will move with "pistonic motion". At high frequencies only the area near to the driving point will move significantly. The problem with this is that the switch between these modes happens in the audible range and that the driver will inevitably have a number of audible resonances. Various techniques have been applied to even out these resonances such as those used in NXT type speakers (e.g. EP1055351, EPI 068770), but a smooth response and even sound dispersion over the full audio frequency range is still not achievable.

In order to avoid the problems of either needing multiple drive units or having to handle the transition out of pistonic motion, inventors such as Lane (US1773910) and Burton (US4924504) have proposed designs where a line region on a flexible diaphragm is driven so that the diaphragm never moves as a piston. A difficulty to overcome with such designs is that in order to have good high frequency response the driven area must have low mass relative to its radiating area, but in order to generate sufficient air movement at low frequencies it must have a large excursion. To avoid generating harmonic distortion, all the forces on the diaphragm must be substantially constant or linear over the full range of this excursion. A constant electromagnetic force for a given input current is easily achieved with conductors moving through a narrow magnetic gap (e.g. US4924504 figure 11), but the conductors and magnetic gap must overlap throughout the full excursion for the force to be constant. The mass of conductors and support needed to achieve large enough low frequency excursions is higher than desirable for good high frequency response. An additional problem is that for a line source, the relatively stiff structure will have structural resonances along its length within the audio range. Also it can't drive the full height of the diaphragm as its stiffness would restrict movement if it came too near to the support frame.

To avoid excessive mass, the conductors may be applied directly to the diaphragm as shown in US4924504 figure 10. In this design however, the force in the conductors increases substantially as the conductors get closer to the magnets. Whilst in US4924504 it is stated that this is an advantage as it "combats the Mylar diaphragm's dynamic compliance", in practice for good low frequency response the tension in the diaphragm must be too low for the increase in tension force to match that of the magnets. A further problem is that the magnets are substantially in the way of sound radiated directly from the conductors, so frequency response and sound dispersion anomalies will be caused at the highest frequencies.

The need for a light diaphragm to give good high frequency response combined with that for a large excursion to give sufficient volume at low frequencies also presents difficulties with the damping of the diaphragm. In order to avoid resonances and audible echoes in the diaphragm, very little of the sound wave coming from the conductors must be reflected from the edges of the diaphragm. This problem is recognised in US4924504 and in figure 2 of that patent, open cell foam damping members (32, 33) are in direct contact with the diaphragm. For a low mass diaphragm however, even the lowest density foam is too stiff in comparison and reflects too much. In figure 9 of US4924504 an air cavity 85 is proposed. Such an air cavity however causes problems of its own. In order to give sufficient damping of low frequencies in the diaphragm it must be relatively deep and as such will have cavity resonances in the enclosed air that come within the mid to upper audible frequency ranges. A further problem is that to give sufficiently high damping, the air gap must be relatively narrow. At the extremes of diaphragm excursion, the air gap on one side of the diaphragm therefore closes down significantly and the damping rises in a very non-linear manner, causing harmonic distortion.

US1773910 and US4924504 both recognise that the frequency response of a constant width, constant mass, constant tension diaphragm will be uneven. Varying the width of the diaphragm is proposed in both, however this causes problems. Reflections off the edges are not returned in the direction they came from, but are bounced towards the areas of larger width. This is particularly the case for the lowest frequencies, which are most difficult to damp heavily without adversely affecting the low frequency output of the speaker. The first problem this creates is that the frequency response is dominated by that of the larger widths. The second is that this generates a component of the wave which travels vertically through the diaphragm. This can cause parts of the conductors to have large excursions whilst other parts have small excursions. This means that the maximum distortion-free volume level at low frequencies is limited by when the first portion of conductor hits the magnets and thus the full excursion over the whole diaphragm cannot be used. In US4924504, a similar problem occurs if slanted members 103 are attached to the diaphragm. The sound wave reflects off the members towards the regions of maximum width between members, creating the same problems previously described, plus the difficulty that waves reflected between the members do not benefit from the damping at the edge of the frame.

This invention is a full range line source loudspeaker with improvements implemented to overcome the frequency response and distortion problems described above.

Statements of Invention

According to a first aspect of the present invention there is provided a loudspeaker comprising: a substantially planar, flexible diaphragm; drive means for driving the diaphragm, the drive means comprising a linear conductor operably connected to the diaphragm and a magnet arrangement mounted in proximity to the linear conductor and arranged to provide a force that is substantially perpendicular to the surface of the diaphragm and substantially proportional to a current within the linear conductor; damping means arranged around the periphery of the diaphragm to damp waves arriving at the periphery from the drive means wherein the diaphragm is tensioned such that the speed of sound within the diaphragm is lower than the speed of sound in air.

The present invention provides a loudspeaker with a localised, linear "ribbon-like" (group of) conductor(s) propagating a wave through a diaphragm at a low speed such that the effective active width of the diaphragm increases continuously throughout the majority of the audio frequency range. Further preferred features of the invention relate to variants of the damping means, the arrangements of the magnets and conductors, corrugations in the conductor and the methods of avoiding peaks and troughs in the frequency response.

Description of the invention.

Figure 1 shows a loudspeaker according to the invention in its most basic form.

Figure 2 shows a horizontal sectional view cut through the loudspeaker of figure 1.

Figures 3 to 5 show sectional views of alternative embodiments to the basic loudspeaker.

Figures 6 to 13 show alternative arrangements of conductors and magnets.

Figure 14 shows an example calculation for the optimum positioning of the conductors illustrated in figure 13.

Figures 15 and 16 show alternative constructions for the support of the magnets as well as improvements to the conductors.

Figures 17 and 18 show methods of improving the smoothness of the frequency response of the loudspeaker as well as one technique for supporting the loudspeaker.

Referring to figures 1 and 2, the audio signal generates a current in a ribbon-like conductor (la) fixed to a diaphragm (2a -shown shaded in figure 1). A bar of magnets (3) is mounted adjacent to the conductor (la) to generate a magnetic field that intersects the conductor so as to provide a force substantially perpendicular to the diaphragm surface, substantially proportional to the current in the conductor (la). The diaphragm (2a) is tensioned such that a sound wave originating from the conductor (la) travels through the diaphragm at a speed much lower than that of sound in air (typically less than 100th of it).

What this means is that a wave travelling through the diaphragm is largely inaudible at any significant distance from the diaphragm as the positive and negative parts of the signal are physically so close together on the diaphragm that when radiated into air they substantially cancel out. Only the original audio signal generated at the conductor is therefore heard. If we imagine either the positive or negative portion of an audio signal travelling through the diaphragm however, the distance it gets before being cancelled out by the opposite polarity increases as the frequency decreases, so a larger area of the diaphragm becomes active, the lower the frequency. If the width "W" of the diaphragm is selected such that for the speed of sound through the diaphragm, half a wave of a (low) frequency to be produced well will fit within it, then the full diaphragm area becomes active at this frequency. Typically a polymer diaphragm of the order of I 3microns thick can easily be tensioned to become fully active at about 20Hz within a width of the order of 0.5-1 metres.

As described in relation to the prior art, for this method to work, it is very important that very little of the sound wave in the diaphragm (2a) is reflected from the edge (4), or the returning wave will generate an audible echo comparable to that of a stone building.

Initially, closed-cell (1.5mm thick neoprene) foam was used for the diaphragm. This had some success in damping out the waves, but because of its relatively high mass and internal damping it also tended to limit the high frequency response from the conductor (1 a). A thin polymer diaphragm (of the order of 1 3microns thick) allows the high frequency response to be good. This sort of material however has little internal damping, so needs damping at its edges to minimise reflections. As this material is very light, even the lowest density foam support at the edge would reflect too much (even if made with shallow (e.g. degree) triangular wedges to force multiple reflections off the foam). A solution was found by using an air cavity (5a) at the edge of the diaphragm lined with an open cell acoustic foam (6) to damp out sound travelling within the air of the cavity. Air gaps (5a) of the order of 1-2mm thick with acoustic foam (6) of the order of 3mm thick have been found to give good results. In order to give optimum damping whilst still allowing a large diaphragm excursion, it can be helpful to have a tapered air cavity (7) with acoustic foam lining (6). Alternatively a tapered (8) or stepped (9) acoustic foam may be used. As well as damping out sound waves within the air of the cavity, the acoustic foam serves another function. At the extremes of diaphragm excursion, as the diaphragm (2a) approaches the foam, air is able to be squeezed through the foam. This means that the damping does not increase as dramatically as it would with a non-porous boundary to the air cavity and therefore harmonic distortion is minimised. Whilst acoustic foam is described here, it will be appreciated that alternative porous damping media may be used such as felt, kapok or bonded acetate fibre.

In the basic design of figure 2, the force on the conductor (la) for a given current increases as it gets closer to the magnets (3), leading to harmonic distortion at large diaphragm excursions. As shown in figure 3, this effect can be reduced by adding another bar of magnets (10) on the opposite side of the diaphragm (2a). A further set of air cavities (5b) also lined with acoustic foam can also be added to increase the damping level and improve the linearity of damping over the full range of diaphragm excursions.

To improve the linearity of large diaphragm excursions without requiring a second set of magnets, figure 4 shows that a second diaphragm (2b) and conductor (ib) may be added.

As one conductor (la) gets closer to the magnet, the other one (ib) gets further away.

The same applies to the damping, so that as the air cavity (5c) on one diaphragm (2a) closes down, the air cavity (5d) on the other diaphragm (2b) opens up. One drawback of this approach is that the diaphragms must be physically separated by a distance of the order of 20-30mm. What this means is that at around 5-8KHz a wave through the air from one diaphragm would at least partially cancel out that from the other. In order to avoid this problem, acoustic foam (11) or any other suitable sound absorbing material may be added adjacent to the magnets (3) to damp out or block the highest frequencies travelling between the two diaphragms (2a,2b). Alternatively or additionally, a capacitor connected in parallel with and/or an inductor connected in series with the conductor (ib) facing away from the listener may be used to attenuate the high frequency electrical signal to it. As the signal to the conductor (Ia) facing the listener is always present, the problems normally inherent in a cross-over are much reduced. A further technique which may be used instead of, or in addition to an electrical attenuation circuit is to increase the mass and/or damping of (or associated with) the rear conductor (1 b). This may be done for example by increasing the thickness of the conductor and/or by adding a layer of non-conductive material. It may seem that removing the sound from one of the diaphragms will result in a reduced sound output at high frequencies, but in practice it has been found that there naturally tends to be a rise in output of each diaphragm at these sorts of frequencies, so the resulting frequency response is reasonably flat. If an electrical attenuator with a capacitor in parallel to the rear conductor (1 b) is used, it is also possible to boost the current to the front conductor (la) to increase high frequency output at the expense of lower electrical impedance at high frequencies.

Figure 5 shows that further foam lined damping cavities (5e, 5f) may be added so that each diaphragm (2a, 2b) has a damping cavity on both sides of it.

The direction of magnetic force in a conductor is in a direction perpendicular to both the current and the magnetic flux. This means that to move a diaphragm back and forth with a current in a vertical conductor, the magnetic flux must be travelling largely horizontally in the plane of the diaphragm at least where the conductor is. In a ribbon loudspeaker, a horseshoe shaped magnetic path (12) as shown in figure 6 is used to provide magnetic flux lines (13a) in the correct direction, but this is only practical for a relatively narrow gap or the magnets become massive, so it is not practical to fit a wide diaphragm between the poles of such a magnet. In "quasi-ribbon" or "orthodynamic" drivers, the magnets are placed as shown in figure 7. The magnet poles face the diaphragm and are either side of the conductor, giving flux lines (13b) which are curved, but approximately in the plane of the diaphragm (2a) near the conductor. The magnets alternate with those on one side of a conductor (la) having North poles facing the diaphragm and the ones on the other side having South poles facing the diaphragm. A perforated steel magnetic return path (14) is provided to complete the magnetic circuit. It is possible to use such a configuration with this invention, but the relatively wide structure can cause diffraction of the higher frequencies.

With the advent of high flux rare earth magnet materials such as Neodymium Iron Boron, it is possible to mount a relatively narrow magnet in front of or behind the diaphragm as shown in figure 8, to give flux lines (1 3c) through the conductor without the need for a separate magnetic return path. Sound is able to travel around such a narrow structure with relatively little diffraction. A problem with such an arrangement however is that the flux lines (1 3c) concentrate as the conductor gets nearer to the magnet, increasing the force on the conductor for a given current. As shown in figure 8, in order to maximise the amount of flux interacting with the conductors when they are distant from the magnets, it is advantageous for the width of the conductors to go beyond the width of the magnets. In order to counteract the increase in force when the conductors approach the magnets, the conductor can be divided into two or more strips (ic, Id) with a gap (15) between them.

This means that as the gap (15) approaches the magnet, the flux within this gap does not generate force, with this "lost force" increasing as the gap (15) gets closer to the magnet.

As the flux lines (1 3c) are curved, there is a bending force on the conductors, so it can be advantageous to link them together with a bridge (16) of insulating material (e.g. Mylar) to stiffen and/or damp the conductors. Such insulation can also help to protect the conductors from short-circuit and the user from the signal voltages. In order to improve linearity further, magnets can be disposed on both sides of the diaphragm (2a) (and conductor ic, ld) as shown in figure 9.

In the magnet arrangements shown in figures 8 and 9, the current in both conductors (ic, id) must be in the same direction (i.e. either both in to the page or both out of it in the orientation shown). Usually the input connectors to the loudspeaker will be at the bottom of the loudspeaker, so one or more return current paths are needed, which may conveniently be through one or both of the magnet bars, or a conductor mounted to them.

In order for a loudspeaker to have an electrical impedance in the 4-8 ohm range (which most amplifiers are nominally designed for) whilst having thick enough conductors to be robust, it is advantageous to have several conductors in series. To avoid needing to make multiple return connections, it is convenient to have the return conductors on the diaphragm and preferably generating additional force on it. As shown in figure 10, to achieve this, additional magnets with opposing poles may be added such than one set of conductors (Ic, id) can have current in the opposite direction to the other set (le, if).

These can be put onto the diaphragm in either the form of a rectangular spiral (17) or two U shapes (18a, 18b). The opposing poles mean that the magnets push each other apart.

To hold them at the correct spacing, ferromagnetic (e.g. steel) strips (19) may be used, separated by and fixed to non-magnetic (e.g. brass or austenitic stainless steel) spacers (20). The magnets are then attracted to the strips (19) more than they are pushed apart by each other. To stop the magnets sliding on the strips (19), it is usually necessary to glue them in place (for example with anaerobic adhesive).

As shown in figure 11, in order to fine-tune the frequency response and dispersion of high frequencies, additional mass (21a, 21b) and/or additional conductors (21a, 21b) may be added adjacent to the outermost conductors (ic, If). These additions may also incorporate damping materials such as flexible adhesives. An example construction could be a layer of adhesive (e.g. silicone or acrylic), a layer of conductor (e.g. copper or aluminium) another layer of adhesive (e.g. silicone or acrylic) and a layer of insulator (e.g. Mylar).

A disadvantage of the magnet arrangements in figures 8-1 1 is that a lot of the flux from the magnets misses the conductors. As shown in figures 12 and 13, to improve this situation two diaphragms (2a, 2b) can be used (in similar ways to figures 4 and 5) with all of the conductor arrangements described previously. As before, acoustic foam (11) or any suitable sound absorbing material may be used to prevent the high frequencies transmitting between the two diaphragms.

Figure 14 illustrates how to calculate the optimum conductor widths and positions for the magnet arrangement of figure 13. The top graph shows a series of lines (22a-22t) representing the cumulative sum of the magnetic flux density times length in the plane of the diaphragm (2a) starting from the central axis. This is also representative of the force that a conductor with a certain current density placed within this magnetic flux would generate perpendicular to the diaphragm. The thickest solid line (22a) represents the diaphragm when closest to the magnet and the thinnest dashed line (22t) represents the diaphragm when furthest away from the magnet. It can be seen that the distribution of force varies significantly with the position of the diaphragm. For a relatively wide conductor that catches as much of the flux as possible, the force would be much higher when the diaphragm is closest to the magnet (line 22a).

The bottom graph shows what happens if the forces able to be generated on both diaphragms (2a, 2b) are added together. For example line 23a is the sum of lines 22a and 22t and line 23j is the sum of lines 22j and 22k. Even if a full width conductor was chosen, it can be seen that variation in force (24) can be reduced to about 5-8%. By choosing suitable positions and widths of the conductors (1 g, 1 h) however it is possible to reduce this variation to negligible levels. The position and width of conductor Ig is chosen such that force Fl on line 23a is equal to force F2 on line 23j. The position and width of conductor lh is chosen such that force F3 on line 23a is equal to force F4 on line 23j.

There are many combinations of positions and widths of conductor for which this will be the case, so the aim should also be to maximise the forces achieved whilst avoiding excessively wide conductors, or gaps between them that are too small to insulate the conductors adequately from each other. By making the forces on the conductors equal at the middle and the extremes of the diaphragm movement, the forces at positions in between are also substantially equal and the harmonic distortion generated by large diaphragm excursions is minimised.

Figure 15 illustrates some improvements to the basic constructions shown so far: The diaphragm and conductors may be corrugated (25) in the region of the conductors.

This serves two functions. One function is to increase the resistance to bending so that the variable force distributions as shown for example in figure 14 do not cause buckling of the conductors. The other function is to make the conductors much more elastic along their length so that large diaphragm excursions can occur without being constrained by the conductor stiffness.

Intermediate support braces (26) may be placed at positions along the length of the magnet bars to minimise vibrations in the magnet bars. These braces may be shaped with an aerofoil like section as shown to minimise disturbances to the sound waves. These may be made from materials such as wood, metal or composites (e.g. carbon fibre reinforced resin). If made from metal, it should ideally be non-magnetic so as to not disturb the flux from the magnets. Austenitic stainless steel provides a good combination of being non-magnetic, with high stiffness and good appearance.

In order to improve the acoustical transparency of the magnets, they may be staggered to gives gaps (27) between them. By doing this, additional layers (28) of magnets may also be used to increase the possible force without blocking the path for sound waves too much.

Figure 16 illustrates similar improvements to figure 15, but for the case where two diaphragms (2a, 2b) are used. In this case the bracing bars (29) must fit within the space occupied by the magnet bars in order not to limit the diaphragm excursions. To be stiff enough, these would usually need to be made from metal or a stiff (e.g. carbon fibre) composite. With two diaphragms, the magnets do not need to be as acoustically transparent, but the layers (30) of magnets may be staggered like the bricks in a wall and glued together to provide a strong and stiff structure, whilst using easily available sizes of magnet.

Reducing the reflected wave as described in relation to figures 2-5 is not enough to give the loudspeaker a smooth frequency response. If the wave speed in and width of the diaphragm is constant, there will be peaks to the frequency response whenever an odd number of half waves fit into the width and corresponding dips when an even number of half waves fit into the width. Figure 17 illustrates a solution to this problem for a loudspeaker of constant width. The tension in the diaphragm is varied throughout its height, so that the speed of sound in it is not constant. In this case, the wave speed in the middle is half of that at the ends. If the loudspeaker is driven at its fundamental resonant frequency at the ends such that half a wave (31) fits within the width, a full wave (32) will fit within the width in the middle. This means that the ends are producing maximum sound (resonance) whilst the middle is producing minimal sound (anti-resonance). At points in between, an intermediate wave (33) between half a wave and full wave fits within the width, giving an intermediate sound level. At all frequencies above this, there will be some parts of the diaphragm in resonance and some parts in anti-resonance, giving a smooth overall frequency response. With a diaphragm of varying tension, the sound waves tend to get bent towards the region of lowest tension, so to avoid this region dominating the frequency response it is advantageous if this is tuned with its fundamental resonance below the normal audible frequency range. For example the ends could be tensioned with a fundamental resonance of 20Hz and the middle could be tuned to a fundamental resonance of 10Hz.

One method of achieving such a variable tension is to initially tension the diaphragm equally with the frame fixed in one position and then to subsequently reduce the width of the frame slightly in the middle to reduce the tension here. Some polymers (such as polyolefin) can have a heat shrinking property, so it is possible to achieve the equal initial tension by applying a fixed shrinking temperature (of the order of 60-90°C) to the diaphragm. The shrinkage varies with temperature, so it is also possible to apply a variable temperature throughout the height of the diaphragm.

Adjusting the tension throughout the height of the diaphragm is not easy to do accurately and as described previously can cause bending of the sound waves into the region of lowest tension. An alternative shown in figure 18 is to vary the effective width of the diaphragm by varying the width of the aperture that the sound comes through. The simplest form is a straight tapered edge (34), but a more even frequency distribution is achieved with a curve (35) where the taper is proportional to the aperture width. Discrete steps (36) in the width can also give a reasonably smooth frequency response if dimensioned so that their main resonances don't coincide with each other. Multiple curves (37) can also be employed if desired. The actual width of the diaphragm within the aperture may be kept constant such that residual reflections (particularly of the lowest frequencies) off the edge support of the diaphragm are reflected back in the direction they came from, minimising the problems of vertical components of waves described previously in relation to the prior art. The lowest frequencies in the diaphragm are the least damped by the cavities at the edges because a lower proportion of a wavelength fits within the depth of the cavity. It is therefore possible to tune the diaphragm to resonate to a certain extent at a very low frequency without causing too much resonance or an audible echo at higher frequencies. This can be desirable because a dipole speaker of finite width reduces in efficiency at bass frequencies as air moves around the edges in the opposite direction to the diaphragm movement. This lightly damped resonance can be used to extend the bass response of the speaker to a lower frequency than would otherwise be achievable. In addition to the aperture, the outer boundary of the speaker may also be made of non-constant width, such as the dashed form (38) shown on figure 18 in order to spread out any resonances caused by the sound waves propagating through the air when they meet this boundary.

In order to minimise excitement of the floor to ceiling resonances and anti-resonances inherent in listening rooms with the ceiling parallel to the floor and in order to minimise sound reflections off the ceiling, it is advantageous that the loudspeaker occupies most of the height of the listening room. In order to give stability to such a tall loudspeaker without the need for a large footprint, the speaker may be braced against the ceiling with a pad (39) attached to a screw jack (40). The pad may for instance be made from a material such as foam rubber to minimise transmission of vibration into the ceiling. Of course a shorter speaker is also possible supported by any of the conventional stand arrangements.

Claims (21)

1. CLAIMS1. A loudspeaker comprising: a substantially planar, flexible diaphragm drive means for driving the diaphragm, the drive means comprising a linear conductor operably connected to the diaphragm and a magnet arrangement mounted in proximity to the linear conductor and arranged to provide a force that is substantially perpendicular to the surface of the diaphragm and substantially proportional to a current within the linear conductor; damping means arranged around the periphery of the diaphragm to damp waves arriving at the periphery from the drive means wherein the diaphragm is tensioned such that the speed of sound within the diaphragm is lower than the speed of sound in air.
2. 2. A loudspeaker as claimed in Claim 1, wherein the effective width of the diaphragm varies continuously throughout the audio frequency range.
3. 3. A loudspeaker as claimed in Claim 2, wherein the width of the diaphragm is selected such that for the speed of sound through the diaphragm the width > half a wavelength of a 20Hz wave generated by the drive means.
4. 4. A loudspeaker as claimed in any preceding claim, wherein the damping means comprises an air cavity at the edge of the diaphragm lined with a porous damping material.
5. 5. A loudspeaker as claimed in Claim 4, wherein the air cavity is tapered.
6. 6. A loudspeaker as claimed in Claim 4 or 5, wherein the porous damping material within the cavity is tapered or stepped.
7. 7. A loudspeaker as claimed in any one of Claims 4 to 6, wherein the porous damping material is an acoustic foam.
8. 8. A loudspeaker as claimed in any preceding claim, wherein the magnet arrangement comprises rare earth magnetic materials.
9. 9. A loudspeaker as claimed in any preceding claim, wherein the magnetic arrangement comprises a bar of magnets substantially aligned with the linear conductor.
10. 10. A loudspeaker as claimed in Claim 9, wherein the bar of magnets comprises a plurality of staggered magnets running the height of the diaphragm.
11. 11. A loudspeaker as claimed in Claim 9 or 10, comprising one or more intermediate support braces placed along the length of the magnet bars
12. 12. A loudspeaker as claimed in any preceding claim, wherein the conductor is corrugated and the diaphragm is corrugated in the region of the conductor
13. 13. A loudspeaker as claimed in any preceding claim, wherein the tension in the diaphragm is varied throughout its height
14. 14. A loudspeaker as claimed in any preceding claim, comprising support structures for bracing the loudspeaker between the floor and ceiling of a room.
15. 15. A loudspeaker as claimed in any preceding claim, comprising two magnetic arrangements arranged either side of the diaphragm.
16. 16. A loudspeaker as claimed in any one of Claims 1 to 14, comprising two diaphragms arranged in parallel with the drive means disposed in between.
17. 17. A loudspeaker as claimed in any preceding claim, wherein the conductor and magnetic arrangement are arranged such that the forces on the conductor at the middle and extremes of diaphragm movement are substantially equal.
18. 18. A loudspeaker as claimed in any preceding claim, further comprising more than one linear conductor
19. 19. A loudspeaker as claimed in Claim 18, wherein each linear conductor is associated with its own magnetic arrangement.
20. 20. A loudspeaker as claimed in Claim 18 or 19, wherein the current through adjacent conductors travels in opposite directions
21. 21. A loudspeaker as claimed in any preceding claim, wherein the diaphragm, drive means and damping means are housed within a casing, the casing having a substantially vertical aperture that varies in width along the height of the casing.