[Aether Audio]

The following s an engineering white paper that will soon be available on our website.  Please pardon the length but a thorough examination of the subject requires such.  We hope it will be of some value to those interested in the claims made by advocates of this technology.

Thanks - Bob


Coincident Driver Technology:
Is This A Superior Development In Loudspeaker Design?


OJECTIVE:

Advocates of currently available coincident driver technology suggest that it is inherently superior to that of the more traditional non-coincident approach.  The claim of true “point source” radiation characteristics is offered as the primary, if not the only reason for their advocacy.  We intend to show that any such benefit can be emulated to a great degree by the proper implementation of the traditional, non-coincident method and that by doing so, the many inherent drawbacks of the coincident technique can be avoided.


BACKGROUND:

From the very beginning of high fidelity sound reproduction, engineers and enthusiasts alike have mused over and elaborated upon the advantages and desirability of the theoretical “point source” loudspeaker.  Numerous attempts have been and continue to be made in the quest for such a device.  Although the continued effort to develop and improve this technology has some merit, when properly implemented, the present state-of-the-art utilizing non-coincident transducers is sufficient to obviate the need for such heroic attempts in virtually all practical applications.


DEFINITIONS AND EXCLUSIONS:

Due to the diverse number of different driver technologies, there exist several that may be initially viewed as exhibiting some form of coincident behavior.  In order to limit this paper to a reasonable size, we are forced to exclude all but that class which most succinctly fits the traditional definition of “coincident.”  Please note that these exclusions are not intended to imply that the excluded typess do not exhibit similar properties to those covered under this analysis.  The following paragraphs have been devoted to a brief summary of the most prevalent sub classes of these excluded designs as well as that of the true coincident device.


Electrostatic Planer Drivers:

Electro-static drivers rely on a transformation of a large electrical potential into an Electrostatic field.  This field is then dispersed over the surface of a relatively large and lightweight membrane and becomes the driving force employed to develop mechanical motion in it.  None of these, save one that we are aware of, can be considered remotely coincident in behavior.  In fact, a more succinct classification of the Electrostatic design would be that of “Line-Source.”  Seeing that Line-Source operation is distinctly different from that of a “Point-Source” or coincident design, we shall not elaborate further as it clearly lies outside the scope and intent of this analysis.


Electromagnetic Planer or “Ribbon” Drivers:

The Ribbon type driver is similar in operation to that of the Electrostatic, the primary difference being how the membrane is driven.  One type Ribbon driver employs a technique of “etching” a conductive path of thin metal upon a lightweight, non-conductive membrane.  This membrane is then stretched in front of a magnetic assembly and amplifier current is passed through the conductive path of the membrane.  The resulting fluctuations of the developed magnetic field induce mechanical motion of the membrane.

The other main type of Ribbon driver uses a membrane of very thin metal and amplifier current is passed through the entirety of it.  Again, the resultant fluctuations of the developed magnetic field drive the membrane’s motion.  Regardless of type, Ribbon drivers fall within the classification of Line Source operation just as Electrostatic drivers and will not be included in our analysis.

Electrodynamic Drivers:

Given the above failure of both Electrostatic and Ribbon drivers to fit the true definition of coincident or Point Source operation, we shall limit our discussion to those that reside in the class of Electrodynamic operation.


Lowther Drivers:

Within the dynamic class of coincident driver technology there are essentially two sub classes.  One of these is what is referred to as the “Lowther” driver. One could consider this design as a refinement of the first original dynamic cone driver.  It is primarily a single cone device that has been specifically engineered to be operated over the full audio range.  It typically incorporates a secondary “whizzer”cone of smaller dimension that is mechanically attached to and centrally located on the main cone assembly.  The entire assembly then is driven by means of a single voice coil and consequently, has no need for any form of traditional crossover filter network.  This type of device relies upon a form of mechanical crossover action that is the result of a mechanical de-coupling of the smaller cone from that of the larger unit.  It can be considered as an “engineered in” form of cone “beak-up mode” operation.

Over the last several decades, the Lowther design has continually been refined  and has amassed a moderately large number of ardent advocates.  Some recent designs have actually attained a fairly high level of performance.  Never the less, the design is fundamentally characterized by a significantly limited bandwidth and dynamic range capability along with poor high frequency dispersion performance.  It also suffers from generally higher levels of static T.H.D. and IM distortion by-products.  Because of these issues, it will undoubtedly remain a relatively minor contender in the battle of loudspeaker technologies.  Do to its significant limitations and out of a sense of fairness, we have chosen not to include it in our analysis here.


The True Coincident Driver:

Finally, we arrive at the definition of the true coincident driver.  The dominant characteristic of this driver is that it employs the use of two distinctly separate driven elements.  The most common combination of elements are a larger cone used for reproduction of lower frequencies and some type of smaller dome for the reproduction of high frequencies.  This dome is centrally located at the middle of the larger cone and its primary axis of sound propagation is parallel to that of the cone.  Both elements are driven independently of one another and therefore require some type of external crossover filter network in order to direct amplifier current of given frequency to its appropriate element.  These independently driven elements each possess their own individual voice coils and may or may not share the same magnetic assembly.  If they do not, then some method of mechanical support for the high frequency assembly is required in order to suspend it safely in front of the larger cone while avoiding mechanical contact with it.

It wasn’t long after dynamic drivers began their meteoric rise in the early days of high fidelity that the first coincident drivers were developed.  One of the earliest and most successful of them was the renowned Altec 604 which is still in production to this day.  The high frequency section is composed of a compression driver based diaphragm assembly attached to a sectoral horn which projects outward from the center of the surrounding low frequency cone.  A common magnet structure is used to provide the static magnetic field needed to drive both voice coils in this design.  Obviously, the 604 along with virtually all others of its type require an external crossover network for directing signals to the appropriate section.  Seeing the design is still in production, it represents one of the more successful designs of this type.  Given its early introduction to the audio market, one can easily see that the more recent designs cannot lay much claim to any suggestion of originality.  As we shall see in the following analysis, many modern attempts to improve upon its design have actually exchanged some weaknesses for others of greater severity.


ADVANTAGES OF COINCIDENT TECHNOLOGY:

At the risk of being redundant, we shall examine the potential advantages offered by coincident driver technology.  The most obvious as well as main performance advantage is that all radiated sound along the primary axis appears to emanate from a single point in space at all operational frequencies.  The consequence of this effect is that the radiated energy has a homogenous and symmetrical dispersion characteristic in all planes radial to the primary axis of propagation.

In practice, a coincident loudspeaker will exhibit the same frequency magnitude and phase response variations as one takes measurements at all equal angles away from the primary axis.  This characteristic is desirable in that the resultant reflected energy of the reverberant sound field will tend to exhibit a frequency magnitude response that is more similar to the direct energy radiated from the primary axis of the loudspeaker.  This is particularly important for the purpose of presenting a cohesive sound field to the listener.  A much more integrated form of reproduction of the recorded material results and the “loudspeaker in a room” effect is reduced.

Another advantage is that that the dispersion pattern of the high frequency drive mechanism integrates with the low frequency section to produce a concentric and uniformly over-lapping “circle of coverage” that facilitates ease of placement and predictable performance. Systems offering coincident drivers will often out-perform many otherwise equivalent, non-coincident systems in these areas of performance.

From at least a theoretical standpoint, the technology also offers the potential to improve inter-driver signal alignment.  If design geometry permits, the high frequency drive element can be optimally placed nearer the acoustic center of the low frequency section.  Combining this with a phase linear crossover network may result in a system that exhibits linear or nearly linear phase response and group delay characteristics throughout the crossover region and beyond.  While there is continuing debate as to the importance and audibility of these parameters, eliminating them from the complete list of potential errors is a worthwhile advantage that obviates the need for further concern.

Finally, there is the potential for significant system cost saving’s.  In practice, a smaller and simpler enclosure would be required for a coincident driver as it represents a self contained single-driver assembly.  Also, driver manufacturing costs could be reduced by taking advantage of the low frequency section’s magnetic system.  A single assembly would reduce material costs, require less factory assembly space  and reduce inventory requirements as compared to comparable non-coincident system production.  All of this could result in a better price/performance ratio for the end consumer as well as a larger profit margin for the manufacturer.


DISADVANTAGES OF COINCIDENT TECHNOLOGY:

An initial review of coincident driver technology seems to suggest that there would be little reason to deviate from that approach if ultimate system fidelity, ease of placement and cost were the objectives.  Never the less, seeing that coincident drivers have been in use for over 50 years, they have yet to replace the non-coincident approach.  In fact, the vast majority of systems in use both professionally and in consumer systems are still of the non-coincident type.  If cost were the issue, then the coincident type should still predominate due to the reduced manufacturing costs outlined in the previous section.  These facts lead one to the conclusion that, at least for the present, coincident technology appears to fall short in its ability to deliver in practice what theory would suggest is possible.

Upon closer inspection, it is possible to discover the likely reasons for this.  A proper analysis requires the breakdown of the coincident driver into its three primary domains of operation and determining the weaknesses of each.


The Electro-Magnetic Domain:

An Overview of Dynamic Drivers.

At the heart of every driver, regardless of type, lies its mechanism for converting electrical energy from the amplifier to the mechanical motion of its vibrating surface.  All dynamic drivers, coincident or otherwise, rely on a transformation of electrical current supplied by the driving amplifier into an alternating magnetic field.  It is in the realm of this transformation that any non-linearities present the potential to negatively impact performance.  The first area for potential trouble resides in the very nature of the magnetic material which lies at the heart of the overall magnetic structure.  Magnets have improved dramatically in recent years but regardless of chemistry, all have their limitations.  

Even the strongest magnets suffer from a gradual decrease in their  magnetomotive force as operational temperatures approach the Curie point. This is the temperature at which a magnetic material looses all residual magnetism.  This loss is a result of heightened molecular activity disrupting the alignment of the magnetic domains that are the source of its retained magnetic field.  The rate at which a magnetic material experiences this loss of strength varies depending on material type.  Never the less, at some point all such materials begin to experience a reduction of magnetic strength at temperatures common to loudspeaker operational conditions and well below their Curie Point.  Upon a lowering of temperature, the former magnetic field strength returns to, or near to its original level, depending on how closely operational temperatures approached the Curie point as  well as the amount of time such temperature was maintained.

The effect upon loudspeaker performance this “sliding magnetization” has, is manifest as a type of “dynamic compression.”  As magnetic field strength decreases, the slope of the transfer characteristic of the driver decreases.  This results in a non-linear acoustic output level with respect to input electrical drive power that is temperature dependant.  The final temperature attained by the magnetic structure of a loudspeaker is the integral of the input power minus the heat shedding characteristics of the structure, which upon first approximation can be represented as a constant..  Therefore, magnetization levels are directly dependent upon the average input power change over time.  The result is a dynamically changing acoustic output with respect to a dynamically changing  power input.  

Designers try to select magnetic materials that tend to manifest their demagnetization characteristics as far as possible to the extreme high end of their anticipated operational temperatures.  Because of this, the resulting dynamic compression is typically manifest only when higher average power levels are experienced by the driver.  The major caveat to this is that the better materials with regard to this issue are usually more expensive than their lesser counterparts.  One then finds that those drivers that suffer least from dynamic compression effects are also more expensive.

The above change in transfer function induced by temperature dependant non-linearities of the magnetic structure is further influenced by the temperature change incurred within the windings of the voice-coil.  As its average temperature increases, there is a corresponding increase of wire resistance making up the coil.  According to Ohms Law, as resistance increases in a circuit, the current flow and consequentially the power developed in the circuit decreases.  Since cone motion is directly proportional to voice-coil current, any rise in voice-coil temperature will decrease the acoustic output of the driver - assuming all other parameters are held constant.

Once again, the voice-coil’s temperature rise is the integral of input power minus it ability to shed heat.  Since it is buried within the heart of the magnetic structure, its ability to dissipate heat is directly tied to that of the entire magnetic and support assembly.  As one can easily envision, the rise in temperature of a dynamic driver assembly carries a profound potential to degrade dynamic performance.  As a result of both magnetic structure and voice-coil heating, we find that at the operating extreme of electrical power input, a point is reached wherein virtually no increase of acoustic power output occurs.

As if the above outlined parameters were not of sufficient challenge to the designer, the following problem adds another dimension of difficulty to overcome.  Hysteresis characteristics are the bane of all dynamic loudspeakers.  Regardless of type, such hysteresis artifacts effect all of the various magnetic materials used in loudspeakers and are not constant, but vary with temperature as well.  Hysteresis arises as a natural artifact of ferromagnetic action and is dependent upon the specific formulation used in the design.  It is manifest as a type of magnetic “memory” in a given sample such that the force required to reverse the dynamically induced flux field is greater than the force required to induce it.  If of sufficient severity, an offset of the voice-coil from its static rest position can occur.  Usually though, this effect is only encountered at higher input power levels and at lower drive frequencies.  Never the less, hysteresis effects are a dominant source of even-ordered harmonic and inter-modulation distortion, and are manifest long before the high power levels required to induce voice-coil offset.

As a final challenge to the designer is the need to achieve magnetic field linearity within the region of the magnetic assembly’s voice-coil gap.  This gap is usually shaped as an annular ring opening formed by an outer ring magnet and an inner ferrous pole-piece.  In its simplest form, the magnetic circuit is bridged from the magnet to the pole-piece by a ferrous back plate at the rear of the driver.  This back plate is also the mechanical support for the centered pole-piece.  In order to complete the magnetic circuit, flux lines are forced to pass through the voice-coil gap in order to bridge between the ring magnet and the pole-piece.  This occurs in the region where the voice-coil is located when at rest.  These flux lines then represent the static magnetic field that drives the voice-coil/cone assembly when current is forced to flow through the coil.

In the basic device described above, the magnetic field developed is generally asymmetric in form.  This is due to the air gap required for the voice-coil’s operation.  This asymmetry is the result of the fact that the magnetic permeability of the back plate is much greater than that of the air in the gap.  As a consequence, the flux lines are tightly confined within the area of the gap but stray outward towards the front of the driver in front of the gap.  This stray field represents a decreased driving force encountered by the voice-coil as it experiences outward excursions as compared to inward ones.  This too is a significant source of voice-coil offset at high drive levels as well as even ordered-harmonic and inter-modulation distortion.  Various methods including shorting rings, double magnets and such have been developed to reduce or even eliminate this effect.  These efforts have been relatively successful but represent another necessary evil in overcoming the typical shortcomings of the dynamic driver.


The Case of the Coincident Driver:

As one can clearly see from the issues outlined above, the electromagnetic domain of the typical dynamic driver is plagued with many potential sources of non-linearity and distortion.  Extrapolating from that, it becomes obvious  that if we are to optimize a design, each driver should be treated as an independent device.  Depending upon whether its intended use is for low, mid or high frequency reproduction, the geometries required for one device will necessarily preclude optimization for those intended to cover a different frequency range.  This fact alone makes the design of a coincident driver one of potentially significant compromise.

Aside from these issues, the coincident driver represents an even greater challenge due to the requirement of physically locating the motor system of the high frequency transducer within or very near to that of the low frequency section.  Because of the high drive currents typically experienced by the low frequency section, the risk of “cross modulation” from it to the high frequency section is quite high.  A type of transformer action due to inductive coupling between voice-coil/magnetic circuits is almost unavoidable, even when attempts at magnetic shielding between sections are employed.  The byproduct of such induced coupling is a type of complex inter-modulation distortion that manifests its greatest effect at frequencies in the vicinity of crossover.

We find that there are several other major caveats in the electromagnetic domain of the typical coincident driver.  The issues of hysteresis byproducts and dynamic compression due to thermal modulation of the transfer function, as well as thermal modulation of hysteresis effects, carry even greater potential to degrade performance in the coincident design.  Due to the far greater energy requirements for low frequency reproduction, that section of a coincident driver will generate more waste heat.  Again, magnetic structure heating due to low frequency power dissipation will unavoidably affect the high frequency section because of their common magnetic assembly.  The potential for high frequency dynamic compression and distortion due to hysteresis within the magnetic assembly is far greater than in the non-coincident design and a major obstacle to high fidelity performance.

In summary, we find the electromagnetic implementation of a coincident driver design is fraught with obstacles and potential for compromise with respect to the typical non-coincident design.  Various coincident designs from several different companies over the years have had varying degrees of success overcoming these issues.  Never the less, these designs all manifest compromises that are not required in the non-coincident design.  As we can see, the issues of the electromagnetic domain alone give sufficient cause to avoid the coincident approach.


The Mechanical Domain:

An Overview of Dynamic Drivers:

The mechanical construction of all dynamic drivers comprises four basic subsystems.  These are the magnet/frame assembly, the voice-coil/bobbin assembly, the cone assembly and the suspension system.  These four are usually combined and treated as one system in the design phase of any driver.  Together, they determine the final performance of any dynamic driver.  A process of delicately balancing the different electrical, magnetic and mechanical parameters is required in order to achieve an optimization over the intended range of use.  A relatively small change in any one of these parameters can result in the need for a significant change in any or all of the others.  Seeing that the mechanical parameters have a direct influence upon all, any change in geometry or material can profoundly impact the final results.  

The bandwidth, power handling, damping, sensitivity, directivity, accuracy and reliability of the final device are all the results of ingenuity and compromise in the art of design.  Such ingenuity and compromise have no greater area of influence than in the mechanical domain of dynamic drivers.  In fact, this area lends itself to a greater degree of manipulation and influence over the other parameters than any other domain.  More than anywhere else, it is the mechanical domain wherein the designer is permitted the greatest degree of creativity.  Seeing that it affects all other parameters, it is also where any attempt to stray from optimal design for the purpose of expanding the range of one parameter runs the greatest risk to overall performance.


The Case of the Coincident Driver:

As we have outlined above, the mechanical domain of the coincident driver is where the designer has the most freedom to experiment.  As also outlined, any attempt to extend the operational range of one parameter increases the likelihood of degrading the performance of the rest.  Seeing that the coincident design is an attempt to extend bandwidth to cover the entire audible range, compromise of most, if not all remaining parameters is fundamentally unavoidable.  Although the intent of this paper is not that of a primer on coincident design, we will attempt to give some generalized examples in order to clarify the matter.

In order to provide sufficient space to accommodate the high frequency section, the voice-coil diameter of the low frequency section may require that it be made larger than would be otherwise optimal.  This carries the potential to increase the voice-coil’s resulting mass and therefore lower efficiency.  To reduce this effect the coil may be made shorter.  The drawback to this is that a shorter coil stroke will result.  To compensate for the shorter voice-coil, a larger cone may be employed in order to avoid the loss of low frequency extension.  A larger cone could result in more mass which again reduces efficiency.  A lighter cone material may be then chosen to compensate for the added cone size.  This lighter cone material may then exhibit a greater number and intensity of undesirable break-up modes which then lead to increased distortion.  On and on the compromises and adjustments must be made in order to come to a workable design.  In the end a functional device may result, but in its wake, a series of significant compromises to optimized performance parameters will undoubtedly remain.


The Acoustic Domain:

An Overview of Dynamic Drivers:

The characteristics defining the energy dispersion properties of a given dynamic driver are governed by the laws of the acoustic domain.  The driver’s cone parameters determine these characteristics and all cones are dominated by three primary variables that affect such dispersion.  The first of these is the driver’s effective cone diameter.  All devices exhibit an omni-directional radiating pattern when producing wavelengths of large dimension with respect to that of the diameter of the radiating surface.  As the operational frequency  is increased from the omni-directional point, then the resultant wavelength is decreased.  When the generated wavelength begins to approach a smaller ratio with respect to cone diameter, a narrowing of the dispersion pattern will begin to be observed.  If the cone is circular in shape, then this narrowing will be axis-symmetric.  If the cone is of any other shape, then such narrowing will be determined by the diameter of the cone that is in parallel to the plane of dispersion.  This process of narrowing continues until the driver reaches the upper limit of its operational bandwidth.

Strictly speaking, the above example of the cone diameter’s effect on dispersion only holds true if one case of another primary variable is encountered.  This leads us to the definition of the second variable, which is cone rigidity.  If the process of dispersion narrowing is to continue seamlessly until the upper bandwidth limit is reached, then the cone must exhibit the property of a “perfect piston.”  This property is a quite simple concept and implies that the only motion experienced by the cone is a direct result of voice-coil action.  Perfect piston behavior is characterized by a lack of secondary cone flexure that would otherwise occur if the cone material does not display sufficient rigidity to prevent it.

Virtually all cone drivers exhibit some form of secondary flexure, whether intentionally designed to or not.  This flexure is a byproduct of cone material type and mechanical design.  In the second case of cone rigidity, manufacturers have taken advantage of this secondary behavior in order to extend driver bandwidth.  If properly done, a controlled form of break-up can be achieved wherein the center of the cone begins to operate somewhat independently from the rest of the surface.  This independent behavior increases with increased frequency until some upper absolute limit is reached.  

It is by this method that many lower frequency drivers are able to operate up to a sufficiently high frequency to be crossed over to the typical high frequency driver.  At the frequency where this independent behavior begins to manifest, whatever narrowing of dispersion that had begun to take place will be reduced or even halted.  In some instances, a widening of dispersion will even be observed as the effective cone diameter becomes smaller with respect to the reproduced wavelength.  In any case, this type of operation constitutes a tradeoff with respect to distortion.  Although this technique is useful for the purpose of extending high frequency bandwidth as well as reducing driver “beaming” effects, this same ability to flex also makes such devices prone to a form of inherent and undesirable flexure.  This type of flexure produces audible byproducts that are very complex and non-harmonically related to the source signal.  As such, they constitute a significant source of increased distortion levels.  Never the less, this method is employed successfully by the majority of manufacturers and dominates the more traditional, non-coincident as well as coincident driver topologies.

The third variable dominating driver radiation characteristics it that of radiation impedance.  If traditional, baffle mounted drivers are employed, then this radiation impedance is dominated by the properties of air and the geometry of the cone/diaphragm.  As stated above, the case of cone diameter is a common example of a special case where all other variables are held constant and the only difference permitted is that of a variation of one cone dimension.

If, on the other hand, a device of sorts is used to assist in an acoustic transformation of radiation impedance, then a variety of parameters may be influenced.  Such of a device as is of common usage is that of the typical horn or waveguide device.  Any of these type devices possess the potential for affecting and/or controlling the dispersion properties of a dynamic driver and have been successfully used to do so for many years.  They also offer the potential for increased efficiency, lower distortion and extended bandwidth if the associated parameters are adjusted accordingly.  In fact, there are few instances where a dynamic driver will not benefit significantly from a properly matched acoustic transformation device.


The Case of the Coincident Driver:

As can be seen from the above outline, the acoustic domain has a profound potential for affecting the final response of any dynamic driver.  All of these effects must also be taken into consideration when designing a coincident device.  Once again, any method developed for the optimization of a non-coincident driver may require mild to sever modification if application to a coincident design is attempted.  In many cases, such methods may not be possible to implement at all.

A somewhat recent variation of the coincident driver attempts to make use of the low frequency cone’s natural flare.  In this type of design, it is utilized as a type of extension to the high frequency driver, which itself may or may not have a short horn assembly loading the diaphragm.  The low frequency cone essentially becomes a type of waveguide which continues to load and direct the wave front emanating from the high frequency section.  Upon an initial analysis, this would seem to be an ideal and synergistic way to take advantage of the low frequency cone’s presence.  While this technique may offer some form of high frequency dispersion control, it also introduces a potential source of significant distortion.

As far back as 1976, Paul W. Klipsch was able to demonstrate that non-coincident drivers placed in close proximity to one another and operating over different frequency bands will interact in the acoustic domain such as to generate undesirable cross inter-modulation byproducts.  It only stands to reason that the closer two sound sources are placed together, the more this effect will be manifested.  The coincident driver represents a “worst case scenario” with regard to such placement as its very design is based on the concept of locating one sound source within the center of another.  It does not require much imagination to visualize the vibrations of the low frequency cone “pushing and pulling” on the wave front generated by the high frequency section as it passes outward from and along the length of this cone.  As such, the coincident driver is probably the worst possible source of acoustically born inter-modulation distortion.


VERDICT OF THE COINCIDENT DRIVER CASE:

After  reviewing the many issues governing dynamic driver performance, it becomes clear that in order to optimize their design one needs to treat them as independently as possible.  Combining two or more such devices so as to create a form of “hybrid,” as in the form of the coincident type, clearly becomes an exercise in extreme compromise.  This fact alone supplies sufficient reason for their lack of generalized use and acceptance by the audio marketplace.  If they truly offered all of the advantages their proponents claim without incurring these compromises, then their very dominance alone would provide sufficient evidence for their case.  But as we can see, even though their introduction to the audio market goes back for over 50 years, they still lay claim to a relatively small percentage of total market sales.  

This fact becomes even more solidified by the creative development of non-coincident systems that offer the lion’s share of advantages promoted by coincident types while avoiding virtually all of their compromises.  We intend to outline one such approach in the following section in such a way as to show that resorting to a coincident topology becomes totally unnecessary for achieving its goals in the majority of practical applications.

Seeing that in any foreseeable future, the laws of physics are unlikely to yield to the desires of designers and marketing personnel engaged in the development and promotion of coincident systems, the coincident driver is destined to remain a relatively minor player in the contest for market share.  As long as fidelity of reproduced sound is the driving force behind design, the coincident driver will undoubtedly be limited to applications of convenience or cost conservation.


THE NON-COINCIDENT ALTERNATIVE:

The design objectives of the coincident driver represent a desirable and worthwhile combination of advantages.  Just because the actual implementation of such a device involves significant compromise, it does not mean its design objectives should be abandoned or are unattainable.  In fact, if one applies the correct combination of technologies and performance parameters, it is possible to achieve the majority of coincident technology’s design goals in most practical situations by a creative application of existing, non-coincident technology.

As any good audio engineering textbook will point out, the best compromise to the “single driver for all frequencies” concept is an optimally designed two-way system.  An optimized combination of two drivers covering the entire audio range and employing a single crossover frequency represents the least compromising and most efficient way to achieve near ideal performance.

One requirement of such a system would be that any radiation lobing errors be reduced to a minimum.  One way to achieve this would be to ensure that the effective inter-driver spacing be small with respect to the wavelength of the crossover frequency.  In order to achieve this, some means of extending the high frequency driver’s lower bandwidth would be in order.  The longer resulting wavelength from employing the lowest possible crossover frequency would relax the inter-driver spacing requirement and thereby permit the use of more practical devices and implementations.

High frequency dispersion control also offers the potential for improved “point-source” type behavior.  By selecting an appropriate waveguide or horn type radiation transformer, both lower frequency performance and controlled dispersion of the high frequency driver can be attained.  Combining this modified high frequency driver with a well designed low frequency driver through a crossover network employing a 4th order Linkwitz-Riley topology will ensure an in-phase condition at crossover and thereby, further improve the system’s point-source characteristics.  This combination of devices and features offers one of the best possible combinations for emulating the coincident device.  The reader may find it interesting to note that this very combination, as represented by a number of variations from different manufacturers in one form or another, has been employed by professional recording engineers and other sound professionals from the earliest days of modern audio engineering.  To be sure, “there is nothing new under the sun.”


CONCLUSIONS:

We have shown that the coincident driver represents an extensive set of compromises in order to achieve its design objectives.  We have also shown that such compromises are not necessary to achieve these goals as long as a proper implementation of existing, non-coincident technology is employed.  In closing, we suggest that any attempts to improve the state of the art should include the early observations made by the founding fathers of modern day audio technology.  Any attempt to significantly stray from the basic foundations laid in those earlier times will probably result in a set of compromises that will not advance the state of the art.

[JLM]

Bob:

I agree with much of what you state, but note:

1.  Many of the points made, such as the limitations of dynamic drivers apply to both coincidental and non-coincidental designs, so I'm wondering why you included those points.

2.  How much heat has to be generated for the deletrious effects to the magnet to be heard?  What kind of volume levels are we talking about here?  Is it a "real world" concern for listeners of classical music that average 75 dB in room playback volume levels?

3.  How do different magnet materials compare as to the affects from heat, etc.?

4.  One speaker design you didn't mention is the single driver that doesn't have a whizzer cone.  Granted, most of these have severe limitations with smaller drivers being unable to produce deep bass and larger ones lacking wide dispersion of high frequencies.  But on the positive side there is no crossover circuitry induced errors, no worry about imaging from multiple sources, no issues regarding how signals are reproduced from two different drivers blend at crossover, and all the benefits from active versus passive amp/driver interaction.  As least one good example of a truely full range whizzerless single driver exists, the Fostex F200A.  It has an 8 inch diameter cone with alnico magnet and is rated 30 - 20,000 Hz, 90 dB/w/m at 8 ohms.

[Aether Audio]

JLM,

1.  The very reason for comparing the limitations of both types of drivers is to point out that, in order to optimize an electrodynamic driver, the frequency range it operates over will inevitably require some reduction.  As in the design of any device, the more you ask it to do, the less effectively it does any one particular thing well.  

A coincident driver must reproduce the entire audio bandwidth, so no single frequency range can be completely optimized.  Take the case of a good subwoofer.  By reducing the frequency range it operates over from 20 to 80 Hz, we're only asking it to cover 2 octaves.  It can be designed to do this very well but then it cannot operate up into the mid and high frequencies.  All materials and costs being similar, in order to extend that same driver's higher operating frequency, we would be forced to "give up" some of the performance advantages in the case before.  Now it has a wider bandwidth but it won't quite belt out the low end as well anymore.  In engineering, we call this the "Gain/Bandwidth Product." (GBP)  The more you try to extend bandwidth, the lower the gain (or some other parameter - i.e. - accuracy) of the device.

These are simple engineering trade offs that go into every product ever designed to some degree.  Of course, as you throw more money at the problem you can minimize some of these compromises.  In the end though, regardless of cost, we run into the brick wall of physical limits and the limitations of present day materials science.  Even money has its limitations.

2. Totally driver dependent.  Impossible to tell unless all electromagnetic and thermal variables are known.  Basically cheap drivers do poorly - good ones do better.

3. Alnico is real good with regard to heat but looses magnetization over time and is not as strong as the rare-earths to begin with.  Cheap ferrite is the worst but then again, if the geometry is optimized such that the structure is vented, it may not suffer too much at all.  The thermal time constant of the assembly is the key here.  If the design sheds excess heat quickly then the magnet won't usually get too hot.  It's when you start pushing hard where we start to see this effect in any decently designed driver.  That's one major area of concern for the coincident design.  Its geometry criteria is based on extending high frequency bandwidth -- not providing superior venting to minimize heat build up.  One almost always negates the other.   Because of that tweeter stuck in the middle, you can't get optimal air flow through the magnet structure.

4. As I stated in the paper, the Lowther types have made progress over the years and I'm sure at moderate levels they can sound pretty darned good - maybe even great.  I've never heard the Fostex unit but I'll bet it is good.  Again, its got to have some limitation or there would be no reason for anyone to build anything else.  I would suppose that would be in its dynamic capabilities.  At some point, it just come down to raw SPL and power handling if you want to cover the entire 90 dB range of human hearing.

Thanks JLM and take care,8)
-Bob