In Part 1 of this article, I described the workings of loudspeaker mechanisms that use the reciprocating (“back-and-forth”) motion of a physical diaphragm to produce sound. I went on to say that where those two elements (reciprocating motion and a physical diaphragm), are present (as in almost every kind of speaker other than the “plasma” types, with no physical diaphragm, and the “continuous motion” types, with no reciprocating motion) the necessary “start-and-stop” motion of the diaphragm as it moves back and forth to produce sound will always be subject to inertia and can never exactly follow the music signal driving it,
This, I said, is because no physical diaphragm can ever accelerate or decelerate to follow its driving signal in zero time; there’s always some inertial lag between changes in the signal and changes in the movement of the diaphragm caused by them. And, because of that, there’s always some inaccuracy (“undershoot” or “overshoot”) in the diaphragm’s motion that affects the quality of the sound.
How much of that there will be depends on, among other things, the volume level at which the speaker is playing — the louder it plays, the more undershoot or overshoot there’s likely to be. That’s because the only way to make more volume from any given speaker is to increase the velocity of its diaphragm (how far it moves in response to each signal change). The effect of doing this will be exponential and will require four times as much control force for every doubling of velocity. A speaker’s control factors, however, tend to act linearly, like conventional springs, and will provide only twice as much force for each doubling of total movement. That’s why the increasing undershoot and overshoot.
Having an exponentially increasing requirement for control but only linearly increasing available control force is obviously a problem at a speaker’s extremes of performance. Perhaps surprisingly, it’s also a problem at every other point, as well.
While amplifier damping energy only operates to oppose back-EMF, which only occurs when the diaphragm is moving other than as directed by the driving signal (undershoot and overshoot), all of the other diaphragm control factors are at work whenever the diaphragm is moving at any volume level and is at anything other than its “at rest” center position. In short, the spider, surround, and trapped air volume, if any, are always providing (except at dead-center) spring force to either resist the movement of the cone away from its rest position or to accelerate it back towards its rest position.
And, because that force is linear while the force necessary to maintain control of the diaphragm varies exponentially, conventional drivers or driver/enclosure combinations always – except at just one single level of movement where the needed force and the force available coincidentally match – either provide too much or too little control to the diaphragm, and the sound is compromised accordingly.
With velocity rather than mass being the principal factor in determining how much force will be necessary to effectively control a loudspeaker’s diaphragm, it might seem that tweeters and not woofers would be the drivers most likely to be troubled by problems of control. After all, in most speaker systems, the tweeter may operate at as much as 100 times the maximum frequency of the woofer. (In a three- or four-way system, for example, if the tweeter “tops out” at 20 kHz and the top frequency of the woofer is 200 Hz, that’s 100 times!). What keep the tweeter from being the problem in most cases are these three factors: 1.) Tweeters almost always have much more powerful motors per unit of driven mass than woofers do, (In cone or dome drivers, this is the mass of the diaphragm plus the mass of the voice coil plus the mass of the moving portions of the spider and surround. In other designs, it may be different or even just the mass of the diaphragm, only). 2) Tweeters move less distance per excursion in producing high frequencies than woofers do in producing bass. Because of this, even though their number of start/travel/stop cycles per second is much higher, their velocity of travel (the distance traveled per excursion times the number of excursions per second) may not be correspondingly greater, even when both are playing at the same volume level. The lesser mass of the tweeter’s moving elements as compared to those of a woofer, will also offset, although just arithmetically, the exponential effect of the difference in velocity, and will serve to lessen the total amount of control force required. 3) The total high-frequency energy content of most music is typically substantially lower than its typical low-frequency content, so, depending on the drivers used, it’s possible that the percentage of available movement actually used by the tweeter may be lower than that of the woofer, which can also serve to lower the tweeter’s velocity, and thus its required control force. (Remember that, in accordance with the formula F = ½ MV2, every lessening of movement (V) will result in a disproportionately greater lessening of the control force required.)
The amount of travel required by any driver – whether a woofer, tweeter, or midrange – to produce any given sound level, can be reduced by simply using more drivers. Remember that, in the final analysis, all any driver is is an air pump, and that, because two identical drivers moving any given distance per excursion will move twice as much air per excursion as any one of the same driver moving the same distance, it’s possible for two drivers to produce the same amount of air movement as one single driver by moving less. This lesser necessary amount of movement means less velocity which, again, means a more than linear lessening of necessary control force. Certainly — especially for bass — more drivers producing the same frequency at any desired sound level will sound more “powerful” and more “natural”, and certainly some part of that may be due simply to having more surface area to produce the sound, but I would contend that a more significant part of the improvement can be credited to a reduction of inertia-related distortion due to the greater control of the drivers made possible by the reduced diaphragm velocity that comes from reduced diaphragm travel.
This same technique – reducing overall travel by moving the same amount of air with a greater total diaphragm area – finds its best expression in panel speakers, like the planar magnetics (Magnepan, Eminent Technology, etc.) and electrostatics (Martin-Logan, Quad, Acoustat, and others). As certain multi-tweeter designs have shown, it can also be used to effect better control of high frequency drivers, but other difficulties, notably “comb filter effects” or “beaminess” (primarily from panel speakers, which tend to be problematic only in the upper frequencies), can limit or even nullify whatever improvement might be gained.
Besides limiting travel, progressive diaphragm suspension (spider and surround) could also be used to make more well-controlled (and thus presumably better-sounding) drivers, the problem with this, as with enclosure-related loading techniques, is that any such control technique would likely be less progressive (either in increase or decrease) than the control requirements of the driver.
The one technique that I think could bring with it near-perfectly-controlled speaker systems would be something similar to, but entirely different in effect from the digital signal processing systems we now use for “flattening” or otherwise correcting speaker frequency response or overall in-room performance. To my knowledge, nothing like this currently exists, but if it could be built it could be a genuine “game-changer”.
Instead of working in the frequency domain, this device would only affect system dynamics, and would – perhaps as some sort of micro-time-delayed “feed-forward” system (most easily possible in the digital domain) – analyze the music to be played; determine its dynamic changes; and, perhaps with a selectable or adjustable “no change” level which would include everything up to the point where the driver’s (or drivers’) own control factors start to be less than totally effective, apply, above that point, dynamic gain changes to the leading edges of the signal’s waveform that would, using the power of the amplifier, supply the additional control necessary.
Such a system could be designed for a specific driver; for a specific combination of drivers (a specific speaker system); or as an outboard accessory suitable or with options to make it suitable to any speaker, and would, if it were ever to become available, bring with it a whole new level of loudspeaker performance accuracy and musical enjoyment. It might never be perfect, but with appropriate design and the ability to feed changes forward to correct “undershoot” in advance, it would certainly be far better than anything we have now.
John, Jeff, Dan, Tim, Thorsten, Bascom, Tony, Bob, Peter – all you other designers out there, whether of tube or solid state gear – now that you know about it, make it happen! We’re all waiting!
Veuillez communiquer avec notre équipe publicitaire afin d’obtenir les renseignements relatifs au placement média, que ce soit sur notre site Internet ou à l’intérieur de notre magazine. Nous nous ferons un plaisir de répondre à vos questions.