US3014096A - Sound reproducing means - Google Patents

Description

Dec. 19, 1961 w. w. CLEMENTS SOUND REPRODUCING MEANS Filed May 29, 1956 AMPLIFIER PROGRAM SIGNAL SOURCE FIG.

EXCURSION GAUGE AMPLIFIER PROGRAM SIGNAL SOURCE FIG. 2

AMPLIFIER INTEGRATOR I NTEGRATO PROGRAM SIGNAL SOURCE FIG. 3

AMPLIFIER INTEGRATOR] PROGRAM SIGNAL SOURCE FIG. 4

ilniteri rates 3,014,096 Patented Dec. 19, 1961 free 3,014,996 SQUND REP ODUCENG MEANS Warner W. Clements, Los Angeles, Calif. (13435 Java Drive, Beverly Hills, Calif.) Filed May 29, 1956, Ser. No. 588,029 5 Qiaims. (Q1. 179-1) My invention relates to systems for sound reproduction which incorporate moving-coil type loudspeakers.

It is possible for a loudspeaker, by itself, to impose serious limitations upon the quality of sound obtainable from an otherwise excellent sound reproducing system. Associated with one such limitation is the resonance which can be defined in the low frequency range of nearly any moving-coil type loudspeaker. This is the principal resonance affecting the operation of such loudspeakers. I will hereinafter refer to it as the resonance and to its frequency as the resonant frequency. The terms are ordinarily used loosely to apply in the case of either mounted or un-mounted loudspeakers. It is when a loudspeaker is mounted or enclosed that its reproduction characteristics become important and it is this condition that I imply. \I will speak in particular of direct-radiator systems, which type is in most general use at this time. My remarks apply with lesser force to horn-loaded systems; however many of the latter operate virtually as direct-radiator systems at low frequencies.

In actual practice the following effects can be-noted in connection with the resonance: In the frequency region extending for some distance above the resonant frequency the sound output is relatively unaffected by loudspeaker characteristics. At and near the resonant frequency the sound output tends to be accentuated. Below the resonant frequency the sound output is rapidly attenuated with falling frequency. At any more than one octave below the resonant frequency the output is usually so reduced by comparison with that at other frequencies as to be virtually useless. If loudspeaker and enclosure be designed to effect a higher resonant frequency, these effects accordingly take quencies and the band of reproduced frequencies is thereby narrowed. Conversely, if a lower resonant frequency can be achieved, then lower sound frequencies are included in the reproduced range. The reproduction of the lowest notes is held to be a desirable objective in high quality sound systems. Also, the lower the frequency of the peak in the response, the less objectionable it is from a standpoint of sound quality and the easier it becomes to deal with it by means of restricting signal transmission at the lowest frequencies. These considerations indicate that loudspeakers intended for high quality reproduction should have resonant frequencies at least as low as the frequencies of the lowest audible sounds. Yet the resonant frequencies found in practice in even the most expensive loudspeakers are much higher than this. Unfortunately it is difficult to make substantial reductions in the resonant frequencies of practical loudspeakers.

The resonance in question is a mechanical phenomenon. Like other simple mechanical resonances it is associated with a weight and a springiness or, more correctly, a mass and a compliance. Both qualities are involved with the operation of the loudspeaker diaphragm, which must vibrate in order to induce sound waves in the air with which it is in contact. The diaphragm is built to be very light, but it still has appreciable mass. Moreover, the diaphragm is only part of the moving system. There is also a voice coil which is necessary to drive the diaphragm and is built integrally with it in conventional equipment. In addition some mass effect is contributed by certain auxiliary structure and by the air that moves back-and-forth with the diaphragm.

place at higher fre- As the diaphragm moves in either direction from its neutral position it encounters a restoring force which would tend to return it to neutral. This restoring force is due at least in part to the suspension system, which is designed to make the vibratory action of the moving system center about a neutral position. If the loudspeaker is mounted in a cabinet that is enclosed totally, or nearly so, the enclosure, too, contributes to the restoring force, due to the compression or rarefaction of the air within it as the diaphragm moves. The total restoring force is substantially proportional to the measure of the excursion from neutral position. The ratio of restoring force to excursion is the measure of the stiffness, which is defined as the reciprocal of the compliance.

The resonant frequency canbe defined as that frequency at which the mechanical reactance due to total effective moving mass is numerically equal to the mechanical reactance due to total effective stiffness. Qualitatively speaking, the respective parameters have the same effeet on frequency as in the case of a weight bouncing on the end of a rubber band. Increasing the stiffness of the rubber band will cause the frequency of bounce to increase and vice-versa. Increasing the mass of the weight will cause said frequency to decrease and vice versa.

it can be seen that, in order to re-design a loudspeaker for a lower resonant frequency, it is necessary either to increase the mass of the moving system or to decrease the effective stiffness. But there are considerations that oppose the employment of either measure.

Considering the moving mass first, it can be shown that to increase this quantity is to cut loudspeaker efficiency sharply. Loudspeakers that achieve improved low frequency reproduction by the method of adding otherwiseunnecessary mass to the moving system require the expenditure of greatly increased amounts of amplifier output power at all frequencies as payment for said improvement, which occurs only at the lowest frequencies. By way of example, to decrease the resonant frequency by one octave by this method would require that total effective mass be increased by a factor of four. This would result, according to efficiency formulas given by Olson and by McLachlan, in loudspeaker efficiency being decreased approximately sixteen times. (See Olson, Harry F., Elements of Acoustical Engineering, New York, D. Van Nostrand, 2nd ed., 1947, pp. -127; also McLa'chlan, N. W., Loudspeakers, New York, Dover, 1960, page 253.) Such a loss of efficiency is efiected in'increased voice coil heating, which can become a problem in its own right. Besides these drawbacks, the method of mass loading has the additional one of placing extra gravitational strain on the suspension system and tending to make mounting position critical. As for decreasing total stiffness, one avenue of approach would be to decrease the suspension stiffness in the loudspeaker proper. To do so and yet maintain strength and stability in the suspension is a mechanically difficult proposition. Meanwhile, if the loudspeaker is intended to have even fair low frequency response, it must be housed in an enclosure of a type that, as a necessary evil, contribues to the total stiffness. An average sized enclosure would supply enough stiffness to hold the resonant frequency of a typical matching loudspeaker well up in the audio band, even if suspension stiffness in the loudspeaker proper could be completely abolished. In other, words, unless the stiffness arising in the enclosure can be dealt with, there is no hope for attaining significantly large reductions in the resonant frequency by the method of reducing the stiffness.

But the only way to reduce stiffness due to the enclosure is to increase the size of the enclosure. And enclosures now in general use for high fidelity reproduction purposes are in most cases already much larger than would be desirable. Large enclosures are costly and inconvenient. Rather than further increasing enclosure sizes, it would be of great benefit if some Way could be found for reducing them.

What has been said so far may be briefly re-stated as follows: Objectives to be sought in a high quality sound reproducing system include the reduction of the resonant frequency of the loudspeaker and the reduction in size of the loudspeaker enclosure. However, these are conflicting objectives. In the present state of the art, the only way to advance toward one of these objectives without retreating from the other is to apply extra mass loading to the moving system of the loudspeaker. But the latter measure is extremely costly from the standpoint of efficiency.

The invention described in my co-pending application, Serial Number 583,106, filed May 7, 1956 (now Fatent Number 2,948,778), presents one solution to this problem. The present subject invention presents another and different solution.

An object of my invention is to provide improved low frequency response characteristics in sound reproducing systems which incorporate moving-coil loudspeakers.

Another object of my invention is to permit the use of smaller loudspeaker enclosures in sound reproducing systems which incorporate moving-coil loudspeakers and which are designed for high-quality low frequency reproduction.

Stiffness makes its effect on the moving system felt through the force which is exerted in response to excursion. The strategy of my invention is to introduce an extra force which acts on the moving system. This extra force is caused to be proportional to excursion, in the manner of a force due to true mechanical stiffness, but is applied to the opposite direction from a force of the latter type. The overall effect can be described as the development of a negative stiffness which appears to reside within the mechanical system. The negative stiffness serves to cancel out a part of the pre-existing positive stiffness. Stated another way, the moving system is caused to behave, while reproducing program material, as if the true stiffness affecting its motion had been reduced. amount of apparent reduction in stiffness thereby defines the quantity of negative stiffness. The resonant frequency will be reduced in accordance with the reduction in apparent net stiffness. Of course the negative stiffness cannot be made so large as to exceed the positive stiffness or instability would result.

The source of the force that creates the effect of negative stiffness is an auxiliary current that is caused to flow in the voice coil. This extra current shares the same winding with the current representing the program material to be reproduced. (The two currents add up to a single resultant current but for the present purposes it is more instructive to consider the two components separately.) The two currents specified may be supplied respectively by two amplifiers. However for illustrative purposes I will describe the general arrangement wherein only one amplifier is used to supply both as l find this arrangement generally preferable.

Two types of electrically-encoded intelligence can be defined as being amplified simultaneously in an amplifier used for the purpose. One of these, embodied as a program signal, controls the current component in the output representing the program material to be reproduced. It may be derived from any conventional source for such a signal, such as radio or phonograph apparatus, wire transmission means, etc. (Such sources are well known to those skilled in the art. The presence of one or more of them is necessary to the operation of my invention but not directly concerned with the structure or principle thereof.) The other type of intelligence referred to is embodied as one or more auxiliary signals and controls the current component in the output that produces the effect of a negative stiffness. The auxiliary signal or sig- The nals may be developed from the actual motion of the moving system through various simple and inexpensive means and may be introduced into the amplifier through extra input connections provided for the purpose. The amplifier used should be designed for this service, but need be only slightly more complicated or expensive than a conventional amplifier.

Since my invention provides a means of negating stiffness, the loudspeaker designer is thereby left free to incorporate extra stiffness where he can reduce bulk, cost, or distortion by so doing. For instance the designer is enabled to make use of smaller loudspeaker enclosures in high quality systems. This matter of increasing the real stiffness points up still another advantage of my invention over the method of incorporating extra mass loading of the moving system: Any time stiffness and mass are both" increased in the design of a loudspeaker, the principal resonance is thereby sharpened and heightened, with consequent deterioration of sound output quality. Accordingly, when a designer in the past has sought to use mass loading to compensate for a smaller enclosure he has been forced to also incorporate extra mechanical damping, which measure carries a penalty of its own in the form of increased cost and further decreased efficiency.

To clearly distinguish my invention from the prior art it is necessary to mention the related matter of elec-- trically mediated dam-ping. There exist several feedback. systems which aim to provide improved loudspeaker damping. They accomplish this end by increasing the ap parent resistive control over the motion of the moving system, thereby snowing under the effect of the mass at frequencies above resonance. But unfortunately conventional loudspeakers require mass control in order to increase the velocity of the moving system at low frequencies and so to maintain the level of sound output at those frequencies. What damping systems accomplish, in short, is to suppress the resonant peak to a greater-or-less degree at the cost of simultaneously suppressing much of the low frequency response in general.

The present invention, by contrast cannot be used to increase resistive control, but is used to decrease the effect of existing stiffness. The latter method is a superior one for dealing with an objectionable resonance since it moves itto a lower frequency rather than suppressing it and thereby maintains and extendslow frequency response in general.

As has been stated, the functioning of my invention involves an extra current in the voice coil. Since it requires power to force this current through the voice coil, the benefits of my invention are achieved at the cost of some extra expenditure of amplifier output power. However this extra power is required only at frequencies below that of the true (as opposed to the apparent) resonance. By contrast the method of mass loading requires extra power at every frequency above that of its lowered resonance. To achieve a substantially lowered resonant frequency in given circumstances the method of my invention requires only a small fraction of the extra power expenditure required by the method of mass loading.

Other objects and advantages will be made apparent to one skilled in the art by the following description and claims, taken in conjunction with the accompanying drawings in which:

FIG. 1 is a diagrammatic representation of an illustrative embodiment in which the source for the auxiliary signal used is a pressure sensitive transducer mounted in the loudspeaker enclosure. The loudspeaker and the enclosure are shown in section; symbols are used for the other elements of the figure. The fine, arrowed lines in this and the other illustrations indicate electrical connections, with the direction of the arrows indicating the direction of the flow of influence.

FIG. 2 is a diagrammatic representation of an illustrative embodiment in which the source for the auxiliary signal used is an excursion gauge coupled to or integral with the loudspeaker. Symbols are used for the elements of the figure. The heavy arrowed line 15 represents the mechanical coupling or integrity.

FIG. 3 is a diagrammatic representation of an illustrative embodiment in which the source for the auxiliary signal used is means associated with the amplifier output circuitry plus an integrating circuit. Symbols are used for the elements shown.

FIG. 4 is a diagrammaticrepresentation of an illustrative embodiment in which the respective sources for the two auxiliary signals used are means associated with the amplifier output circuitry plus integrating circuits. Symbols are used for all elements.

General The amount of force, f required to produce the effect of a given negative stiffness, S is given by,

where x is the excursion in the affected mechanical system and the minus sign indicates that the force developed is backward from that associated with a positive stiffness. In my invention a force of the strength thus specified is developed by means of an auxiliary current, i that is caused to flow in the voice coil of the loudspeaker used. From loudspeaker theory,

fx= x (B) where B and L are loudspeaker design constants. Combining Equations A and B:

r x BL Equation C gives broad instructions for putting my in- Vention into practice in connection with amoving-coil loudspeaker employed in the sound reproducing situation. The excursion of the moving system must be made to govern an extra current proportional to itself that is caused to flow in the voice coil.

This task must be accomplished without imposing appreciable mechanical load on the moving system, for one of the purposes of the invention is to avoid such loading. Since little power can thus be derived from the controlling source and since it takes power to force the required current through the voice coil, amplification is required. Furthermore, since mechanical information must be trans lated into electrical information, the process of transduction is also required. In preferred embodiments of my invention, equipment normally present in sound reproducing apparatus is enlisted to help out in one or both of these tasks, amplification and transduction, while simultaneously continuing to fulfill its normal role.

Re-arranging Equation C:

x n a: Equation D shows that the amount of negative stiffness developed is a function of loudspeaker characteristics and of the ratio of auxiliary current to excursion. One skilled in the art will be able to control this ratio in given equipment in the design process. It is also a simple matter to incorporate means for varying the ratio at will in equipment in use, thereby making the amount of negative stiffness developed a matter of the users discretion.

Naturally there is a limit to the amount of negative stiffness that can safely be developed in a given situation. If the negative stiffness were made so large as to equal the positive stiffness, the apparent resonant frequency would be zero and the system would be unstable. In practical equipment instability may occur somewhat short of the limit thus defined. That is to say that oscillation of one kind or another may occur at very low frequencies when the negative stiffness developed at other frequencies is not yet equal to the positive stiffness. This can occur because of phase-shift at the low frequency extreme in conventional equipment. It would be theoretically possi-ble to avoid this effect by using throughout equipment responsive to DC, but in practice it isnt necessary to go to such an extreme. There could be no advantage to moving the resonant frequency much beyond the lower limit of audibility; as long as no attempt is made to do so stability can be maintained in conventional A.C. responsive equipment, with the aid of proper design. Proper design in this situation means the employment of the same skills that are required to provide ample margins of stability in conventional feedback systems.

The principle difference between the respective illustrativeem bodiments to be described is in the means employed for transduction; that is to say in the means employed for developing electrically encoded intelligence depicting the excursion of the moving system. Since amplifier considerations are much the same for all embodiments, said considerations will be discussed first. Specific embodiments will be described in connection with the particular means of transduction used in each.

Amplifier Any sound reproducing apparatus employing a loudspeaker must include or be served by an amplifier. In preferred embodiments of my invention this amplifier is employed to develop the required auxiliary current while continuing to amplify the program signal. (in the case of multi-channel systems that use separate amplifiers and loudspeakers for various frequency ranges, my invention would preferably involve only the low frequency units.) It will be shown that the electrically encoded intelligence depicting the excursion of the moving system can be embodied as one or as two auxiliary signals. The amplifier accordingly must be provided with a plurality of inputs, one for the program signal and either one or two for the auxiliary intelligence. Regardless of the number of inputs, the amplifier has only one output so all signals are combined therein. Combination may, with equal validity, be regarded as taking place in the output or at any point in the pro-output signal circuitry where all signals are present. In any event, in accordance with the principle of superposition, one current component in the output can be ascribed to the program signal and another can be ascribed to the auxiliary intelligence.

The auxiliary intelligence may not require as much amplification as the program signal. Hence the representation for the amplifier used in FIGS. 1, 2, 3, and 4 is doesnt present an improper load to the source of said signal.

The portion of the amplifier'involved in amplifying the auxiliary signal or signals should be designed to keep phase-shift small to down below the lowest frequency to which it might be desired to move the resonance. In order to accomplish this a specially designed output transformer may be required. Alternatively, an output-transformerless amplifier may be utilized.

One skilled in the art of amplifier design can easily provide almost any desired amplifier output impedance (or apparent impedance). The influence of this parameter upon the operation of the invention will be briefly discussed, to facilitate the choice of a suitable value in a given case. The simplified theory already given, as exemplified by Equation D, assumes that the magnitude of the extra component of output current, i is immediately determined by the magnitude of the auxiliary signal or signals introduced into the amplifier, and says nothing about the effect upon said current of variation with fre- 7 quency of the load imposed upon the amplifier by the loudspeaker. Just what this effect is can be most clearly shown by means of the equivalent output-loudspeaker circuit transformed into a convenient form.

As usually given, this in a shunt-series circuit; the amplifier being depicted as a voltage generator having the output resistance in series, and the loudspeaker being'de' picted as a damped shunt resonant circuit representing the mot-ional impedance (the electrical reflection of the mechanical parameters) plus a series inductance and resistance representing the voice coil. For the present purposes the apparent output impedance and the voice coil impedance can be lumped together; their effects as felt in the vibratory system are indistinguishable. Also, as a simplifying assumption which is perfectly valid at low frequencies, the inductance of the voice coil can be ignored. Finally, by means of Nortons theorem, the voltage generator with its total series resistance (output plus voice coil) can be replaced by a current generator having the same total resistance in shunt. In this final form the equivalent circuit has all its elements connected in shunt. The current generator may be considered to produce the same auxiliary current, i regardless of the value of the output impedance.

As thus arranged, the equivalent circuit clearly shows the effect of the various parameters. It can be seen that whatever the value of the output impedance it is always possible to adjust the value of i to supply the required amount of negative stiffness, the latter quantity being independent of frequency. The effects of a less-thaninfinite output impedance do not interfere with the operation of the invention, but are superposed thereupon. The specific drawback of a too-low output impedance is that it introduces electrically-mediateddamping of the mechanical resonance. In some cases, as when other-thansimple loudspeaker enclosures are used, the latter elfect may be tolerated to some extent. In general, however, it will be found that excessive damping will interfere with the increase in diaphragm velocity that should take place with falling frequency, and will thus cause a drooping of the low frequency response. The lower the apparent resonant frequency is moved, the more objectionable this effect will be. A relatively high output impedance is to be preferred.

I have found in practice that some conventional vacuum tube output circuit designs work satisfactorily in my invention without special measures being required to boost the apparent value of the output impedance. However such circuits should not include any feedback which reduccs said impedance to low values or makes it negative. Where it is desirable to increase the apparent output impedance of a given amplifier, this may be accomplished very simply by separately incorporating negative current feedback.

The auxiliary current called for by Equation C can be caused to flow through the voice coil in either of two directions. In order to determine which is the proper direction it is necessary to know which convention for positive flow in the voice coil would correspond with a given convention for positive excursion of the moving system. It happens that any set of conventions applicable in the case of the program component of output current would also apply in the case of the auxiliary component. So the proper polarity in which to introduce the auxiliary current must be that which yields a given phase relationship with the program current; the direction of excursion can thus be ignored. If the auxiliary current should be caused to flow with reversed polarity, the apparent added stiffness affecting the moving system would be positive instead of negative. Obviously, then, if the resonant frequency is raised instead of lowered when the auxiliary circuitry is first connected up, the polarity is incorrect and should be reversed. The proper phase relationships of the program and auxiliary components of output current can be rigorously stated as follows: The two components should be generally opposing at frequencies above the apparent resonant frequency and generally aiding at frequeneies below the apparent resonant frequency.

In some embodiments the auxiliary output current will result from the introduction into the amplifier of two auxiliary signals. It is necessary first to determine that the currents in the output attributable respectively to each of these auxiliary signals have the correct polarity with regard to each other. After that the combined resultant of the two should be taken as the auxiliary output current which fulfills the purpose of Equations C and D and which must have the relationship with the program current specified in the previous paragraph.

Air pressure transducer" If a loudspeaker is mounted in a total enclosure and if the dimensions of the enclosure are small with regar to the wavelength of reproduced sound over a given frequency range, then the instantaneous increase or decrease of the pressure of the airwithin the enclosure is substantially proportional to the excursion of the loudspeaker diaphragm over the given frequency range. Certain pressure microphones are capable, over a given frequency range (which may be made to correspond with the one just mentioned), of producing an electrical signal whose magnitude is instantaneously proportional to the departure of the incident air pressure from ambient. if such a microphone is placed within such an enclosure, the electrical output of the microphone will depict the excursion of the moving system. Said output will therefore meet the requirements for use in my invention as an auxiliary signal, over the specified frequency range.

FIG. 1 represents an embodiment of my invention in which a suitable pressure microphone is used as the auxiliary signal source. resented by the rectangle 10. it is housed in the enclosure 2-; along with the loudspeaker 11. The left end of the rectangle representing the amplifier in the figure is the input end of same; the output is at the right end. The auxiliary signal produced by the transducer is introduced into the signal circuitry of the amplifier somewhere prior to the output so that it will control (or be amplified to become) the auxiliary component of current in the output.

It is important that this arrangement not be confused with any scheme for feeding back a signal based upon the sound output from a loudspeaker. Sound output is a very different thing from the pressure within a total enclosure having small dimensions with respect to wavelength. The sound output, at a given level of program signal input, ideally will remain constant at all frequencies over the useful range. On the other hand the level of the aforesaid enclosure pressure will vary inversely as the square of frequency (as an approximate rule where piston action prevails).

According to the rule just given the amplitude of the auxiliary signal quickly becomes very small with rising frequency. -At medium and high frequencies the system functions much like a conventional one not having the benefits of my invention. This situation has two'important consequences. First, the limitation of cabinet size with regard to wavelength of reproduced sound applies only at the low frequencies where the auxiliary signal is large enough to be important. The loudspeaker to which the transducer is applied can be used to cover a wide range of frequencies if proper enclosure size is initially chosen and if means can be incorporated for preventing standing waves from developing within the enclosure at any frequency.

As a second consequence, it can be shown that a virtue of the embodiment of HG. l is that it can be converted to a multi-channel arrangement wherein a crossover neti work is provided in the output of the amplifier to divide the frequency spectrum between two or more loudspeak ers. All that is necessary is to add the network and the extra loudspeaker or loudspeakers in the usual manner to The pressure microphone is rep-.

the system as represented. Certain obvious precautions should be taken. The loudspeaker to which the transducer is applied should, of course, be the low frequency unit. The lowest crossover frequency chosen should be well above the true resonant frequency of this unit. Also, the high frequency loudspeaker or loudspeakers must be prevented from creating a response in the pressure transducer.

A different arrangement is possible in which any extra loudspeakers used are driven by one or more extra amplifiers provided for the purpose, the division of the frequency spectrum being made prior to the amplifier inputs. In the event that such a system is used the added equip ment would have no bearing on my invention. The program signal source shown in the drawings can immediately be construed to be one adapted to supply a limited range of frequencies.

Excursion gauge transducer An excursion gauge is here defined as an instrument adapted to develop an electrical signal of amplitude instantaneously proportional to the departure from a given mean position of the body to which the instrument is mechanically connected. Such an instrument need only be properly attached to the moving system of a given loudspeaker in order to provide an auxiliary signal of the type required in my invention. FIG. 2 represents an embodiment utilizing an excursion gauge. The heavy line 15 indicates mechanical integrity or a mechanical link between the moving system of the loudspeaker 11 and the moving component of the excursion gauge.

There exists a great variety of possible designs for suitable excursion gauges. Such designs may be adapted from designs for strain gauges, meter deflection transducers, deflection-responsive pickups and certain microphones. Some suitable designs are simple and self-contained. Others are more elaborate and require associated equipment such as oscillators and demodulators. Tte rectangle representing the excursion gauge in FIG. 2 should be construed to include any necessary auxiliary apparatus.

One design of particular interest is the moving-coil, integrating excursion gauge. This instrument is basically a velocity meter but integrating its output signal in a circuit provided for the purpose changes said signal into one depicting excursion. (This process is based on the fact that excursion is the integral of velocity. Electrical integration does not as a rule take into account the conpresent purposes it is undesirable that it should do so.) If an excursion gauge of this type is used in the FIG. 2 embodiment, certain condensations are possible because of the fact that the loudspeaker, too, is a moving-coil instrument. For instance the velocity meter part of the excursion gauge (the integrator being the other part) can be built integrally with the loudspeaker, with the magnet that supplies the necessary magnetic field for the loudspeaker also being used to supply the field required by the velocity meter. As a further condensation the winding for the velocity meter can be made to move in the same magnetic gap with the loud speakers voice coil. This is equivalent to saying that the excursion gauge represented in FIG. 2 may consist simply of an extra winding on the voice coil of the loudspeaker, plus an integrating circuit.

Like the FIG. 1 embodiment, the FIG. 2 embodiment may be altered to include a crossover network and one or more extra loudspeakers connected in the output circuity. However in making such alteration'the precautions mentioned in connection with the former embodiment should be observed.

Loudspeaker self-transducer It is known that a voltage, called the motional voltage, is developed in the voice coil of an ordinary loudspeaker as the latter drives the diaphragm in the sound reproducing situation. This voltage is proportional to velocity The difference between the amplifier output voltage and the motional voltage is the drop due to the output current across the internal impedance of the loudspeaker. On the basis of these considerations it is possible to write the equation,

where e is the output voltage, t is the output current, c is the motional voltage, K is an arbltrary constant and 2,, is the internal (non-motional) loudspeaker impedance.

The quantity 2,. is more commonly called the blocked voice coil impedance.

Each term of Equation E can be used to represent quantitatively the magnitude of an electrical signal. The quantity KZ can be regarded as a second constant in the equation. Accordingly, the implications of Equation E can be stated as follows: The combination in proper ratio and in a generally opposing sense of a signal of magnitude proportional to the amplifier output voltage with a signal of magnitude proportional to the output current will yield a composite signal of magnitude proportional to the velocity of the moving system of the loudspeaker. The signal thus arrived at is similar to one that might be provided by a separate-unit velocity meter. Given such a velocity-indicating signal it is possible to turn it into an excursion-indicating signal by means of an integrator, as previously mentioned. Equation E thus suggests an additional method of fulfilling the function of transduction, as required in my invention.

The proper ratio referred to above in the italicized sentence is dependent upon the value'of Z Strictly speaking this ratio should vary with frequency, as Z has a reactive component due to the inductance of the voice coil. However, at medium and low frequencies for a given loudspeaker, the value of Z is mostly resistive. Meanwhile, since the excursion of the moving system of a loudspeaker falls off rapidly with rising frequency for constant sound response, the auxiliary signal to be furnished by the means suggested will be incon-' sequential at medium and high frequencies. negligible error will be introduced by considering the constant KZ to be all real and independent of frequency. Only the resistance of the blocked voice coil need be considered; the inductance can be ignored.

An embodiment utilizing a loudspeaker as a velocity meter for its own moving system is represented in FIG. 3. The rectangle 13 represents a means for developing a signal of magnitude proportional to the output voltage. The rectangle 14 represents a means for developing a signal of magnitude proportional to the output current. Means 13 and 14 are located in the output circuitry of the amplifier, preferably housed with the amplifier proper rather than with the loudspeaker. The signals from means 13 and 14 are combined, integrated, and introduced into the pre-output signal circuitry of the amplifier, as shown.

The combination of these two signals must bein a generally opposing sense and should preferably be in the.

ratio suggested by Equation E (although some error in this ratio can be tolerated and might even be desirable in certain conditions in order to achieve special effects upon the frequency response). Provided the specified conditions are fulfilled, one of the means 13 and 14 could be located on the primary side of the output transformer (if an output transformer is used) and the other on the secondary side, provided further that precautions be taken to keep non-signal (D.C.) currents where they belong. As regards balance between the two signals, one the means 13 or 14 may be made adjustable to permit Therefore.

made zero under otherwise operative conditions, no

auxiliary signal should be produced. That is to say that the signals from means 13 and 14 should exactly cancel each other out. Excursion may be temporarily eliminated by blocking the movement of the voice coil or by connecting a dummy load having the electrical characteristics of the blocked voice coil in place of the loudspeaker.

Loudspeaker self-transducer, dual signal The velocity-proportional signal generated by the means suggested by Equation E and embodied in the version of FIG. 3 is only a means to an end. That end is the provision of an excursion-proportional signal. It will now be shown that it is possible to get a signal of the latter type without having a definable velocity signal as an intermediate product. Integrating Equation E:

The term on the right of Equation F, being proportional to the integral of a voltage depicting the velocity of the moving system, represents the excursion. The equation shows that the signal of magnitude proportional to output voltage and the signal of magnitude proportional to output-current may be integrated before they are combined rather than after. Note that the first two terms of Equation F taken separately say nothing about velocity and taken together indicate excursion.

FIG. 4 represents an embodiment which utilizes the principle of Equation F. As before, 13'is a means for developing a signal of magnitude proportional to the output voltage and 14 is the means for developing a signal of magnitude proportional to the output current. Note that there are two integrators, one each for means 13 and 14.

This arrangement has the advantage that each of the two integrators used can be tailored to the characteristics of its respective signal source. As a matter of fact the means 13 and 14 will in some cases be combined, at least to some extent, with their respective integrating circuits. For instance if means 13 is used with a capacitor integrator, a series resistor is required to malte the capacitor current proportional to the amplifier output voltage. That series resistor might be considered as constituting means 13 or it might be considered a part of the integrating circuit.

The two signals derived from the integrators in FIG. 4 become the necessary auxiliary signals which between them convey intelligence representing the excursion of the moving system. If these signals do not have the proper ratio to each other demanded by Equation F, a gross correction can be made by introducing each into the amplifier at a different level of amplification.

For resolving the polarity and magnitude relationships As suggested in connection with the FIG. 3 embodiment, it might be advisable to provide a conveniently adjustable balance control for setting the relative intensities of the signals emanating from means 13 and 14. In the present instance a control fulfilling the same purposes might also be located between an integrator and the amplifier proper. Adjustment can be made by the operator or installer according to the following directions: First temporarily eliminate the motional impedance by one of the methods previously suggested. Supply as a program signal a steady test signal of fixed frequency. Arrange to connect and disconnect both auxiliary signals from the amplifier while measuring the output voltage. Adjust the value of one or the other of the auxiliary signals by the means provided until the output remains constant regardless of whether both are connected or both are disconnected. As before, some error in the balance can be tolerated and might even be introduced deliberately.

Integrator design Integrating circuits are available in either active or passive types. Both types find their basis in voltageto-current relationships in reactive circuit elements; the difference between the two types is that the active circuit incorporates an amplifying device, such as a vacuum tube, and the passive circuit does not. Either type is suitable for use in my invention. One skilled in the art is familiar with the design of said circuits.

No electrical circuit is capable of performing integration that is perfect in the mathematical sense. Integration by circuit means is achieved only at the cost of a loss in signal strength and the amount of that loss depends on how closely perfection must be approached. Perfect integration would require infinite attenuation, which is to say that the signal output from the integrator would be zero. For any given type of circuit, the less perfect the job of integrating that is performed, the less signal strength is Wasted in the process. ccordingly, for the sake of design economy, a user will want to know how poor a job of integrating can be tolerated when an integrator is called for in my invention.

The first thing to be noted in this connection is that the integrating operation need take place only at certain frequencies, namely, those frequencies at which an appreciable force simulating stiffness is to be developed. Once given a specified range of frequencies, it is possible to rate the performance of an integrator in terms of the phase lag it produces. An ideal integrator would be one which would produce exactly 90 phase lag over the specified bandwidth; however, no known between the auxiliary signals some point of reference wherein both signals are present is required. Since both must be introduced into the amplifier prior to the output circuitry, the current components in the output respectively ascribable to each may be used as the first two terms of Equation F. It can be seen that these latter two components must be generally opposing and should have a given ratio. Once this is attended to, they can be regarded as a single component constituting the necessary auxiliary output current. The said current thus arrived at must have the proper magnitude to produce the desired amount of negative stiffness as shown by Equation D. It must also have the proper polarity as it relates to the phase of the program current, as was discussed in connection with amplifier considerations.

circuit is capable of achieving this performance unless the bandwidth is zero.

Phase shift can readily be calculated or measured. In actual tests of embodiments of appropriate species of my invention, no difference in performance can be noted when the phase lag (at critical frequencies) departs from the ideal by as much as 20. Beyond that point the performance falls off gradually as the phase of the affected signal departs from quadrature. There is no point at which the ability of the apparatus as a whole to produce stiffness effects suddenly departs. However, as stiffness effects decrease, damping effects increase. As a result, by the time phase shift has been reduced to within 30 or so of zero (or damping effects are apt to be so masked as to not be readily discernable.

Therefore, herein and in the claims, when I speak of rendering [a signal] into integral form, or of integrating a signal, I do not mean to imply the achievement of mathematically ideal integration, for that would be impossible. I intend to imply only an approximation of that ideal; moreover, that approximation need not be a close one, but only good enough to permit the stated combination of components to produce a useful stiffness effect. When I speak of an integrator or an integrating circuit, I means a circuit capable of producing integration to the standards just defined.

Conclusion By the word signal as used in the specification and claims I mean specific electrically encoded intelligence measurable at a given circuit point. It is convenient and in accordance with current practice to speak of a signal as passing through an amplifier or integrator and being amplified or integrated in the process. However it should not be inferred from such usage that a signal retains any literal identity before and after being so operated upon. The identity intended is that of a response with the particular stimulus that evokes it. The character of the information transmitted can be changed in the process.

When I speak of the value of a given voltage, current, or excursion, or use the general term magnitude, I mean the instantaneous value unless otherwise indicated. The magnitude of a signal may be the measure of its voltage, of its current, or, even of its carrier frequency, whichever has the desired eflFect in the apparatus responsive to it. When I speak of combining two signals in common circuitry I imply that vector combination of their audio frequency intelligence will take place. In speaking of relationships between combined signals I take the 90 phase angle mark as the borderline between generally aiding and generally opposing, or between generally in phase and generally out of phase.

I use the word circuitry in a sense in which it is commonly used in electronic and allied arts, namely as generic to both circuit elements and circuit wiring. Examples of circuit elements are resistors, capacitors, transformers, and vacuum tubes.

It should be apparent that my invention resides largely in the realm of principle. My disclosure has necessarily been somewhat abstract. Where I have become specific to the extent of suggesting definite components and arrangement details, it has-been only for illustrative purposes; a great variety of specific constructions may be resorted to without departing from the spirit and the scope of the invention as hereinafter claimed.

I claim:

1. Sound reproducing apparatus comprising: an amplifier; a moving-coil type loudspeaker connected to the output of said amplifier; means for developing a signal of magnitude instantaneously proportional to the output voltage of said amplifier; means for developing a signal of magnitude instantaneously proportional to the output current of said amplifier; connections for combining the two signals thus developed in a generally opposing sense to produce a combined signal; an integrator circuit arranged to integrate said combined signal; connections for introducing the integrated signal into the pre-output signal circuitry of said amplifier; a source for a program signal; and connections for introducing said program signal into the pre-output signal circuitry of said amplifier.

2. Sound reproducing apparatus comprising: an amplifier; a moving-coil type loudspeaker connected to the output of said amplifier; means for developing a first signal component having magnitude proportional to the integral of the output voltage of said amplifier; means for developing a second signal component having magnitude proportional to the integral of the output current of said amplifier; connections for introducing said first and sec ond signal components into the pre-output signal circuitry of said amplifier; a source for a program signal; and connections for introducing said program signal into the pre-out'put signal circuitry of said amplifier.

3. The combination set forth in claim 2 further characterized in that said means for developing said first signal component and said means for developing said second signal component employ respective circuit elements for rendering said signal components into integral form; and still further characterized in that said connections for introducing said signal components into said amplifier are arranged to introduce each of said components at a different point, said signal components thereby being characteriza'ole as two auxiliary signals.

4. In sound reproducing apparatus: an amplifier; means for developing a signal component of magnitude proportional to the integral of the output voltage of said amplifier; means for developing a signal component of magnitude proportional to the integral of the output current of said amplifier; connections for introducing both of said signal components into the pro-output signal circuitry of said amplifier; and connections for introducing a program signal into the pre-output circuitry of said amplifier.

5. The combination set forth in claim 4 further characterized in that said developed connections for introducmg said signal components into said amplifier are arranged to introduce both of said components at the same point, said components thereby being characterizable as a single auxiliary signal.

References Cited in the file of this patent UNITED STATES PATENTS OTHER REFERENCES Clements: A New Approach to Loudspeaker Damping, Audio Engineering, August 1951 (pgs. 20-22 and 54-55).

Thaler et al.: Servo-Mechanism Analysis (pg. 209).

Images