US 2134059 A
Description (OCR text may contain errors)
Oct. 25, 1938. o. H. scr-IADE LOUDSPEKER CIRCUITS 2 Sheets-Sheet l Filed March 15, 1957 7'0 PWER OUTPUT 7055.9
CURVE H VER/1U REJPU/VSE UFE MT//P/MS'E IND REJUNNCE FURREUf/ON 6em/6mm wwf/Imis u m f 234 5 www INVEN TOR y] l1'. SCHADE A TTORNE Y Oct. 25, 1938. o. H. SCHADE LOUDSPEAKEB CIRCUITS Filed March l5, 1937 2 Sheets-Sheet 2 WHR/VfR TUBE 50 INVENTOR 0770 H. 5CH/10E STR/KING RANGE WITH CORRECT/0N ATTORNEY v 'Patented Oct. 1938 PATENT OFFICE,
Loupsrsaxan cmcul'rs om n. schade, west oneven, N. J., signoria Radio Corporation of America, a' corporation of Delaware Application March 15, 1937, serial No. 130,8345
4- claims. ,y (ci. 17a-,1).
My present invention relates to sound reproducer circuits of the loudspeaker type, and more particularly to a loudspeaker circuit capable of utilizing` large power sources with displacement n amplitudesfof the mobile elementhof the loudspeaker limited to maximum permissible values at low audio-frequencies. l
The safe'displacement amplitude of they driving mechanism of any given type ofloudspeakcr 1 is limited to a certain maximumvalue'. If the latter is exceeded, ,t the coil amature, or4 diaphragm, strikes against stationary parts. Again, the elastic limitv of the returning element, as the spring, may be exceeded thereby causing possible damage to the speaker. The electro-acoustic efliciency of commercial dynamic loudspeakers is nearly constant below 200 cycles, except at, near or below their resonance frequency. The displacement amplitude rises at the mechanical resonance frequency to a value which may be many times higher than the amplitude obtained at the same frequency for a non-resonant system. If the resonant frequency of the speaker is placed high (100 cycles or above), the maximum power inputs, the moving parts do not usually strike against stationary parts. If the resonant frequency is low (60 cycles or below), the amplitude is limited bythe free space for the moving coilwhich may strike against stationary. parts with large power inputs. e
Power `sources of high. internal impedance, such as pentode or class B amplifiers,"will emphasize these conditions as the power in the moving coil does not decrease with rising impedance (higher efficiency) of the coil around resonance. 'I'he acoustic effect of resonance is an increase (boom) in sound output around the resonance frequency 40 and is undesirable for reproduction of high quality broadcasts, as the range of increased output is confined to a narrow frequency band. For good acoustic performancev the resonance should be below the frequencyrange used for music and speech. A low resonance frequency restricts the power output into the speaker to certain maximum values at low frequencies without striking of the voicel coil against stationary members.
Since theamplitudes of voice coil displacement must be limited to safe values at lowfrequencies, various devices haverbeen used to `secure such limitation. For example, a high resonance frequency has been employed to sacrifice low fre- Cai j quency sound output. Again.' the diaphragm amplitude is usually limited by the spring or the outside seal, which may bel strained with large` area, and consequently the air load on the speakver, has been increased to reduce the amplitudel of displacement; this remedy, however, affects the high frequency response. Another device which has beenfused in the past involves the artificial increase of the air lo'ad on the dia-V phragm at low frequencies 4by coupling it to acoustic absorption chambers which increase the. air pressure at the diaphragm. Finally, the electricalinput tothe speaker -may be limitedfby lo dividing larger powers between two or more speakers of different resonance frequencies.
It is one of the main objects of the present invention to provide a novel, and highly effective, method of'Vv compensating for the mechanical 15 resonance of a loudspeaker; a series resonant network, tuned substantially to the mechanical resonance, being .connected in shunt with the speaker to maintain the' total impedance or sound output constant. 20 K Another important object isto overcome the shift in speaker mechanical resonance tn higher` frequencies, due to increasing power and'caused by electro-mechanical overload conditions on the loudspeaker, by using a series resonant-network 25 across the speaker input which is constructed so as automatically to trail the speaker resonance variation.
Still other objects ofthe invention are to improve the operating characteristics of `dynamic loudspeaker network, and more especially' to provide eflicientand acoustically satisfactory speaker circuits which'are economically manufactured and assembled. f
The novel lfeatures which I believe to be char- 35 acteristic ofmy-inven'tion are 'set'forthin particularity in the appended claims; the inventionV itself, however, as to both its organization andY .method of operation will best be understood by reference to the following description taken in 40 connection with the drawingsin which I have indicated diagrammatically several circuit orn ganizations whereby my invention may be carried into effect.
In the drawings: v o 45 Fig. l shows one form of the invention,
Fig. 2 graphicallyillustrates the eifect of using the invention with fixed frequency resonance correction, K 50 Fig. 3 shows speaker response curves showing` the striking range whencorrection isV used, y
Fig. 4 is an audio output circuit employing another form of the invention, and
Fig. 5 shows curves similar to those of Fig. 3 55 when self-adjusting ronance correction il em- Ploved.
Referring now to the accompanying drawings. wherein like reference characters in the different figures designate similar circuit elements, ingrll'ig. 1 there is shown a power output network for driving a loudspeaker of the dynamic type. '111e speaker is conventionally represented as including a diaphragm I and voice coil 2; those skilled in the art are fully acquainted with the details of constructing this type of speaker. The transformer 3 adapts the electrical speaker constants to the characteristics of the power tubes. The latter are not shown, but it will beYV understood that the plates thereof, assuming` a push-pulloutput stage, are connected to the primary P. j A` series resonant path, including` condenser l, re-
sistor 5 and coil t, is connected across a portion` of the primary P; although the path can be con nected across the entire primary if design considerations warrant it. The resistor l may be the resistance of coil i. i
The series resonant network 4 5-l functions to counteract the mechanical resonance of the speaker. The constants of the elements are adjusted to give a current rise at the mechanical resonance frequency of the speaker suchthat the total sound output remains constant. In Fig. 2 there is shown the effect of using a nxed frequency speaker resonance corrector network: 'I'he curves in Fig. 2 were obtained on 'a circuit using a medium sized speaker with a maximum power input of 'approximately 10 to 15 watts. The power source utilised' a 227 driver and a pair of 146 type power output tubes. The curves were taken for approximately 5.5 watts input into the speaker. The transformer I was of com mercial size and silicon steel.
Curve I of Fig. 2 shows the overall response curve of the audio network with a pure resistance load being connected across the secondary of transformer I. The dotted curve 1I is the overall response of the network when the speaker is the load. However, no correction for its reactanee or resonance has been made. It will be noted that at about 60 cycles there exists a resonance voltagerise due to the mechanicalresonance of the particular speaker; The rise above 200 C. P. S. is due to the inductive reactance of the voice coil.
Curve m shows the overall response curve for the network,including the speaker load, after correction ismade for inductive reactance and me'` chanical resonance.` It will be observed that the shunt. section i-I absorbs the resonance peak at 60 cycles. The high frequency response of curvelI above 200 C. P. B. is brought down by a special .shunt section which comprises a condenser Ci in series with a resistor R1. The element I consisted of an iron core coil having an air gap;.the coil had a fairly constant inductance.
As larger powers arekused to drive a speaker,v it is found that the so-called "striking" range of the voice coil near resonance is widened. Fig. 3
shows measured curves of the voltage on the primary of the coupling transformer between the plate circuit of a class B pentode stage and a loudspeaker of low resonance frequency for various constant A. C. grid voltages on the power stage versus frequency. The range of striking of the voice coiiis marked. The section ,violent striking" is dangerous to the speaker. This is partly due to the shift in mechanical resonance to higher audio frequencies with increasing power,` and caused .by electromechanical overload conditions on the loudspeaker. (Bee the loops in Fig. 3).
steel; this determination is made by bridge measurements for the particular grade of steel. Ban., in turn'is directly proportional to the peak magnetizing voltage, and inversely proportional to the frequency. It follows that the resonance frequency of a tuned circuit having a closed iron core inductance varies with the applied voltage, as well as with frequency. As the voltage on the speaker, and also the mechanical resonance frequency, increase with power input rise. the resonance frequency of an electrical shunt section the voltage by absorptionvof power, if by design the ux density in the inductor has increased beyond 4000 g'auss to a value where the inductance'decreases rapidly with increasing voltage.
In Fig. iis shown a power output stage using a pair of I2A5 pentodes in push-pull. The stage is of the class B type. The output transformer l has the voice coil Ilofthe speaker across its' secondary. The resonance correction network includes coil Il and condenser I2; they are in series relation and tuned to approximately the mechanical resonance of the speaker. The con,
denser il and resistor Il are connected in series, and both are in shunt with the series resonant network Il-l2. Both series paths are in shunt across the primary P'. The condenser i8 and resistor Il function to correct for the' inductive reactance ofthe voice coil, as explained in connec-H tion with Fig. l.A The function of path ik", then, is to reduce the high audio response of the audio`output network. y Y
`In a circuit of the type shown lin Fig. 4 that was tested, a value of 0.25 mfd. was used for condenserfil. 'I'he choke coil l I was of such density, and
. the resonant frequency of the absorption circuit Ii-g-II was so selected, as to secure the desired vtrailing action. The choke "used contained approximately 0.25 lb. of silicon steel and 0.08 lb. of copper wire; this showing the relatively small cost of the correction circuit. Actual results observed with this type of circuit were as follows:v (see curves of Fig. 5): For low powers (0.3 w.) the low frequency response ofthe speaker is excellent as substantial absorption of power by the resonance corrector takes place only below 35 cycles. For a somewhat larger power (0.7 w.) more absorptionv takes place below 45 cycles. At higher power levels w.) the absorption of power by the network I I-I! is considerable thus preventing damage to the speaker; and commences at about '15 cycles; the resonance frequency of speaker and circuit having increased to this value due to increased voltage. Considering ther increase in sou'nd output at speaker, resonance, the reproduction of low frequencies at this power level is maintained substantially con-4 stant down to approximately 65 cycles, although the voltage and power input to the 4voice coil are actually decreased as compared to higher frequencies. 'I'he amount of absorbed power is ad- Iiusted by the choke coil resistance.
using a closed Viron core inductor will 4automatically trail the speaker resonance and lower The operating range of the iron core chokes I will be in the `region of steel saturation. This may give rise to a certain amount of harmonic voltages which will cause corresponding distortion of the sound output. Such distortion is of slight importance since the speaker would otherwise strike and produce rattling sounds.
KAgain, the sound output at the very low frequencies is distorted anyway even at low volumes because of the insufficient baffle size at low frequencies. It is pointed out that undistorted reproduction of a 60 cycle tone requires a distance of approximately 2.8 meters from back to front of the speaker diaphragm around the cabinet. The shift of the resonance curve in the iron core corrector circuit may be such as to take care of variations of the resonance frequency of loudspeakers as occur usually in production. By'employing the correction circuit of the present invention it is possible to employ a medium size, or small, loudspeaker successfully with large power inputs, and avoid a high resonant frequency with consequent gain in the reproduction of low frequencies and elimination of low frequency resonance transients (boom).
possesses a .natural resonant frequency whichV varies in value with increasing power input to the loudspeaker, a series resonant circuit connected in shunt with the speaker input, said resonant circuit being adjustably tuned automatically to a frequency in the vicinity of y said variable re`so nant frequency of the diaphragm system.
The self-adjusting resonance frequency feature ,does not require careful matching of the correcting k3 2. In combination with a loudspeaker oi' the type having a mobile diaphragm system which possesses a resonant frequency which varies ln' value with increasing power input to the loudspeaker, a series resonant circuit connected in shunt with the speaker input, said resonant circuit being adjustably tuned automatically to a frequency in the vicinity of said resonant fre' quency, and said circuit including reactive` elements of such construction that the resonant fre'- quency thereof changes with the power suppliedA y thereto.
3. In combination, in an audio reproducer system, a loudspeaker of the dynamic type whose mobile system has a relatively low resonant frequency, an input transformer having its secondary coupled to the speaker moving coil, an absorption circuit, resonant to said frequency, connected across at least a portion of the transformer primary, said absorption circuit including at least a condenserin series with a choke coil,
andthe constants of said choke coil being so chosen that the resonant frequency of the absorption circuit varies with the power input to vthe transformer whereby the resonant frequency of the absorption circuit automatically trails changes in resonance yfrequency of said mobile system. 7
4. In an audio reproducer system of the type employing mobile elements having a mechanical resonance which causes rattling and striking of the mobile elements against immobile portions of vthe reproducer, an electrical absorption circuit coupled with the reproducer input, said absorp- Y tion circuit being resonant to a frequency in the vicinity of said mechanical resonance whereby the effect of said mechanical resonance -on the acoustic character` of the reproducer is greatly diminished, and the constants of said absorption circuit being'so chosen that changes in the frequency value of said mechanical resonanceV due to increased power input to the reproducer are compensated for by corresponding changes in the resonant frequency ofV said absorption circuit.
O'ITO H. SCHADE.