|Publication number||USRE33333 E|
|Application number||US 07/211,217|
|Publication date||Sep 11, 1990|
|Filing date||Mar 11, 1988|
|Priority date||Aug 21, 1984|
|Publication number||07211217, 211217, US RE33333 E, US RE33333E, US-E-RE33333, USRE33333 E, USRE33333E|
|Inventors||Wilson E. Taylor, Jr., Larry E. Hand|
|Original Assignee||Peavey Electronics Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (4), Non-Patent Citations (6), Referenced by (16), Classifications (9), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention relates to the field of high fidelity audio-amplifiers, and more particularly such amplifiers employing digital techniques.
Audio-amplifiers have been marketed in the past, wherein the analog audio signal to be amplified is converted into pulse trains which are modulated in accordance with the analog signal, such pulse trains being amplified by power switching amplifiers, and the output thereof being applied to low-pass filters for reconstructing the audio signal thus amplified. However, the results produced by such amplifiers currently on the market are deemed to be below the level of excellence of sound reproduction desired.
It is an object of the present invention to provide an extremely high fidelity audio-amplifier employing audio signal modulated pulse trains which are amplified to high power output levels, and yet producing very low distortion (less than 0.1%) and a minimal degree of feedback (approximately 22 dB).
In conventional audio-amplifiers, the input and output waveforms are scaled and compared for differences, and these differences are thresholded to achieve distortion detection. In accordance with the amplifier of the present invention, this process is simplified whereby the input audio signal is converted into a sequence of high frequency pulses which control power switches, which in turn apply amplified pulse trains to a reconstruction filter to reconstruct the amplified audio input. The power switches employed for amplification function in a digital manner with their transistors fully saturated or fully cutoff, so as to operate most efficiently, and the various features of the invention result in the aforesaid level of excellence.
The above stated objects of the invention are accomplished in the preferred embodiment by providing a modulator for transforming the analog audio input signal into two complementary trains of pulse width modulated signals for driving the power switches, which modulator utilizes a precision triangular waveform to control the transformation. The modulator employs all differential processing without saturated transistors to achieve a minimal distortion modulation system. Upon the occurrence of over-modulation, pulses from the width modulated pulse train produced by the modulator are dropped, and the resulting loss of pulses is detected and fed back to a variable gain amplifier stage of the system. The signal resulting from the detection of missing pulses by the missing pulse or distortion detector reduces the system input gain and signal level and can limit system distortion for up to 20 dB of input overdrive. The distortion detection system for detecting the missing pulses is coupled between the modulator and a voltage controlled attenuator and utilizes a timing network to establish the missing pulse period necessary to begin input attenuation of the audio signal by the variable gain amplifier stage.
The reconstruction filter includes a low-pass section for recovering the amplified audio and a notch filter section having notches centered to eliminate the fundamental of the switching frequency. A variable inductor is employed in connection with a section of the notch filter for aiding in the proper notching of the signals to maximize surpression of the switching frequency, such inductor being coupled to the feedback network, and is adjusted by the factory while observing output distortion until the distortion attains a minimum level. The reconstruction filter also includes a tapped inductor functioning as an autotransformer which introduces offsets in the output switching waveform of the power amplifier to effectively compensate for crossover distortion due to energy recovery diode overswings and power switch voltage drops. This circuit increases system efficiency by eliminating series diodes formerly necessary to prevent forward conduction of integral body drain diodes in the power MOSFETS used as switching elements in the power amplifier. Also, a short circuit protection circuit employs a mono-stable multi-vibrator (MSMV) which is triggered upon the detection of excessive audio current and inhibits the operation of the modulator within 2 microseconds, thereby to immediately inhibit the application of input pulses to the power switching amplifiers until the multi-vibrator is re-enabled after a fixed time delay.
Other objects, features and advantages of the present invention will become apparent upon study of the following description taken in conjunction with the drawings in which:
FIG. 1 discloses the overall organization of the circuitry of the preferred embodiment of the invention;
FIG. 2 illustrates the voltage controlled attenuator together with the missing pulse detector for controlling the attenuator;
FIG. 3 illustrates the feedback network;
FIG. 4 illustrates the modulator;
FIG. 5 illustrates the power switching circuitry;
FIG. 6 illustrates the reconstruction filter and R.F. filter at the output of the system;
FIG. 7 illustrates the protection circuit; and
FIGS. 8A-8D illustrates certain waveforms which are helpful in the understanding of the operation of the modulator of FIG. 4.
Referring now to FIG. 1, illustrating the overall digital energy conversion amplifier (DECA), an audio input is amplified by input amplifier 1 and is transmitted to voltage controlled attenuator 3 which is normally set in the "no attenuation state". In this state this circuit contributes less than 0.001% of the total distortion, although it is capable of up to 20 dB of attenuation, with less than 1% distortion. The audio signal is applied to the DECA modulator 7 via a feedback network 9. Network 9 consists of a series of lag/lead circuits for shaping the overall response of the amplifier which are designed to give an overall amplifier bessel response. Modulator 7 converts the incoming analog signal into a series of digital pulses which are pulse width modulated. This conversion is accomplished by comparing the input analog data to a fixed frequency sampling signal produced by oscillator 8. The signal is preferably a triangular wave, which permits double edge modulation of the applied signal, thus reducing distortion for a given sampling frequency. A level slicing technique of the triangular waveform is produced as described below to produce two pulse width modulated trains which are applied to power switch 11 which amplifies these pulse trains. The power switches are employed in a digital manner with their transistors fully saturated or fully off to maintain maximum efficiency of operation. The resulting amplified signals, having slew rates in excess of 5000 volts/microsecond, are applied to reconstruction filter 13 which functions to retrieve the D.C. component of the amplified pulse trains which is a replica of the audio-input signal. Filter 13 employs a low pass filter network to eliminate the undesirable high frequency content of the amplified pulse trains while retaining the high quality analog data, and includes several notch filters, centered to eliminate the fundamental and third harmonic of the pulse train, and also employs an autotransformer co-acting with energy recovering diodes, to produce circulating current which forces the high power switching transistors in power switch circuitry 11 to operate in their most efficient switching region. A radio frequency filter 16 is coupled to the output circuit of filter 13 to completely eliminate the residual pulse train signals which may otherwise be present.
A drop in the frequency of the digital pulse trains developed in modulator 7 occurs when the analog input from feedback network 9 exceeds the amplitude of the triangular waveform. This action is due to output clipping caused by the application of an analog signal which is greater than the amplifier can reproduce at nominal gain. The occurrence of clipping, reduces the frequency of the digital pulse train developed by modulator 7, and a timing circuit within distortion detection unit 5, is employed to forward a signal to voltage controlled attenuator 3 which attenuates the input audio signal pursuant to the reduction of frequency of the pulse trains developed in modulator 7. Modulator 7 utilizes differential processing without employing saturated transistors, thereby to achieve a minimal distortion modulation system. Protection and monitoring circuitry 17 disables the drive signal pulse trains generated in modulator 7, within 1 microsecond, in the event of a fault condition. For example, a short circuit condition produces excessive current through the reconstruction filter, which is employed to actuate a MSMV to inhibit the operation of the digital pulse train amplifying circuitry in modulator 7. The MSMV is automatically reset after a short time, and is again set upon the occurrence of a new fault condition.
Referring now to FIG. 2, input amplifier 21 receives the audio-input signal which is transmitted through amplifier 22 and 23 of the feedback network designated 9 in FIG. 2. FIG. 4, which is directed toward the DECA modulator, includes a pair of transistors 24 and 26 coupled to current source 28 and functioning as a comparator. The amplified audio signal produced at the output terminal 31 of amplifier 23 of the feedback network is applied to the base 32 of comparator transistor 24. Triangular waveform pulse generator 36 is coupled to the base of transistor 26, and in accordance with the level slicing modulation operation of the present invention, a series of pulse trains will be produced at the collectors of transistors 24 and 26 which are fed to the inputs 38 and 40 of the amplifier of FIG. 4 comprising transistors 42 and 44 which transistors function to amplify the pulse width modulated trains produced by transistors 24 and 26. The emitters of transistors 42 and 44 are coupled to a current source 47 which may be employed to rapidly shut off the supply of current to transistors 42 and 44 in accordance with fault conditions to be described below.
The drive waveforms are produced in accordance with the level slicing technique by comparator 25 which generates the pair of pulse width modulated output signals of FIG. 8B and 8C which serve as complementary drive signals for transistors 42 and 44 which in turn will ultimately drive the power switching amplifiers discussed below. The variable audio signal is applied through amplifiers 22 and 23 to the base of transistor 24 which functions in a non-saturating mode, while the triangular waveform show in FIG. 8A from generator 36 is applied to the base of transistor 26 of comparator 25. As indicated by FIGS. 8A through 8C, the comparator 25 produces pairs of pulse width modulated signals having transistions corresponding to the crossover points of the triangular waveform with regard to the variable audio signal shown in FIG. 8A. As audio signal 51 increases as shown at the left-hand portion of FIG. 8A, the duty cycle of the positive going drive pulses of FIG. 8B will be reduced although their frequency will remain the same; likewise with respect to the negative going pulse train of FIG. 8C generated by the level slicing technique of comparator 25. The audio signal 51 may at some point such as at 52 exceed the amplitude of triangular waveform 53, and the operation of comparator 25 will result in the absence of a pulse as indicated at 56 of FIG. 8B and 57 in FIG. 8C. This condition is caused by output clipping due to the application of an audio-input signal greater than the level that the amplifier may reproduce at nominal gain and the result is a drop in the frequency of the pulse train indicated at 56 and 57 occurring when the audio signal exceeds the peak-to-peak amplitude of 53. This condition will be detected by the distortion detection circuit 5 of FIG. 1 to be explained below. Should the audio signal 52 become more negative than the bottom portion of the triangular waveforms, pulses will again be dropped to indicate an overdrive condition of the amplifier. The pulse trains of FIG. 8B and 8C are applied to the input terminals 38 and 40 of the amplifier including transistors 42 and 44, wherein the pulse train is amplified and transmitted to the input circuits of the push-pull solid state pulse train amplifiers of FIG. 5.
Thus in accordance with the level slicing technique of the present invention executed by the structure of FIG. 4, a periodic waveform is applied to one input terminal of comparator 25 and a signal proportional to the audio input signal is applied to the input terminal of transistor 24. It is important to note that the modulator of FIG. 4 does not employ saturated transistors, and differential processing without saturation achieves minimal distortion with respect to the modulation of the audio signal.
The circuit comprising the lower portion of FIG. 2 has its input terminal B coupled to terminal B of transistor 42. As a result, the waveform generated by the modulator and amplified by amplifier 42, is applied to the base of transistor 51 via capacitor 52 at the lower portion of FIG. 2. Distortion detector 5 of FIG. 1 may be found at the lower portion of FIG. 2, which functions to forward a signal to the voltage controlled attenuator 20 upon the detection of missing pulses due to overdrive as explained above. This action causes attenuation of the input signal to eliminate the over modulation condition. The nature of attenuator 20 is known to workers in the art; see U.S. Pat. No. 4,318,053.
The generation of pulses indicated in FIGS. 8B and 8C cause transistor 51 to turn on and off. The negative voltage initially stored in integrating capacitor 53 is determined by a voltage divider consisting of resistors 54, 56 and 57. This reference voltage stored by capacitor 53 will prevent transistor 55 from turning on to cause the voltage controlled attenuator 20 to attenuate the audio signal applied to the feedback network. The continued application of pulses of the pulse train, where pulses are not missing, causes the actuation of transistor 51 to in turn prevent substantial buildup of a negative voltage across capacitor 60 so that diode 61 remains back biased. However, upon the occurrence of a missing pulse condition or state, transistor 51 will remain off and capacitor 60 will charge up through resistor 62 to forward bias diode 61 and to cause a further buildup of the negative voltage across capacitor 53 via diode 61 which in turn results in the increase in collector current of transistor 55 coupled to the voltage controlled attenuator. This action will cause the attenuator to reduce the magnitude of the audio input signal to prevent the overdrive condition. Thus it may be seen that resistor 62 and capacitor 60 will form a timing network which determine the missing pulse period necessary to actuate transistor 55 to in turn produce attenuation of the audio input signal. With no current from transistor 55 the signal gain of the voltage controlled attenuator is unity and as the current produced by transistor 55 increases, the gain is reduced by up to 20 dB at maximum attenuation.
The power switches of FIG. 5 function to amplify the output waveforms produced upon the collectors of transistors 42 and 44 of FIG. 4. Terminal D of transistor 42 is coupled to terminal D of FIG. 5 whereas terminal E of transistor 44 is coupled to terminal E of FIG. 5. Optical isolation switching elements 71 and 72, actuated in accordance with the output pulse trains produced by the modulator transistors 42 and 44, are coupled, via intermediate amplifiers, to the input circuits of high power MOSFET elements 80 and 81 in the push-pull power amplifier circuit 82. The drain of field effect element 80 is coupled to +75 volts, whereas the source terminal of element 81 is coupled to a -75 volt bus. The output of the push-pull high current power amplifier is taken at terminal G which is the center tapped junction point between inductors 83 and 84.
Referring now to FIG. 6 which discloses the various sections of reconstruction filter 13 of FIG. 1, terminal G is coupled to a tap 88 of tapped inductor or autotransformer 91. The left-hand terminal of inductor 91 is coupled to terminal G via resistor 92 and capacitor 93, while the right-hand terminal of autotransformer 91 is coupled to resistor 96 which is in turn coupled to the input 97 of the notch filter portion of low-pass filter network; sections 101 and 102 thereof include inductors 103, 104 and capacitors 106 and 107. FIG. 6 further includes low-pass filter stage 110 terminating at an R.F. frequency filtering stage 111, acting in accordance with the invention, to completely eliminate any residual pulse train signals which might be present at the system output. The right hand portion of inductor 112 is coupled to output terminal 113 of the system. The low-pass filter, having the components listed below, consists of a fundamental low pass filter with a Chebyshev characteristic coupled to a series of notch filters, centered to eliminate the fundamental frequency of the triangular waveform (500 kHz) and the third harmonic of the pulse train.
A further feature of the invention employs variable inductor 121, coupled in parallel with capacitor 122, to form a high Q sharp notch filter which is centered approximately at 500 kHz. There is a tendency however for the notch filters to become phase shifted at about the 500 KHz center line and this is used to cause an increase in the efficiency of suppression of the pulse train frequency. Thus, variable inductor 121, is coupled to terminal J of the FIG. 3 feedback network and coupled in turn to the upper input terminal of amplifier 23 via resistor 35. The output 113 of the system is examined for distortion, and inductor 121 is adjusted to change its inductance which affects the feedback signal, until the scope observed distortion is reduced to a minimum during testing. This process is facilitated by including a damping resistor 123, which reduces the Q of the LC circuit including inductor 121. Thus, the signal fed back to terminal J of the feedback network by adjustable inductor 121, is employed to substantially eliminate the closed loop distortion of the output at 113 of FIG. 6. The notch rejection section of the reconstruction filter also includes fixed capacitors 106 and 107, coupled in parallel with inductors 103 and 104 as shown.
In accordance with an additional feature of the invention, tapped inductance 91 of FIG. 6, functioning as an autotransformer, performs crossover distortion correction in the output switching waveform generated by the push-pull power MOSFETS 80 and 81 by transformer action which causes the tapped left hand-portion of the inductor to insert offsets into the switched waveform to effectively buck the crossover distortions resulting from a number of sources such as energy recovery diode overshoots and the MOSFET "on" resistances. System efficiency is increased by eliminating series diodes necessary to prevent forward conduction of the integral body drain diodes in the power MOSFETS. The crossover distortion is produced by energy recovery diode overswings and the on resistances of the MOSFETS. The preferred compensation circuit includes having tap 88 of the autotransformer 91 coupled to amplifier center-tap 86 of the amplifier output reactance comprising inductors 83 and 84. Back biased diodes 64 and 66, coupled across the supply busses and coupled to autotransformer 91 via resistor 92, function to insert the compensation offsets into the waveform, cooperating back biased diodes 67 and 68 being coupled across the amplifier components as shown. Thus, the autotransformer functions as a floating battery source for inserting circulating currents back into the power amplifier via resistor 92, diodes 64 and 66, and the remaining amplifier switching circuit elements.
The protection and monitoring circuitry 17 of FIG. 1, corresponds to FIG. 7 disclosing circuitry for providing protection against the failure of transistor output devices such as the power MOSFETS 80 and 81 or due to other abnormally high currents. In accordance with this feature of the invention, terminals L and M of FIG. 7 are coupled to terminals L and M of FIG. 6 which will produce a voltage across them indicative of the current passing through resistor 96, of the reconstructed DC signal. Should the current through resistor 96 indicating a serious fault be excessive, the resulting voltage impressed across terminals L and M in FIG. 7 will cause MSMV 142 to become triggered, since its input terminal 145 is coupled to the output terminal of switching device 141 which in turn is coupled to terminals L and M via switching device 140. Thus a threshold voltage, produced by excessive current through the reconstruction filter resistor 96, will trigger the MSMV 142, to cause a change of state in switching devices 143 and 147 in tandom therewith. The change of state of switching device 147 will actuate switching devices 148 and 149 in the modulator section of FIG. 4; the collector of switching device 147, namely terminal P, is coupled to terminal P of the switching device 148. The result is an interruption of the current supply normally provided by transistor 149 to the modulator amplifier at the right-hand portion of FIG. 4, and as a result, no further pulses will be impressed upon terminals D and E which are coupled to the input circuits of the power switches of FIG. 5 discussed above. The described arrangement will thus sense excessive current in the amplified audio signal and shut down the system modulator within two microseconds to prevent output device failure. Owing to the recovery period of MSMV 142 the modulator is re-enabled after a fixed time delay, and will of course again inactivate the modulator should the high current fault condition persist. The result is a complete fast acting protection against output short circuit faults and the like.
Thus it should now be appreciated that the aforesaid inventive features of the novel digital energy conversion amplifier, cooperate to provide a very high efficiency system for audio amplification whereby high output power levels, with extremely low distortion (less than 0.1%), and a minimal amount of feedback (approximately 22 dB), have been achieved.
The following is a listing of selected component values utilized in the product to be marketed by the assignee of the present invention:
Capacitor 52: 180 pf
Capacitor 60: 0.047 mfd
Resistance 62: 10K
Resistance 54: 220K
Capacitor 55: 0.015 mfd
Inductor 33: 8 uH
Inductor 84: 0.8 uH
Tapped Inductor 91: 18 uH
Resistor 92: 3 ohms
Resistor 96: 0.03 ohms
Variable Inductor 121: 0.3 uH to 1 uH
Capacitor 122: 0.22 mfd
Resistor 123: 47 ohms
Capacitor 106: 0.2 mfd
Inductor 103: 0.46 uH
Resistor 126: 100 ohms
Capacitor 107: 0.33 mfd
Inductor 104: 0.6 uH
Autotransformer 91: 18 uH
Resistor 92: 0.3 ohms
Capacitor 93: 470 pf
Diodes 64,66,67,68: rur 820
FETS 80,81: irf 240
Inductors 83,86: 0.8 uH.
The above described embodiments of the invention are merely exemplary, and it should be understood that numerous equivalents to the above described circuitry may be employed by those skilled in the art rather than those specified. Thus the scope of the invention is to be limited only by the language of the following claims and art recognized equivalents.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4178556 *||Apr 19, 1978||Dec 11, 1979||Attwood Brian E||Class D amplifier system|
|US4182992 *||Apr 14, 1978||Jan 8, 1980||Attwood Brian E||Pulse width modulated signal amplifier|
|US4476436 *||Aug 25, 1981||Oct 9, 1984||Sony Corporation||Power amplifier|
|US4509101 *||Aug 15, 1983||Apr 2, 1985||Nippon Gakki Seizo Kabushiki Kaisha||Protection circuit for switching power amplifier|
|1||*||A 0.5 MHz Switching DC AC Inverter Topology Provides Low EMI Environments for Critical Spacecraft Applications; Attwood, Apr. 10 12, 1984.|
|2||A 0.5 MHz Switching DC-AC Inverter Topology Provides Low EMI Environments for Critical Spacecraft Applications; Attwood, Apr. 10-12, 1984.|
|3||*||Design Parameters Important for Optimization of Very High Fidelity PWM (Class D) Audio Amplifiers; Attwood; AES; Mar. 2 5, 1982.|
|4||Design Parameters Important for Optimization of Very High Fidelity PWM (Class D) Audio Amplifiers; Attwood; AES; Mar. 2-5, 1982.|
|5||*||Very High Fidelity Quartz Controlled PWM (Class D) Stereo Amplifiers for Consumer & Professional Use; Attwood; Feb. Mar. 1978, AES.|
|6||Very High Fidelity Quartz Controlled PWM (Class D) Stereo Amplifiers for Consumer & Professional Use; Attwood; Feb.-Mar. 1978, AES.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5115203 *||Jun 25, 1990||May 19, 1992||Rockwell International Corporation||Digital power amplifier|
|US5479337 *||Nov 30, 1993||Dec 26, 1995||Kaiser Aerospace And Electronics Corporation||Very low power loss amplifier for analog signals utilizing constant-frequency zero-voltage-switching multi-resonant converter|
|US6640092 *||Feb 22, 1999||Oct 28, 2003||Bae Systems Australia (Mbk) Pty Limited||Transmitter|
|US7138864 *||Nov 15, 2004||Nov 21, 2006||Tripath Technology, Inc.||Overcurrent protection in amplifier topologies employing DC isolation|
|US7200187||Jul 25, 2002||Apr 3, 2007||O'brien Thomas J||Modulator for digital amplifier|
|US8036402||Oct 11, 2011||Harman International Industries, Incorporated||Distortion compensation|
|US8054129||Nov 8, 2011||Yamaha Corporation||Class D amplifier|
|US8299853||Oct 30, 2012||Yamaha Corporation||Class D amplifier|
|US8942391||Sep 9, 2011||Jan 27, 2015||Harman International Industries, Incorporated||Distortion compensation|
|US20030031245 *||Jul 25, 2002||Feb 13, 2003||O'brien Thomas J.||Modulator for digital amplifier|
|US20050122168 *||Nov 15, 2004||Jun 9, 2005||Tripath Technology, Inc.||Overcurrent protection in amplifier topologies employing DC isolation|
|US20070140513 *||Dec 15, 2005||Jun 21, 2007||Harman International Industries, Incorporated||Distortion compensation|
|US20110158428 *||Dec 22, 2010||Jun 30, 2011||Kenji Hashizume||Sound distortion suppression control method and acoustic control apparatus using the method|
|US20110221529 *||Sep 15, 2011||Yamaha Corporation||Class d amplifier|
|WO2005050829A2 *||Nov 16, 2004||Jun 2, 2005||Tripath Technology, Inc.||Overcurrent protection in amplifier topologies employing dc isolation|
|WO2005050829A3 *||Nov 16, 2004||Nov 2, 2006||Babak Mazda||Overcurrent protection in amplifier topologies employing dc isolation|
|U.S. Classification||330/10, 330/207.00A, 330/251|
|International Classification||H03F1/52, H03F3/217|
|Cooperative Classification||H03F3/2171, H03F1/523|
|European Classification||H03F1/52B, H03F3/217B|
|Jan 5, 1994||FPAY||Fee payment|
Year of fee payment: 8
|Feb 17, 1998||REMI||Maintenance fee reminder mailed|
|Jul 14, 1998||SULP||Surcharge for late payment|
|Jul 14, 1998||FPAY||Fee payment|
Year of fee payment: 12