METHOD AND APPARATUS FOR PULSE WIDTH MODULATION IN A SWITCHING AMPLIFIER

Technical Field

[1] The present invention relates to an audiosignalprocess, and more particularly, to a method and apparatus for pulsewidthmodulation in a digital switching amplifier.

Background Art

[2] For amplifying an audio signal, an analogamplification system and a digital amplification system are used. The analogamplification system converts a pulse code modulation (PCM) audio signal intoan analog signal and performs current amplification and voltage amplificationhaving an intermediate voltage stage, so amplification efficiency decreases. Onthe other hand, the digital amplification system performs pulse widthmodulation on the PCM audio signal and also performs switching amplification byrepeating ON/OFF, and the PCM audio signal is output through an analog filter. Therefore, the digital amplification system has amplification efficien- cygreater than that of the analog amplification system. In addition, theamplification circuit of the digital amplification system can be embodied in asmall size and has low power consumption. There are many kinds of the digitalamplification systems. In a digital differential drive system used in thedigital amplification system, there are also many kinds of systems including anAD drive system and BD drive system.

[3] FIG. 1 is a view showing a pulse width modulationcircuit in a conventional AD drive system;

[4] Referring to FIG. 1, a PCM audio signal is input toa (+) port of a comparator 100, and a carrier signal is input to a (-) port ofthe comparator 100. The carrier signal, for example, has a triangle waveform. The comparator 100 outputs a value of +1 or -1 according to a result ofcomparing the two signals input to the (+) port and the (-) port. For example,if a value of the PCM audio signal input to the (+) port is larger than a valueof the carrier signal input to the (-) port, a value of +1 is output,otherwise, -1 is output. Accordingly, output waveform of a PWM+ signal, whichis a positive PWM signal, is a pulse width modulation signal having a differentpulse width according to a value of the input PCM audio signal. A PWM- signal,which is a negative PWM signal, is the inverted PWM+ signal by an inverterlO2.

[5] FIG. 2 is a view showing a pulse width modulationcircuit in a conventional BD drive system;

[6] In the BD drive system, two comparators 110 and 112are used, a PCM signal input to a (+) port of the comparator 110 is inverted byan inverter 114 and input to a (+) port

of the other comparator 112. The sameTriangle waveforms are used for carrier signals in both comparators 110 andl 12. Therefore, a PWM+ signal, which is a positive PCM signal, and a PWM-signal, which is a negative signal, are output.

[7] FlG. 3 is a view showing a pulse width modulationoutput signal according to the conventional BD drive system;

[8] The PCM audio signal 200 is input to the (+) port ofthe comparator 110, and the carrier signal 210 is input to the (-) port of thecomparator 110. The comparator 110 outputs a value of +1 or -1 according to aresult of comparing the two signals input to the (+) port and the (-) port, sothat the PWM+ signal 220 which is a positive PWM signal is output. The invertedPCM audio signal 200 is input to the (+) port of the other comparator 112, sothe PWM- signal 230 which is a negative PWM signal is output. Referring to FTG.3, modulated pulse widths of the PWM+ and the PWM- signals 220 and 230 varyaccording to a value of the PCM audio signal.

[9] When the PWM+ and the PWM- signals 220 and 230 areconnected to both t erminals of load, a differential output 240 subtracting thePWM- signal 230 from the PWM+ signal 220 is applied to the load. A waveform ofthe differential output 240 is the same as the output waveform denoted byreference numeral 240 shown in FlG. 3.

[10] According to the conventional method of pulse widthmodulation, when the input

PCM audio signal is 0, bias currents of the loadconnected to the differential output can be controlled. However, residualnoises due to switching exist in each of the PWM+ and the PWM-signals. Further,some currents flow through a common ground, and unnecessary heat is generated.Thus, amplification efficiency decreases, and pop noises, which may damage aspeaker and offend a user's ear, are generated when the amplifier operates.

Disclosure of Invention Technical Problem

[11] The present invention provides a method and apparatus for output pulse width modulation in a digital switching amplifier capable of dividing an input audio signal according to a sign of the input audio signal value, converting values of the divided audio signals into positive, and performing pulse width modulation on each of the positive audio signals, thereby reducing pulse width of a differential output applied to load as much as possible in a case where a value of the input audio signal is practically

0.

Technical Solution

[12] According to an aspect of the present invention, there is provided a method of output pulse width modulation in a digital amplifier comprising: a determining step of adding or subtracting a value corresponding to a predetermined offset to or from an

input audio signal alternately according to a period of the input audio signal, thereby determining a sign of a resulting value of adding or subtracting; a switching step of outputting the input audio signal to a first port when the determined sign is positive, and outputting the input audio signal to a second port after inverting the input audio signal into a positive number when the determined sign is negative; an adjusting step of adding a value corresponding to a pulse width for a minimum time required for switching to the audio signals output from the first port and the second port, and outputting a resulting value of adding; and a modulating step of performing pulse width modulation on the audio signals output from the adjusting step and outputting a first modulation signal and a second modulation signal.

[13] According to another aspect of the present invention, there is provided an apparatus for output pulse width modulation in a digital amplifier comprising: an offset adder adding or subtracting a value corresponding to a predetermined offset to or from an input audio signal alternately according to a period of the input audio signal; a sign determining unit determining a sign of a value output from the offset adder; an inverter and a switching unit outputting the audio signal to a first port when the determined sign is positive, and outputting the input audio signal to a second port after inverting the input audio signal into a positive number when the detected sign is negative; a minimum pulse width compensator adding a value corresponding to a pulse width for a minimum time required for switching to the audio signals output from the first port and the second port, and outputting resulting values of adding; and a modulation unit performing pulse width modulation on the audio signals output from the minimum pulse width compensator and outputting a first modulation signal and a second modulation signal. Brief Description of the Drawings

[14] FlG. 1 is a view showing a pulse widthmodulationcircuit in a conventional AD drive system;

[15] FlG. 2 is a view showing a pulse widthmodulationcircuit in a conventional BD drive system;

[16] FlG. 3 is a view showing a pulse widthmodulationoutput signal according to a conventional BD drive system;

[17] FlG. 4 is a view showing a structure of adigitals witching amplifier according to a preferred embodiment of thepresentinvention;

[18] Fig. 5 is a circuit diagram showing a postfilteraccording to an embodiment of the present invention;

[19] FlG. 6 is a view showing a structure of a PWMunitaccording to a first embodiment of the present invention;

[20] FlG. 7 is a view showing a structure of a PWMunitaccording to a second embodiment of the present invention;

[21] FlG. 8 is a view showing a switching unitaccordingto an embodiment of the present invention;

[22] FlG. 9 is a view showing a structure of adigitalswitching amplifier using a minimum pulse width according to a switchingtimefor a switching unit;

[23] FlG. 10 is a view showing a structure of adigitalswitching amplifier adding an offset of which signs are alternatelychangedaccording to an input period of an input audio signal to the inputaudiosignal;

[24] FlG. 11 is a view showing an output signal oftheswitching amplifier according to the FlG. 10;

[25] FlG. 12 is a flowchart showing a method ofpulsewidth modulation according to a preferred embodiment of the presentinvention;and

[26] FlG. 13 is a flowchart showing a method ofpulsewidth modulation according to another embodiment of the presentinvention. Best Mode for Carrying Out the Invention

[27] Now, exemplary embodiments according to thepresentinvention will be described in detail with reference to theaccompanyingdrawings.

[28] FlG. 4 is a view showing a structure of adigitalswitching amplifier according to a preferred embodiment of thepresentinvention;

[29] A digital switching amplifier includesanoversampling filter 310, a pulse width modulation (PWM) unit 320, anH-bridge330, a post filter 340, and a single power supply 360. The oversamplingfilter310 receives a pulse code modulation (PCM) audio signal and increasessamplingrate of the PCM audio signal, so that the oversampling filter 310outputs theoversampled PCM audio signal. For example, when a 16-bit PCM audiosignalhaving a sampling rate of 48 KHz is received, the oversampling filter310performs octuple oversampling and converts the 16-bit PCM audio signalinto24-bit PCM audio signal having a sampling rate of 384 KHz and output- stheoversampled PCM audio signal. The PWM unit 320 receives the oversam- pledPCMaudio signal and performs pulse width modulation, so that the PWM unit320outputs PWM+ and PWM- signals. Detailed structure and operations of thePWMunit 320 will be described later.

[30] The H-bridge 330 is a switching device ofwhichpower is supplied by the single power supply 360 to output the amplifiedPWM+and the PWM- signals. The post filter 340 is a kind of low pass filter.Thepost filter 340 removes carriers from the pulse width modulated signalsandoutputs the carrier-removed pulse width modulated signals. The- serecoveredsignals are input into a speaker 350.

[31] FlG. 5 is a circuit diagram showing the post filter340 according to the embodiment of the present invention.

[32] In FlG. 5, the post filter 340 includes an inductor(L) 410 and a capacitor (C) 420. In some cases, the post filter 340 may beomitted.

[33] FlG. 6 is a view showing a structure of the PWMunit 320 according to a first embodiment of the present invention;

[34] The PWM unit includes a sign determining unit 510,a switching unit 520-1, an inverter 530, a first pulse width modulator 540, anda second pulse width modulator 550. The switching unit 520-1 and the inverter530 are included in a signal separating unit 50. The first and the second pulsewidth modulators 540 and 550 are included in a modulator 55. The signdetermining unit 510 determines a sign of the input PCM audio signal. The mostsignificant bit of the PCM audio signal represented as 2's comple- mentsindicates a sign. For example, when the most significant bit of a PCM value isO, the PCM value is a positive number. When the most significant bit of a PCMvalue is 1, the PCM value is a negative number. Therefore, the sign determiningunit 510 reads the most significant bit of an input PCM value and determines asign of the input PCM audio signal. The determined sign information istransmitted to the switching unit 520-1.

[35] The input PCM audio signal is not only transmittedto the switching unit 520-1 but also inverted in the inverter 530. By theinversion, the negative signal is inverted into a positive signal having thesame magnitude. The inversion is, for example, multiplication a PCM audiosignal by -1. The inverted PCM signal is also transmitted to the switching unit520-l. The switching unit 520-1 selectively switches the original PCM signaland the inverted PCM signal according to the sign information transmitted fromthe signal determining unit 510. For example, when a value of the PCM is apositive number, that is, the most significant bit is 0, the original PCMsignal is output. When a value of the PCM is a negative number, that is, themost significant number is 1, the inverted PCM signal is output. Therefore, allthe waveforms output from the switching unit 520-1 have positive PCMvalues.

[36] The first and the second pulse width modulators 540and 550 perform pulse width modulation on the positive PCM signal and theinverted PCM signal output from the switching unit 520-1, respectively. In thepulse width modulation, there is no need to allocate the pulse width to a valueof a negative PCM. Therefore, in order to increase precision, the maximum pulsewidth is allocated to the maximum value of the input signal, and the pulsewidth of 0 is allocated to the input signal having a value of 0. Ac- cordingly,the first pulse width modulator 540 outputs a PWM+ signal, and the second pulsewidth modulator 550 outputs a PWM- signal. Referring to FlG. 6, both the outputPWM+ and PWM- signals are in the positive direction.

[37] FlG. 7 is a view showing a structure of the PWMunit 320 according to a second embodiment of the present invention;

[38] referring to FlG. 7, the inverter 530 may bedisposed at the back of the switching unit 520-2 not in front of the switchingunit 520-2. The signals input to the first and second pulse width modulators540 and 550 are the same ones in FlG. 6, so the output PWM+ and PWM- signalsare the same as those in FlG. 6 in the end. Alternatively, a function of theinverter 530 may be included in the second pulse width modulator 550, so thatthe second pulse width modulator 550 inverts a negative audio signal into apositive audio signal and performs pulse width modulation equal to that of thefirst pulse width modulator 540.

[39] FlG. 8 is a view showing a switching unit accordingto an embodiment of the present invention;

[40] In FlG. 8, the switching unit can be embodied by amultiplexer. The PWM unit shown in FlG. 6 can generate PCM signals input to thefirst and the second pulse width modulators 540 and 550 by using a first and asecond MUX 610 and 620 instead of a switch. That is, the PCM audio signal istransmitted to '0' input port of the first MUX 610, and the PCM audio signalinverted by the inverter 520 is transmitted to T input port of the second MUX620. A value of '0' is input to the T input port of the first MUX 610 and the'O' input port of the second MUX 620. Sign information of the PCM signal isinput to a selecting port SEL.

[41] Accordingly, when the sign information is ¹ O', theinput PCM signal is output to an output port of the first MUX 610 by the firstMUX 610, and the second MUX 620 outputs a value of '0' which is connected tothe '0' input port. On the other hand, when the sign information is T, avalue of '0' which is connected to the T input port is output by the firstMUX 610, and the second MUX 620 outputs the inverted PCM signal which isconnected to the T input port. Therefore, the signals are the same as thepositive and negative PCM signals output by switching (the signal input tosecond MUX is a positive PCM signal which is the inverted original PCMsignal).

[42] FIG. 9 is a view showing a structure of a digitalswitching amplifier using a minimum pulse width according to a switching timefor a switching unit;

[43] In outputs of the switching unit 520-1, there is aminimum pulse width which can be represented due to a delay of a switching unititself when ON/OFF are repeated. That is, a pulse having a smaller width than alength of a minimum time required for switching cannot be output correctly byswitching. Therefore, in order to prevent signal from being distorted byswitching, pulse widths output by switching have to be larger than the minimumpulse width. So, the minimum pulse width compensator 70 adds a value of a PCMfor the minimum pulse width to the positive PCM signal and the inverted PCMsignal output from the switching unit 520-1, respectively, and transmits themto

the first and the second pulse width modulators 540 and 550. When pulsewidth modulation is performed on the signals compensated for the minimum pulsewidth, even if a value of an input PCM audio signal is 0, a pulse having pulsewidth corresponding to the minimum pulse width is output.

[44] FlG. 10 is a view showing a structure of a digitalswitching amplifier adding an offset of which signs are alternately changedaccording to an input period of the input audio signal to the input audiosignal;

[45] When outputs of the positive and negative PCMsignals are switched substantially simultaneously by the switching unit 520-1, the switching device in the H-bridge 330 generates interference, so the outputsignal is distorted. Due to distortion of the output signals, noises occur inan effective bandwidth, so that a signal to noise ratio decreases. Therefore, so as not to simultaneously switch the outputs of the positive and negative PCMsignals, offsets with alternating signs are provided to the positive andnegative PCM signals, so that the pulse widths of the PCM signals are differentfrom each other. An offset adder 80 adds or subtracts a predetermined value ofoffset to or from the input PCM audio signals alternately according to an inputperiod of the input PCM audio signal. Pulse width modulation signals, which area PWM+ and a PWM-, in a case where the aforementioned offset and the minimumpulse width are added, are described in detail with reference to FlG. 11

[46] FlG. 11 is a view showing an output signal of theswitching amplifier according to the FlG. 10;

[47] Referring to FlG. 11, in the case where theaforementioned offset and the minimum pulse width by switching are added, pulseshapes of the PWM+ and the PWM- signals become different from preceding shapesas the positive and negative offsets are added to the input PCM signalalternately. A pulse shape of a differential output resulting of subtractingthe PWM- signal from the PWM+ signal become a pulse having a narrow pulse widthvarying according to a value of the offset.

[48] FlG. 12 is a flowchart showing a method of pulsewidth modulation according to a preferred embodiment of the present invention;and

[49] First, a PCM signal is received (SlOlO), and it isdetermined whether a sign of the

PCM is positive or negative (S 1020). Thedetermination is performed by reading a most significant bit of the input PCMsignal. In a case where the PCM signal is represented as 2's complement, if themost significant bit is ¹ O', the PCM signal is a positive number. If the mostsignificant bit is T, the PCM signal is a negative number. When it isdetermined that the PCM is a positive number (S 1030), the input PCM signal is- modulated and output as a PWM+ signal. When it is determined that the PCM isnot a positive number, the received PCM signal is inverted by multiplying -1(S 1050), and pulse width modulated and output as PWM- signal (S 1060).

[50] FlG. 13 is a flowchart showing a method of pulsewidth modulation according to another embodiment of the present invention.

[51] Referring to FIGS. 9 and 10, the method of pulsewidth modulation, in a case where a pulse width and an offset are added, isdescribed. After a PCM signal is received (SlOlO), a step of adding andsubtracting a predetermined offset to or from the input PCM audio signalalternately according to an input period of the signal is added (S 1015). It isdetermined whether a sign of adding or subtracting the offset is positive omegative (S 1020). It is determined whether or not a resulting value of theoffset addition or subtraction is positive PCM value (S1030). If the resultingvalue is positive PCM value, a value of PCM corresponding to a minimum pulsewidth for switching is added thereto (S 1032), and the pulse width modulation isperformed thereon, so that a PWM+ signal is output (S 1040). If a resultingvalue of the offset addition or subtraction is not positive PCM value, a valueof PCM corresponding to a minimum pulse width for switching is also addedthereto (S 1034), and the received PCM signal is inverted so as to be convertedinto a positive PCM value (S 1050), and the pulse width modulation is performedthereon, so that a PWM- signal is output (S 1060).

[52] The method of pulse width modulation describedabove can be written as computer programs. Codes and code segments foraccomplishing the present invention can be easily construed by programmersskilled in the art to which the present invention pertains. Also, the programsare stored in computer readable media, and embody the method of pulse widthmodulation by being read and executed by a computer system. Examples of thecomputer readable media include magnetic recording media, optical recordingmedia, and carrier wave media.

[53] While the present invention has been particularlyshown and described with reference to exemplary embodiments thereof, it will beunderstood by those skilled in the art that various changes in form and detailsmay be made therein without departing from the spirit and scope of theinvention as defined by the appended claims. The exemplary embodiments shouldbe considered in descriptive sense only and not for purposes of limitation.Therefore, the scope of the invention is defined not by the de- taileddescription of the invention but by the appended claims, and all differenceswithin the scope will be construed as being included in the presentinvention. Industrial Applicability

[54] As describe above, according to thepresentinvention, residual noises output from a switching amplifier decreaseand biascurrents applied to load decrease, thereby reducing heat generationandenhancing amplification efficiency. Furthermore, pop noises generatedatoperation (ON/OFF) of the amplifier decrease.