WO2020081220A1

WO2020081220A1 - Chopped triangular wave pwm quantizer and pwm modulator having quantizer with controllable analog gain and calibratable for multi-non-ideal gain-affecting characteristics

Info

Publication number: WO2020081220A1
Application number: PCT/US2019/053735
Authority: WO
Inventors: Ramin Zanbaghi; John Melanson; Kyehyung Lee; Anuradha PARSI
Original assignee: Cirrus Logic International Semiconductor Ltd
Current assignee: Cirrus Logic International Semiconductor Ltd
Priority date: 2018-10-14
Filing date: 2019-09-30
Publication date: 2020-04-23
Anticipated expiration: 2021-04-14
Also published as: GB2610917B; US10819328B2; GB2610917A; US20210044285A1; GB202104544D0; GB2590585B; CN113016138A; US20200119702A1; GB2610918A; GB202210456D0; KR20210090166A; GB2590585A; CN113016138B; GB202210457D0; GB2610918B; US20200119725A1; KR102690263B1; US11296685B2; US10659029B2

Abstract

An apparatus in a PWM modulator includes a triangular wave generator that generates a triangular wave and a comparator that is responsive to a signal input to generate a signal output. An output of the PWM modulator is responsive to the comparator signal output. A polarity inversion circuit, coupled between the triangular wave generator and the comparator, is configured in one of the following ways: to provide the triangular wave to the comparator when the triangular wave has a first slope polarity and to provide a polarity-inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and to provide the signal input to the comparator when the triangular wave has the first slope polarity and to provide a polarity-inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

Description

CHOPPED TRIANGULAR WAVE PWM QUANTIZER AND PWM MODULATOR HAVING QUANTIZER WITH CONTROLLABLE ANALOG GAIN AND

C ALIBRAT ABLE FOR MULTI-NON-IDEAL GAIN- AFFECTING CHARACTERISTICS

BACKGROUND

[0001] Personal audio devices, including wireless telephones, such as mobile/cellular telephones, cordless telephones, mp3 players, and other consumer audio devices, are in widespread use. Such personal audio devices may include circuitry for driving a pair of headphones or one or more speakers. Such circuitry often includes a power amplifier for driving an audio output signal to headphones or speakers. Generally speaking, a power amplifier amplifies an audio signal by taking energy from a power supply and controlling an audio output signal to match an input signal shape but with a larger amplitude.

[0002] One example of an audio amplifier is a class-D amplifier. A class-D amplifier (also known as a“switching amplifier”) may comprise an electronic amplifier in which the amplifying devices (e.g., transistors, typically metal-oxide-semiconductor field effect transistors) operate as electronic switches. In a class-D amplifier, a signal to be amplified may be converted to a series of pulses by pulse-width modulation, pulse-density modulation, or another method of modulation, such that the signal is converted into a modulated signal in which a characteristic of the pulses of the modulated signal (e.g., pulse widths, pulse density, etc.) is a function of the magnitude of the signal. After amplification with a class-D amplifier, the output pulse train may be converted to an unmodulated analog signal by passing through a passive low-pass filter, wherein such low-pass filter may be inherent in the class-D amplifier or a load driven by the class-D amplifier. Class-D amplifiers are often used due to the fact that they may be more power efficient than linear analog amplifiers, in that class-D amplifiers may dissipate less power as heat in active devices as compared to linear analog amplifiers. Typically, a pulse-width modulation (PWM) amplifier is chosen in order to provide accurate load voltage with desirable Total Harmonic Distortion (THD) and Power Supply Rejection Ratio (PSRR).

SUMMARY

[0003] In one embodiment, the present disclosure provides an apparatus in a pulse width modulation (PWM) modulator. The apparatus includes a triangular wave generator that generates a triangular wave and a comparator that is responsive to a signal input to generate a signal output. An output of the PWM modulator is responsive to the comparator signal output. The apparatus also includes a polarity inversion circuit, coupled between the triangular wave generator and the comparator, that is configured in one of the following ways: to provide the triangular wave to the comparator when the triangular wave has a first slope polarity and to provide a polarity-inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and to provide the signal input to the comparator when the triangular wave has the first slope polarity and to provide a polarity-inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

[0004] In one embodiment, the present disclosure provides a method in a pulse width modulation (PWM) modulator including a triangular wave generator that generates a triangular wave, a comparator that is responsive to a signal input to generate a signal output, wherein an output of the PWM modulator is responsive to the comparator signal output, and a polarity inversion circuit coupled between the triangular wave generator and the comparator. The method includes performing one of the following actions by the polarity inversion circuit: providing the triangular wave to the comparator when the triangular wave has a first slope polarity and providing a polarity-inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and providing the signal input to the comparator when the triangular wave has the first slope polarity and providing a polarity-inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

[0005] In one embodiment, the present disclosure provides a non-transitory computer-readable medium having instructions stored thereon that are capable of causing or configuring a pulse width modulation (PWM) modulator including a triangular wave generator that generates a triangular wave, a comparator that is responsive to a signal input to generate a signal output, wherein an output of the PWM modulator is responsive to the comparator signal output, and a polarity inversion circuit coupled between the triangular wave generator and the comparator, to perform one of the following actions by the polarity inversion circuit: providing the triangular wave to the comparator when the triangular wave has a first slope polarity and providing a polarity-inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and providing the signal input to the comparator when the triangular wave has the first slope polarity and providing a polarity-inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

[0006] In one embodiment, the present disclosure provides a closed loop pulse width modulation (PWM) modulator includes a speaker driver that selectively receives distinct first and second PWM drive voltage swing ranges and a quantizer having an analog gain and that generates a PWM output signal to the speaker driver. The quantizer and speaker driver have a combined gain. While operating in a first mode in which the first PWM drive voltage swing range is supplied to the speaker driver, the analog gain of the quantizer is controlled to be a first gain value. While operating in a second mode in which the second PWM drive voltage swing range is supplied to the speaker driver, the analog gain of the quantizer is controlled to be a second gain value distinct from the first gain value. The first and second gain values of the analog gain of the quantizer cause the combined gain of the quantizer and driver to be approximately equal in the first and second modes.

[0007] In one embodiment, the present disclosure provides a method that includes operations performed in a closed loop pulse width modulation (PWM) modulator including a speaker driver that selectively receives distinct first and second PWM drive voltage swing ranges and a quantizer having an analog gain and that generates a PWM output signal to the speaker driver, wherein the quantizer and speaker driver have a combined gain. The method includes controlling the analog gain of the quantizer to be a first gain value while operating in a first mode in which the first PWM drive voltage swing range is supplied to the speaker driver and controlling the analog gain of the quantizer to be a second gain value while operating in a second mode in which the second PWM drive voltage swing range is supplied to the speaker driver. The second gain value is distinct from the first gain value. The first and second gain values of the analog gain of the quantizer cause the combined gain of the quantizer and driver to be approximately equal in the first and second modes.

[0008] In one embodiment, the present disclosure provides a pulse width modulation (PWM) modulator includes a quantizer having a ramp generator and a comparator. The quantizer has a gain and at least two measurable non-ideal characteristics that affect the gain. The quantizer is adjustable using measured first and second values to correct for respective first and second of the at least two non-ideal characteristics. The gain of the quantizer is calibratable while the quantizer is adjusted using the measured first and second measured values. [0009] In one embodiment, the present disclosure provides a method performed in a pulse width modulation (PWM) modulator including a quantizer having a ramp generator and a comparator, the quantizer having a gain and at least two non-ideal characteristics that affect the gain. The method includes measuring a first value of a first of the at least two non-ideal characteristics, measuring a second value of a second of the at least two non-ideal characteristics, adjusting the quantizer using the measured first and second values to correct for the first and second of the at least two non-ideal characteristics, and calibrating the gain of the quantizer while the quantizer is adjusted using the measured first and second measured values.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 is an illustration of an example personal audio device.

[0011] Figure 2 is a block diagram of selected components of an example audio IC of a personal audio device.

[0012] Figure 3 is a block diagram illustrating an example class-D modulator based on a pulse-width modulation (PWM) signal.

[0013] Figure 4 is a block diagram illustrating in more detail the PWM quantizer of Figure 3.

[0014] Figure 5A is a block diagram illustrating the composition of an example combined linear gain of a quantizer and speaker driver.

[0015] Figure 5B is a block diagram illustrating the composition of an example combined linear gain of a quantizer and speaker driver in a quantizer having an adjustable analog gain capability.

[0016] Figure 6 is a table describing example gains and gain-related parameters for different PWM modulator speaker voltage operating modes.

[0017] Figure 7 is a block diagram illustrating an example embodiment of the quantizer of Figure 3 that enables the analog gain KANA to be adjusted in order to compensate for a change in the speaker voltage between operating modes.

[0018] Figure 8 is an example timing diagram that describes operation of the switched capacitor network of Figure 7 to accomplish different analog gain values over different speaker voltage operating modes.

[0019] Figure 9 is a block diagram of an example conventional ramp generator.

[0020] Figure 10 is an example timing diagram illustrating operation of the ramp generator of Figure 9. [0021] Figure 11 is a block diagram illustrating an example portion of a ramp generator of Figure 4 that employs a chopping technique to generate a triangular wave for use in a quantizer of Figure 3 of a PWM modulator of Figure 3.

[0022] Figure 12 is an example timing diagram illustrating operation of the ramp generator of Figure 11

[0023] Figure 13 is an example timing diagram describing operation of the switches of the ramp generator of Figure 11 to accomplish the waveforms described with respect to Figure 12.

[0024] Figure 14 is an example block diagram illustrating a system for calibrating a gain of the quantizer of Figure 3.

[0025] Figures 15A through 15E are example timing diagrams illustrating aspects of gain calibration of the quantizer of Figure 14.

[0026] Figure 16 is a block diagram illustrating an example system for calibrating a gain of the quantizer of Figure 3.

[0027] Figures 17A through 17C are example timing diagrams illustrating calibration of the gain of the quantizer of Figure 3.

[0028] Figure 18 is a flowchart illustrating an example method for calibrating the quantizer of Figure 3.

DETAILED DESCRIPTION

[0029] Referring now to Figure 1, an illustration of an example personal audio device 1, in accordance with embodiments of the present disclosure is shown. Figure 1 depicts personal audio device 1 coupled to a headset 3 in the form of a pair of earbud speakers 8 A and 8B. Headset 3 depicted in Figure 1 is merely an example, and it is understood that personal audio device 1 may be used in connection with a variety of audio transducers, including without limitation, headphones, earbuds, in- ear earphones, and external speakers. A plug 4 may provide for connection of headset 3 to an electrical terminal of personal audio device 1. Personal audio device 1 may provide a display to a user and receive user input using a touch screen 2, or alternatively, a standard liquid crystal display (LCD) may be combined with various buttons, sliders, and/or dials disposed on the face and/or sides of personal audio device 1. As also shown in Figure 1, personal audio device 1 may include an audio integrated circuit (IC) 9 for generating an analog audio signal for transmission to headset 3 and/or another audio transducer (e.g., a loudspeaker). [0030] Referring now to Figure 2, a block diagram of selected components of an example audio IC 9 of a personal audio device, in accordance with embodiments of the present disclosure is shown. In some embodiments, example audio IC 9 may be used to implement audio IC 9 of Figure 1. As shown in Figure 2, a microcontroller core 18 (e.g., a digital signal processor or“DSP”) may supply a digital audio input signal DIG IN to a digital -to-analog converter (DAC) 14, which may convert the digital audio input signal to an analog input signal VIN. DAC 14 may supply analog signal VIN to an amplifier 16 which may amplify or attenuate analog input signal VIN to provide an audio output signal VOUT, which may operate a speaker, headphone transducer, a line level signal output, and/or other suitable output.

[0031] Referring now to Figure 3, a block diagram illustrating a class-D modulator 100 based on a pulse-width modulation (PWM) signal is shown. The class-D modulator 100 (e.g., included in amplifier 16 of Figure 2) receives an input signal VIN (e.g., of Figure 2) provided to a summing element 301, whose output is provided to a loop filter 302, whose output is provided to a quantizer 304, whose output is provided to a speaker driver 306. The output VOUT (e.g., of Figure 2) of the speaker driver 306 is fed back to a negating input of the summing element 301 to form a closed loop. The speaker driver 306 is coupled to a switch 308 that provides a driver supply voltage VSPK to the speaker driver 306. The driver supply voltage VSPK may also be referred to as the drive voltage swing range. The switch 308 is controlled to select either a high-voltage (HV) supply VBAT (e.g., battery supply) or a low-voltage (LV) supply VBST (e.g., boosted battery supply) to provide as the driver supply voltage VSPK, or drive voltage swing range VSPK, to the speaker driver 306. In one embodiment, VBST is 12V for HV mode operation and VBAT is 4V for LV mode operation, although other embodiments are contemplated in which the values of VBST and VBAT are different. In one embodiment, the modulator 100 has HV and LV modes for the purpose of saving power.

[0032] In one embodiment, the class-D modulator 100 loop operates in the analog domain similar to a continuous-time delta-sigma analog-to-digital converter (ADC). Unlike a double-sampling ADC, which is based on pulse-density modulation (PDM), the class-D modulator 100 is based on PWM. The PWM quantizer 304 converts an analog signal into a PWM signal. The quantizer 304 and the speaker driver 306 have respective gain values. In particular, the gain of the speaker driver 306 is proportional to the driver supply voltage VSPK provided to the speaker driver 306.

[0033] Embodiments are described below in which the gain of the quantizer 304 - more specifically an analog gain of the quantizer 304 - is advantageously adjusted commensurate with a change in the speaker driver gain when the driver supply voltage VSPK changes in order to maintain a fixed combined gain of the quantizer 304 and speaker driver 306. As a result, a voltage range and associated ramp slope of a sawtooth waveform (compared with a signal input to generate a PWM output signal) generated by the quantizer 304 may advantageously be kept close and in some embodiments fixed. Embodiments are also described below in which the sawtooth waveform generator employs a chopping technique on an internally generated triangular wave to generate the sawtooth wave which results in reduced ramp capacitor reset times and swing voltages. Finally, embodiments of a calibration method are described in which multiple non-ideal characteristics (e.g., comparator time delay and offset, RC time constant of the ramp generator, etc.) of the quantizer 304 are measured and adjusted to improve the accuracy of the gain calibration of the quantizer 304.

[0034] Referring now to Figure 4, a block diagram illustrating in more detail the PWM quantizer 304 of Figure 3 is shown. The PWM quantizer 304 includes a ramp generator 402 that generates a sawtooth wave-shaped ramp voltage Vramp. In one embodiment, the ramp voltage Vramp is a differential voltage pair comprising a positive ramp voltage Vrampp and negative ramp voltage Vrampm. The PWM quantizer 304 also includes a first switched-capacitor network (“SC Net.”) 404p and a second switched-capacitor network (“SC Net.”) 404n, each of which receives the positive and negative ramp voltages Vrampp and Vrampm. The switched-capacitor network 404p/404n also receive a positive signal input voltage VIP and a negative signal input voltage VIM. The switched-capacitor network 404p/404n operate on the ramp voltages the signal input voltages VIP and VIM to generate respective

outputs provided to respective comparators 406p and 406n, which generate respective outputs Dp and D_n that are provided to a central edge modulation (CEM) block 408. The CEM block 408 generates a differential PWM output signal DPWM having a positive component denoted DPWM_P and a negative component denoted Dpw_Mn which are provided to the speaker driver 306 of Figure 3. The switched- capacitor network 404p/404n according to one embodiment are described in more detail with respect to Figure 7 below.

[0035] Referring now to Figure 5 A, a block diagram illustrating the composition of a combined linear gain KTOTAL of the quantizer 304 and driver 306 gain is shown. As shown, components that contribute to the combined linear gain KTOTAL include an analog gain KANA of the quantizer 304 (e.g., of switched capacitor network 404), a gain KDRV of the speaker driver 306, and a gain KCEM of the CEM block 408, as shown in Figure 5A and in equation (1) below. That is, the combined gain KTOTAL of the quantizer 304 and driver 306 may be defined from linear analysis of the modulator 100 system as proportional to a product of the analog gain KANA, the driver gain KDRV, and the CEM gain KCEM.

KTOTAL ⁼ KANA * K_DRV * K_CEM (1)

[0036] As shown in equation (2) below, the driver gain KDRV is the ratio of the speaker voltage VSPK and the ramp voltage swing range Vramp(fd). Thus, when the operating mode is transitioned from HV mode to LV mode (or vice versa), the driver gain KDRV will be changed (assuming the ramp voltage swing range Vramp(fd) is maintained), and a change in the driver gain KDRV changes the combined gain KTOTAL. A change in the combined gain KTOTAL is undesirable because it changes the dynamics of the modulator 100 loop, e.g., changes the loop dynamics to be non-linear.

[0037] One solution is to keep the driver gain KDRV fixed across mode changes in order to maintain a fixed combined gain KTOTAL, which requires a change of the ramp voltage swing range Vramp₍fd₎ to match the change in the speaker voltage VSPK in order to maintain the combined gain KTOTAL across the two modes, as may be observed from equation (3) which rearranges equation (2) to specify the ramp voltage swing range Vramp₍fd₎ as the ratio of the speaker voltage VSPK and the speaker driver gain KDRV.

[0038] Per equations (4) below, in the example, it is assumed the analog gain KANA is 0.5, the drive gain KDRV is 10, the CEM gain KCEM is 2, the total gain KTOTAL is 10, VBST is 12V for HV mode, and VBAT is 4V for LV mode. Consequently, as shown in equation (5), the ramp voltage swing range Vramp₍fd₎ is l.2vpp in HV mode and 0.4vpp in LV mode, which are summarized in the two left-most columns of the table of Figure 6 denoted HV and LV (conv.).

Assume : K_ANA = 0.5 , K_DRV = 10 , K_CEM = 2 , K_T0TAL = 10, V_BST = 12V , V_BAT = 4V (4)

[0039] However, such a large difference in the ramp voltage swing range Vramp(fd) between the HV and LV modes, l.2vpp to 0.4vpp, i.e., a factor of 3, is also highly undesirable because it may introduce sensitivity in the system. More specifically, because circuits in the PWM modulator 100 that perform voltage to time conversion (e.g., comparators and summing networks), for example, operate based on the sawtooth ramp characteristics, it may be difficult to design such components to operate properly in the two different modes. [0040] Embodiments are described that advantageously, while maintaining a fixed combined gain KTOTAL, reduce the variation (or in some embodiments eliminate it) in the ramp voltage swing range Vramp₍fd₎ and associated ramp slope between the two modes by adjusting the analog gain KANA of the quantizer 304, as shown in Figure 5B, to compensate for changes in the gain of the speaker driver 306 caused by a change of the speaker voltage VS_PK values of the two modes.

[0041] Referring now to Figure 7, a block diagram illustrating an embodiment of the quantizer 304 that enables the analog gain KANA to be adjusted in order to compensate for a change in the speaker voltage VS_PK between operating modes is shown. In the embodiment of Figure 7, the sawtooth wave- shaped ramp voltage V_ramp is coupled to the comparator 406 through a switched capacitor network 404 (e.g., switched capacitor network 404 of Figure 4). The negative ramp voltage V_rampm is coupled to the negative input of the comparator 406 through a switch controlled by a clock signal CLK_f and then through a capacitor C2. The positive ramp voltage is coupled to the positive input of the

comparator 406 through a switch controlled by clock signal CLK_f and then through a capacitor C3. The node at the switch side of capacitor C2 is denoted node X, and the node at the switch side of capacitor C3 is denoted node Y. The other terminal of capacitor C2 is coupled to the negative ramp voltage V_rampm through a switch controlled by a clock signal CLK_r, and the other terminal of capacitor C3 is coupled to the positive ramp voltage through a switch controlled by clock signal CLK_r.

A common mode voltage V_cm is also coupled to the positive and negative inputs of the comparator 406 through switches controlled by clock signal CLK_r.

[0042] The negative input to the comparator 406 is also coupled through a switch controlled by clock signal CLK_f to a node that is coupled to a capacitor Cl and to positive signal input V_IP through a switch controlled by clock signal CLK_r. The positive input to the comparator 406 is also coupled through a switch controlled by clock signal CLK_f to a node that is coupled to a capacitor C4 and to negative signal input V_IM through a switch controlled by clock signal CLK_r. The other terminal of capacitor Cl is coupled to node X, and the other terminal of capacitor C4 is coupled to node Y. Common mode voltage V_cm is also coupled to node X and to node Y through switches controlled by a clock signal CLK_r-hv. Negative signal input V_IM is coupled to node X through a switch controlled by a clock signal CLK_r-i_v, and positive signal input V_IP is coupled to node Y through a switch controlled by a clock signal CLK_r-iv.

[0043] Operation of the switched capacitor network 404 of Figure 7 is described by the timing diagram of Figure 8 to accomplish an analog gain KA_NA of the switched capacitor network 404 having a value of 0.5 when operating in HV mode and a value of 1.0 when operating in LV mode (assuming respective values of 12V and 4V for the speaker voltage VSPK), as shown in the second-from -right column of the table of Figure 6 denoted LV (emb. 1). Advantageously, controlling the respective analog gain KANA values of 0.5 and 1.0 maintains a combined gain KTOTAL of 10 for both operating modes, as shown in the table of Figure 6. That is, generally speaking, the analog gain KANA has a gain value (first gain value) while operating in the HV mode (first mode) and a different gain value (second gain value) while operating in the LV mode (second mode), and the driver gain KDRV has a gain value (third gain value) while operating in the first mode and a different gain value (fourth gain value) while operating in the second mode, and the product of the first and third gain values is approximately the same as the product of the second and fourth gain values advantageously resulting in an approximately fixed combined gain KTOTAL across the two modes. Further advantageously, controlling the respective analog gain KANA values of 0.5 and 1.0 accomplishes ramp voltage swing ranges Vramp(fd) of l .2vpp and 0.8vpp, respectively, for the two modes, which represents a significant reduction in the variation of the ramp voltage swing range V_ramp(fd) over a conventional solution that does not adjust the analog gain KANA. More specifically, as shown in the table of Figure 6, the variation is advantageously reduced from a factor of 3 to a factor of 1.5.

[0044] In the embodiment of Figure 7, the analog gain KANA is adjustable by powers of two, yet the ratio of the speaker voltages VSPK may be non-powers of two, in which case the slope of the ramp voltage swing range V_ramp(fd) may vary somewhat between the HV and LV modes; however, advantageously, the variation is significantly reduced with respect to a conventional amplifier not having the benefit of the variable analog gain KANA embodiments, and the benefits described may still be realized by the variable analog gain KANA embodiments. In other embodiments, the analog gain KANA may be implemented to accomplish a non-power of two, as shown in the right-most column of the table of Figure 6 denoted LV (emb. 2).

[0045] Advantageously, keeping the combined gain of the quantizer 304 and driver 306 close to the same or the same in both HV and LV mode avoids exacerbating non-idealities of the quantizer and exposing its design trade-offs and makes the modulator 100 loop behave similarly independent of the different speaker voltage values VSPK.

[0046] Referring now to Figures 9 and 10, a prior art block diagram illustrating a portion of a ramp generator 902 used to generate a sawtooth wave and a timing diagram illustrating operation of the ramp generator 902, respectively, are shown. The ramp generator 902 includes an amplifier 904 having positive and negative inputs and positive and negative outputs. Coupled in parallel between the positive input and the negative output of the amplifier 904 are a first ramp capacitor C_rampm and a first switch controlled by CLK_rst; additionally, coupled in parallel between the negative input and the positive output of the amplifier 904 are a second ramp capacitor C_rampp and a second switch controlled by CLKrst. When the switches are open (e.g., CLK_rst is deasserted), two push-pull reference currents Iref (generated by current sources) are passing through the ramp capacitors Crampm and generating

respectively negative and positive sloping ramp voltages Vrampm and Vrampp at the respective negative and positive outputs of the amplifier 904. The reference current I_ref is the quotient of a reference voltage Vref and a reference resistance R_ref. Assertion of CLK_rst causes energy stored in the ramp capacitors C_rampm and C_rampp to dissipate causing the ramp voltages Vra_mpp to be reset to a

common mode voltage V_cm before the next ramp cycle, resulting in a pair of complementary polarity sawtooth waves, as shown in Figure 10. A resulting sawtooth wave is generated as the difference between the positive ramp voltage Vrampp and the negative ramp voltage Vrampm. A ramp voltage swing range Vramp(fd) is shown as the difference between the maximum value of Vrampp and the minimum value of Vrampm, which is also shown as the difference between the maximum value of the sawtooth wave and the common mode voltage V_cm.

[0047] A disadvantage of the prior art ramp generator 902 of Figures 9 and 10 that resets the sawtooth wave to the common mode voltage V_cm is that it requires on each cycle a relatively wide reset time that is required to discharge the ramp capacitors Crampm and Crampp, which may reduce the modulation index (MI) associated with a PWM modulator that employs the ramp generator 902. Another disadvantage is that it requires the ramp voltages to reach relatively high voltage swing values. Advantageously, embodiments of an improved ramp generator are described below that address these disadvantages.

[0048] Referring now to Figure 11, a block diagram illustrating a portion of a ramp generator 402 (e.g., ramp generator 402 of Figure 4) that employs a chopping technique to generate a triangular wave for use in a quantizer (e.g., quantizer 304 of Figure 3) of a PWM modulator (e.g., PWM modulator 100 of Figure 3) is shown. The ramp generator 402 includes an amplifier 1104 having positive and negative inputs and positive and negative outputs. Coupled between the positive input and the negative output of the amplifier 1104 is a ramp capacitor C_rampm and coupled between the negative input and the positive output of the amplifier 1104 is a ramp capacitor Crampp. [0049] The ramp generator 402 also includes a chopping block 1106, or chopping switch 1106, whose first and second outputs are coupled to the positive and negative inputs of the amplifier 1104, respectively. The two inputs to the chopping switch 1106 are coupled to receive respective positive and negative reference currents I_refp and Irefm. The positive reference current I_refp is the quotient of positive reference voltage V_refp and a reference resistance R_ref, and the negative reference current I_refm is the quotient of negative reference voltage V_refm and the reference resistance R_ref. The chopping switch 1106 operates as a crossbar switch that may be dynamically controlled to operate in either a pass-through configuration or a cross configuration. In the pass-through configuration, the chopping switch 1106 connects the positive reference current I_ref_P to the positive input of the amplifier 1104 and the negative reference current I_refm to the negative input of the amplifier 1104. In the cross configuration, the chopping switch 1106 connects the positive reference current I_ref_P to the negative input of the amplifier 1104 and the negative reference current I_refm to the positive input of the amplifier 1104. Depending upon the configuration of the chopping switch 1106, the reference currents I_refp and Irefm selectively push-pull through the ramp capacitors C_rampp and C_rampm to generate respective negative and positive ramp voltages V_rampm and V,-.-,m_PP at the respective negative and positive outputs of the amplifier 1104, as described in more detail below.

[0050] A common mode voltage V_cm is selectively coupled to the positive and negative inputs of the amplifier 1104 through switches controlled by a clock signal CLK_rst. The node holding the negative ramp voltage V_rampm is selectively coupled to the negative reference voltage Vrefm through a switch controlled by a clock signal CLK_rst _¾ii and is selectively coupled to the positive reference voltage V_refp through a switch controlled by a clock signal CLK_rst rise. The node holding the positive ramp voltage Vrampp is selectively coupled to the negative reference voltage Vrefm through a switch controlled by the clock signal CLK_rst rise and is selectively coupled to the positive reference voltage V_refp through a switch controlled by the clock signal CLK_rst f_aii.

[0051] The ramp generator 402 also includes a de-chopping block 1108, or de-chopping switch 1108, whose first and second inputs are coupled to the negative and positive outputs of the amplifier 1104, respectively. The two outputs of the de-chopping switch 1108 are coupled to nodes that hold respective negative and positive output voltages V_outm and V_outp that may be provided to other portions of a quantizer (e.g., to switched capacitor network 404 for provision to comparator 406 of Figure 4 of quantizer 304 of Figure 3) for use in generating a PWM signal, e.g., for provision to a speaker driver (e.g., speaker driver 306 of Figure 3) for driving a speaker for audio generation. The de-chopping switch 1108 operates as a crossbar switch that may be dynamically controlled to operate in either a pass-through configuration or a cross configuration, similar to the chopping switch 1106. In the pass- through configuration, the de-chopping switch 1108 connects the negative ramp voltage Vrampm to the negative output voltage Voutm node and the positive ramp voltage V,-.-,m_PP to the positive output voltage Voutp node. In the cross configuration, the de-chopping switch 1108 connects the negative ramp voltage Vrampm to the positive output voltage Voutp node and the positive ramp voltage V_r.-,m_PP to the negative output voltage V_outm node.

[0052] Referring now to Figure 12, a timing diagram illustrating operation of the ramp generator 402 of Figure 11 is shown. At the beginning of a first sampling period, CLK_rst is asserted to close the switches to apply the common mode voltage V_cm to the inputs of the amplifier 1104. Additionally, CLKrst fail is asserted to apply the negative reference voltage V_refm to the node holding the negative ramp voltage to apply the positive reference voltage V_refp to the node holding the positive

ramp voltage Vrampp. Still further, the chopping switch 1106 is controlled to be in the cross configuration to connect the positive reference current I_ref_P to the negative input of the amplifier 1104 and to connect the negative reference current I_refm to the positive input of the amplifier 1104. The chopping switch 1106 is maintained in the cross configuration through the sampling period to cause the voltage across ramp capacitor C_rampm to increase which increases the negative ramp voltage

from its negative peak (approximately V_cm minus V_refm) to its positive peak (approximately V_cm plus Vrefp) and to cause the voltage across ramp capacitor Cram_PP to decrease which decreases the positive ramp voltage Vrampp from its positive peak (approximately V_cm plus V_refp) to its negative peak (approximately V_cm minus V_refm), as shown. The result is a positively sloping ramp voltage Vramp (shown in Figure 12 as a non-differential voltage Vrampm minus Vrampp) having a swing of approximately the difference between the positive reference voltage V_refp minus the negative reference voltage Vrefm, referred to as Vramp(fd), which is a first period of what will become a triangular wave, as shown. Further during the first sampling period, the de-chopping switch 1108 is controlled to be in the pass-through configuration to cause the negative ramp voltage V_rampm to be provided as the negative output voltage Voutm and to cause the positive ramp voltage Vrampp to be provided as the positive output voltage V_outp. The operation of the de-chopping switch 1108 in pass-through configuration results in the positively sloping ramp voltage Vramp of swing Vramp(fd) being passed through and provided on the output voltage VOUT, which is a first period of what will become a sawtooth wave, as shown. [0053] As will be understood from the following description, the ramp voltage Vramp is a negatively sloping ramp also having a swing of Vramp(fd) during the next sampling period, which will be polarity- inverted to form another positively sloping ramp of the sawtooth wave of the output voltage VOUT. This pattern repeats for subsequent sampling period pairs, resulting in a triangular wave ramp voltage Vramp, which is polarity-swapped on alternating periods (e.g., on periods in which the triangular wave ramp voltage negatively sloping) by the de-chopping switch 1 108 to form the sawtooth wave

on the output voltage VOUT, as shown. As a result, the ramp generator 402 advantageously enjoys the benefits perceived by smaller swings and reset times of the ramp capacitors, as described in more detail below.

[0054] In the next (second) sampling period, CLK_rst is again asserted to apply the common mode voltage Vcm to the inputs of the amplifier 1 104. Additionally, CLK_rst rise is asserted to apply the positive reference voltage V_refp to the node holding the negative ramp voltage V_rampm and to apply the negative reference voltage V_refm to the node holding the positive ramp voltage V_rampp. This manner of operation has the advantage of resetting the respective ramp voltages to the relevant reference voltage in the event that the respective ramp voltages at their peak did not reach the relevant reference voltage during their run up/down, which may help avoid drift away from the relevant reference voltages. Still further, the chopping switch 1 106 is controlled to be in the pass-through configuration to connect the positive reference current I_refp to the positive input of the amplifier 1 104 and to connect the negative reference current I_refm to the negative input of the amplifier 1 104. The chopping switch 1 106 is maintained in the pass-through configuration through the second sampling period to cause the voltage across ramp capacitor C_rampm to decrease which decreases the negative ramp voltage Vrampm from its positive peak to its negative peak and to cause the voltage across ramp capacitor C_rampp to increase which increases the positive ramp voltage Vrampp from its negative peak to its positive peak, as shown. Further during the second sampling period, the de-chopping switch 1 108 is controlled to be in the cross configuration to cause the negative ramp voltage Vrampm to be provided as the positive output voltage Voutp and to cause the positive ramp voltage Vrampp to be provided as the negative output voltage Voutm. As shown in Figure 12, a negatively sloping ramp voltage Vramp is provided to the de- chopping switch 1 108 as a second period of a triangular wave having a swing of approximately Vramp(fd). Furthermore, operating in the cross configuration, the de-chopping switch 1 108 polarity- inverts the received negatively sloping ramp voltage Vra_mp and provides the polarity-inverted version thereof as a positively sloping ramp having a swing of approximately V_ramp(fd) on the output voltage VOUT as a second period of the sawtooth wave. [0055] In the next (third) sampling period, as in the first sampling period, CLK_rst is again asserted to apply the common mode voltage V_cm to the inputs of the amplifier 1104, CLK_{rst fail} is asserted to apply the negative reference voltage V_refm to the node holding the negative ramp voltage V_rampm and to apply the positive reference voltage V_refp to the node holding the positive ramp voltage V_rampp (advantageously resetting the respective ramp voltages to avoid drift away from the reference values), and the chopping switch 1106 is controlled to be in the cross configuration to connect the negative reference current I_refm to the positive input of the amplifier 1104 and to connect the positive reference current I_refP to the negative input of the amplifier 1104, which is maintained through the third sampling period to increase the negative ramp voltage V_rampm from its negative peak to its positive peak and to decrease the positive ramp voltage V_rampp from its positive peak to its negative peak, as shown. Further during the third sampling period, the de-chopping switch 1108 is controlled to be in the pass-through configuration to cause the positive ramp voltage V_rampp to be provided as the positive output voltage Vou_tp and to cause the negative ramp voltage Vrampm to be provided as the negative output voltage V_outm. As shown in Figure 12, a positively sloping ramp voltage V_ramp of a third period of a triangular wave is provided to and passed through by the de-chopping switch 1108 to provide a positively sloping ramp having a swing of approximately Vramp_(fd) on the output voltage VOUT as a third period of the sawtooth wave.

[0056] The operation of the various switches of the ramp generator 402 of Figure 11 to accomplish the waveforms described with respect to Figure 12 is shown in the timing diagram of Figure 13.

[0057] Advantages of the use of a chopping technique to convert a triangular wave, employed internal to the ramp generator 402, to produce a sawtooth wave may now be described. First, the purpose of the PWM modulator is to convert continuous voltage domain information into time domain information in which the time domain has a period T. A large reset time in the conventional ramp generator 902 of Figure 9 (i.e., time required to reset the capacitors to the common mode voltage) reduces the usable portion of the period T, which may result in a loss of dynamic range, for example. In a lower voltage mode in which the duty cycle may need to be increased to provide sufficient pulse energy, the large reset time may particularly exacerbate operation since it may detrimentally affect the ability to increase the duty cycle, for example. Advantageously, embodiments have been described that reduce the reset time. Because in the embodiment of Figures 11 and 12, the ramp polarity is flipped, or inverted, at the end of each period, no sudden discharge of the ramp capacitors occurs and no large reset time is required. The reduced capacitor reset time regains the portion of the period lost by the conventional approach and may therefore improve the dynamic range of the PWM modulator 100. Second, due to the pre-charging of the ramp capacitors, the voltage swing of the ramp is reduced to approximately half the maximum voltage swing, as may be observed from Figure 12. This voltage swing reduction may result in power savings especially for low voltage designs. Finally, as may be observed from Figure 12, a sawtooth wave of swing Vramp(fd) (Vrefp minus V_refm) is provided on the output voltage VOU_T of the ramp generator 402 similar to the sawtooth wave of Figure 10, which advantageously enables reuse of various system elements (e.g., comparator 406 or the switch capacitor network 404) with the added benefits of reduced capacitor reset time and swing values as a result of the manner in which the ramp generator 402 produces a sawtooth wave from an internally-generated triangular wave.

[0058] In an alternate embodiment, the de-chopping switch 1108 provides a polarity-inverted version of the signal input V_IN to the comparator 406 on alternating sampling periods, rather than a polarity- inverted version of the triangular wave, which may effectively accomplish a similar result.

[0059] Referring now to Figure 14, a block diagram illustrating a system for calibrating a gain of a quantizer (e.g., quantizer 304 of Figure 3) is shown. The system includes a ramp generator 402, switched capacitor network 404, and comparator 406 (e.g., of Figure 4). The system also includes a D flip-flop 1407 and a voltage reference generator 1405 that provides the reference voltage V_refp and Vrefm to the ramp generator 402. The ramp generator 402 provides the ramp voltage Vrampp and Vrampm to the switched capacitor network 404 which also receives the input signal V_IP and V_IM. The comparator 406 outputs a value Dp which is provides as the data input to the D flip-flop 1407. The output of the D flip-flop 1407 is value Deal. The D flip-flop 1407 is clocked by a calibration clock CLKcal.

[0060] Referring now to Figures 15A through 15E, timing diagrams illustrating aspects of gain calibration of the quantizer 304 of Figure 14 are shown. Process, voltage, and temperature variations may change the characteristics of the operation of the quantizer 304, such as the characteristics of the ramp generated by the ramp generator 402. The changes may be due to changes in the characteristics of the reference resistance R_ref and ramp capacitance C_ramp, for example. The ramp voltage as a function of time V_ramp(t) is proportional to the ratio of the reference voltage V_ref and the product of the reference resistance R_ref and the ramp capacitance C_ramp, as shown in equation (6). Gain calibration is performed in order to capture the gain error in the time domain and correct the gain error.

' ref

V R, amp (t) = (6) ref C_ramp [0061] During calibration, the common mode voltage V_cm is connected to the signal inputs, and the comparator 406 compares the ramp voltage Vramp to the common mode voltage V_cm. The calibration reference clock CLK_cai causes the D flip-flop 1407 to latch the comparator 406 output Dp, as shown in Figure 15 A.

[0062] For an ideal ramp voltage Vramp, the crossing occurs at the midpoint of the ramp. If the ramp slope is higher than normal (i.e., higher gain), a trim bit is changed in a step by step manner to reduce the slope of the ramp voltage Vramp until the polarity of D_cai flips, as shown in Figure 15B. Conversely, if the ramp slope is lower than normal (i.e., lower gain), the trim bit is changed in a step by step manner to increase the slope of the ramp voltage the until the polarity of D_cai flips, as shown in Figure

15C.

[0063] Various non-idealities may exist in the quantizer 304. For example, the comparator 406 may have an offset which creates an error term on the gain calibration. The gain calibration process only looks for the crossing and perceives the offset as a gain error even when the slope of the ramp voltage Vmm_p is correct. Adjusting to make the crossing mid-ramp creates an undesired gain error, shown in Figure 15D as an offset voltage Vos.

[0064] For another example, the comparator 406 may have a time delay T_d, shown in Figure 15E, which creates an error term on the gain calibration. The gain calibration process may perceive the time delay Td as a gain error even when the slope of the ramp voltage V_ramp is correct. More specifically, the delay error T_d is treated like a voltage domain error. Adjusting to offset the time delay T_d creates an undesired error in the slope of the ramp voltage as shown.

[0065] Advantageously, embodiments of an improved calibration process are now described that removes the comparator offset V_os and compactor delay Td so that they do not create a gain error term which may advantageously result in a more accurate calibration of the gain of the quantizer 304, e.g., the ramp gain/slope.

[0066] Referring now to Figure 16, a block diagram illustrating an improved system for calibrating a gain of a quantizer (e.g., quantizer 304 of Figure 3) is shown. The system is similar in many respects to the system of Figure 14. However, the system of Figure 16 includes a chopping switch 1605 coupled between the output of the ramp generator 402 and the input of the switched capacitor network 404. Additionally, in the system of Figure 16, the reference voltages V_refp and V_refm are passed through the ramp generator 402 to the chopping switch 1605 such that during the offset calibration process of quantizer 304, ramp voltages are not generated. That is, the ramp generator 402 is transparent to the reference voltages V_refp and V_refm. The chopping switch 1605 may operate in a pass-through configuration to connect the positive voltage reference V_refp to the positive ramp voltage input

of the switched capacitor network 404 and to connect the negative voltage reference V_refm to the negative ramp voltage V_rampm input. Alternatively, the chopping switch 1605 may operate in a cross configuration to connect the positive voltage reference V_refp to the negative ramp voltage Vrampm input of the switched capacitor network 404 and to connect the negative voltage reference V_refm to the positive ramp voltage V_rampp input. The chopping switch 1605 is controlled by a clock signal CLK_{r ramp} and CLK_{f ramp} according to the normal operation of those signals within the ramp generator 402 as described with respect to Figure 13. Thus, by operation of the chopping switch 1605, the ramp voltage inputs to the switched capacitor network 404 are a chopped version of the reference voltage V_refp and

Vrefm.

[0067] During calibration, the common mode voltage V_cm is connected to the signal inputs. Initially, higher than normal values of the reference voltage V_refp and V_refm are provided through the ramp generator 402 to the chopping switch 1605 and are scaled down over time until two consecutive values of the comparator 406 output Dp are the same, at which time the comparator 406 offset Vos is measured, as shown in the timing diagram of Figure 17A.

[0068] Referring now to Figure 17B, a timing diagram illustrating measurement of time delay T_d of the comparator 406. The system set up is similar to that of the offset V_os measurement. However, the smallest value of the reference voltage V_refp and V_refm is provided and remains fixed through the measurement process. The output Dp of comparator 406 is skewed from CLK_{r ramp} by the comparator 406 delay T_d. The calibration clock CLK_cai, whose rising edge is trimmable, retimes the comparator 406 output Dp. The rising edge of the calibration clock CLK_cai starts from the rising edge of CLK_{r ramp} and gets incremented. When the polarity of D_cai flips, the delay T_d is measured, as shown in Figure 17B.

[0069] Referring now to Figure 17C, a timing diagram illustrating calibration of the gain of the quantizer 304 is shown. First, the offset voltage Vos of comparator 406 measured according to the process described with respect to Figure 17A is applied to the reference voltage V_refp and V_refm to code shift them up or down. Additionally, time delay T_d of the comparator 406 measured according to the process described with respect to Figure 17B is applied to the calibration clock CLK_cai to delay the edge of the calibration clock CLK_cai to offset the effect of the comparator 406 delay T_d. Then, when the offset V_os and delay T_d have been adjusted, the gain error of the quantizer 304 (e.g., the ramp slope) is calibrated.

[0070] In addition to the offset and the time delay of the comparator, other non-ideal characteristics of the quantizer 304 that may be separately measured and adjusted may include the RC time constant used by the ramp generator to generate the ramp voltage; a bandgap used in a current source of the quantizer; a resistor, capacitor and/or transistor size ratio that controls a gain of the quantizer; a clock speed of the quantizer; and various circuit parasitic. These other non-ideal characteristics of the quantizer 304 may be measured in addition to, in place of, and/or in combination with the comparator 406 offset V_os and time delay T_d and adjusted for while the gain error of the quantizer 304 is calibrated.

[0071] Referring now to Figure 18, a flowchart illustrating a method for calibrating a quantizer is shown. Operation begins at block 1802.

[0072] At block 1802, a first non-ideality is measured (e.g., comparator offset V_os per Figure 17A). The operation proceeds to block 1804.

[0073] At block 1804, a second non-ideality is measured (e.g., comparator delay T_d per Figure 17B). The operation proceeds to block 1806.

[0074] At block 1806, additional non-idealities may be measured (e.g., RC time constant, current source bandgap, resistor/capaci tor/transistor size ratio clock speed). The operation proceeds to block 1808.

[0075] At block 1808, the quantizer 304 is adjusted using the values of the non-idealities measured at blocks 1802 through 1806. The operation proceeds to block 1812.

[0076] At block 1812, the gain of the quantizer 304 is calibrated while the quantizer 304 is adjusted using the non-ideality measured values.

[0077] It should be understood - especially by those having ordinary skill in the art with the benefit of this disclosure - that the various operations described herein, particularly in connection with the figures, may be implemented by other circuitry or other hardware components. The order in which each operation of a given method is performed may be changed, unless otherwise indicated, and various elements of the systems illustrated herein may be added, reordered, combined, omitted, modified, etc. It is intended that this disclosure embrace all such modifications and changes and, accordingly, the above description should be regarded in an illustrative rather than a restrictive sense.

[0078] Similarly, although this disclosure refers to specific embodiments, certain modifications and changes can be made to those embodiments without departing from the scope and coverage of this disclosure. Moreover, any benefits, advantages, or solutions to problems that are described herein with regard to specific embodiments are not intended to be construed as a critical, required, or essential feature or element.

[0079] Further embodiments likewise, with the benefit of this disclosure, will be apparent to those having ordinary skill in the art, and such embodiments should be deemed as being encompassed herein. All examples and conditional language recited herein are intended for pedagogical objects to aid the reader in understanding the disclosure and the concepts contributed by the inventor to furthering the art and are construed as being without limitation to such specifically recited examples and conditions.

[0080] This disclosure encompasses all changes, substitutions, variations, alterations, and modifications to the example embodiments herein that a person having ordinary skill in the art would comprehend. Similarly, where appropriate, the appended claims encompass all changes, substitutions, variations, alterations, and modifications to the example embodiments herein that a person having ordinary skill in the art would comprehend. Moreover, reference in the appended claims to an apparatus or system or a component of an apparatus or system being adapted to, arranged to, capable of, configured to, enabled to, operable to, or operative to perform a particular function encompasses that apparatus, system, or component, whether or not it or that particular function is activated, turned on, or unlocked, as long as that apparatus, system, or component is so adapted, arranged, capable, configured, enabled, operable, or operative.

Claims

1. An apparatus in a pulse width modulation (PWM) modulator, comprising:

a triangular wave generator that generates a triangular wave;

a comparator that is responsive to a signal input to generate a signal output, wherein an

output of the PWM modulator is responsive to the comparator signal output;

a polarity inversion circuit coupled between the triangular wave generator and the

comparator; and

wherein the polarity inversion circuit is configured in one of the following ways:

to provide the triangular wave to the comparator when the triangular wave has a first slope polarity and to provide a polarity -inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and

to provide the signal input to the comparator when the triangular wave has the first slope polarity and to provide a polarity -inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

2. The apparatus of claim 1,

wherein the triangular wave generator generates the triangular wave having a period; and wherein the polarity inversion circuit provides the triangular wave/signal input and the

polarity -inverted version of the triangular wave/signal input to the comparator during respective halves of the period.

3. The apparatus of claim 2, wherein the triangular wave period corresponds to twice a period of a PWM signal generated by the PWM modulator.

4. The apparatus of claim 2,

wherein the triangular wave generator comprises a chopping switch that selectively couples respective opposite polarity reference current sources to respective opposite polarity inputs of an operational amplifier on alternating periods of the triangular wave.

5. The apparatus of claim 4, wherein the triangular wave generator is configured to reset the inputs of the operational amplifier to a common mode voltage at a beginning and at a middle of the period of the triangular wave.

6. The apparatus of claim 2,

wherein the triangular wave comprises a differential voltage having positive and negative portions; and

wherein the triangular wave generator is configured to reset the positive and negative

portions of the triangular wave differential voltage to respective positive and negative reference voltages on alternating periods of the triangular wave.

7. The apparatus of claim 1,

wherein the polarity-inversion circuit comprises a de-chopping switch.

8. A method, comprising:

in a pulse width modulation (PWM) modulator including a triangular wave generator that generates a triangular wave, a comparator that is responsive to a signal input to generate a signal output, wherein an output of the PWM modulator is responsive to the comparator signal output, and a polarity inversion circuit coupled between the triangular wave generator and the comparator:

performing one of the following actions by the polarity inversion circuit:

providing the triangular wave to the comparator when the triangular wave has a first slope polarity and providing a polarity-inverted version of the triangular wave to the comparator when the triangular wave has a second slope polarity opposite the first slope polarity; and

providing the signal input to the comparator when the triangular wave has the first slope polarity and providing a polarity-inverted version of the signal input to the comparator when the triangular wave has the second slope polarity.

9. The method of claim 8, further comprising:

generating, by the triangular wave generator, the triangular wave having a period; and providing, by the polarity inversion circuit, the triangular wave/signal input and the polarity- inverted version of the triangular wave/signal input to the comparator during respective halves of the period.

10. The method of claim 9,

wherein the triangular wave period corresponds to twice a period of a PWM signal generated by the PWM modulator.

11. The method of claim 9, further comprising:

wherein the triangular wave generator comprises a chopping switch; and

selectively coupling, by the chopping switch, respective opposite polarity reference current sources to respective opposite polarity inputs of an operational amplifier on alternating periods of the triangular wave.

12. The method of claim 11,

resetting, by the triangular wave generator, the inputs of the operational amplifier to a

common mode voltage at a beginning and at a middle of the period of the triangular wave.

13. The method of claim 9, further comprising:

resetting, by the triangular wave generator, the positive and negative portions of the

triangular wave differential voltage to respective positive and negative reference voltages on alternating periods of the triangular wave.

14. The method of claim 8,

wherein the polarity-inversion circuit comprises a de-chopping switch.

15. A non-transitory computer-readable medium having instructions stored thereon that are capable of causing or configuring a pulse width modulation (PWM) modulator including a triangular wave generator that generates a triangular wave, a comparator that is responsive to a signal input to generate a signal output, wherein an output of the PWM modulator is responsive to the comparator signal output, and a polarity inversion circuit coupled between the triangular wave generator and the comparator, to perform operations comprising:

performing one of the following actions by the polarity inversion circuit:

16. The non-transitory computer-readable medium of claim 15, further comprising:

17. The non-transitory computer-readable medium of claim 16,

18. The non-transitory computer-readable medium of claim 16, further comprising:

wherein the triangular wave generator comprises a chopping switch; and

19. The non-transitory computer-readable medium of claim 18,

20. The non-transitory computer-readable medium of claim 16, further comprising: wherein the triangular wave comprises a differential voltage having positive and negative portions; and

21. The non-transitory computer-readable medium of claim 15,

wherein the polarity-inversion circuit comprises a de-chopping switch.

22. A closed loop pulse width modulation (PWM) modulator, comprising:

a speaker driver that selectively receives distinct first and second PWM drive voltage swing ranges;

a quantizer having an analog gain and that generates a PWM output signal to the speaker driver;

wherein the quantizer and speaker driver have a combined gain; wherein while operating in a first mode in which the first PWM drive voltage swing range is supplied to the speaker driver, the analog gain of the quantizer is controlled to be a first gain value;

wherein while operating in a second mode in which the second PWM drive voltage swing range is supplied to the speaker driver, the analog gain of the quantizer is controlled to be a second gain value distinct from the first gain value; and

wherein the first and second gain values of the analog gain of the quantizer cause the

combined gain of the quantizer and driver to be approximately equal in the first and second modes.

23. The closed loop PWM modulator of claim 22,

wherein the speaker driver has a gain that is a third gain value while operating in the first mode and that is a fourth gain value while operating in the second mode; and wherein a product of the first and third gain values is approximately equal to a product of the second and fourth gain values.

24. The closed loop PWM modulator of claim 22, further comprising: a switched capacitor network that generates the first/second gain value to control the analog gain of the quantizer to be the first/second gain value.

25. The closed loop PWM modulator of claim 24,

wherein the switched capacitor network receives an input signal voltage and a ramp voltage and responsively generates a voltage used to generate the PWM output signal, the ramp voltage having a swing range; and

wherein a ratio of the ramp voltage swing range while operating in the first mode to the ramp voltage swing range while operating in the second mode is less than a ratio of the first PWM drive voltage swing range to the second PWM drive voltage swing range.

26. The closed loop PWM modulator of claim 24,

wherein the switch capacitor network is controllable such that a ratio of the first and second analog gain values is a power of two.

27. The closed loop PWM modulator of claim 24,

wherein the switch capacitor network further receives a common mode voltage and generates the voltage used to generate the PWM output signal responsively to the input signal voltage and the ramp voltage and the common mode voltage.

28. A method, comprising:

in a closed loop pulse width modulation (PWM) modulator including a speaker driver that selectively receives distinct first and second PWM drive voltage swing ranges and a quantizer having an analog gain and that generates a PWM output signal to the speaker driver, wherein the quantizer and speaker driver have a combined gain: controlling the analog gain of the quantizer to be a first gain value while operating in a first mode in which the first PWM drive voltage swing range is supplied to the speaker driver;

controlling the analog gain of the quantizer to be a second gain value while operating in a second mode in which the second PWM drive voltage swing range is supplied to the speaker driver, wherein the second gain value is distinct from the first gain value; and wherein the first and second gain values of the analog gain of the quantizer cause the combined gain of the quantizer and driver to be approximately equal in the first and second modes.

29. The method of claim 28,

30. The method of claim 28,

wherein said controlling the analog gain of the quantizer to be the first/second gain value comprises operating a switched capacitor network to generate the first/second gain value.

31. The method of claim 30,

wherein said operating the switched capacitor network to generate the first/second gain value comprises:

receiving, by the switched capacitor network, an input signal voltage and a ramp voltage and responsively generating a voltage used to generate the PWM output signal, the ramp voltage having a swing range; and wherein a ratio of the ramp voltage swing range while operating in the first mode to the ramp voltage swing range while operating in the second mode is less than a ratio of the first PWM drive voltage swing range to the second PWM drive voltage swing range.

32. The method of claim 30,

33. The method of claim 30, further comprising wherein said operating the switched capacitor network to generate the first/second gain value further comprises: receiving, by the switch capacitor network, further a common mode voltage and generating the voltage used to generate the PWM output signal responsively to the input signal voltage and the ramp voltage and the common mode voltage.

34. A pulse width modulation (PWM) modulator, comprising:

a quantizer having a ramp generator and a comparator;

wherein the quantizer has a gain and at least two measurable non-ideal characteristics that affect the gain;

wherein the quantizer is adjustable using measured first and second values to correct for respective first and second of the at least two non-ideal characteristics; and wherein the gain of the quantizer is calibratable while the quantizer is adjusted using the measured first and second measured values.

35. The PWM modulator of claim 34, wherein the at least two non-ideal characteristics are from the set comprising:

a time delay of the comparator;

an offset of the comparator;

a process variation-dependent time constant of a resistor-capacitor (RC) circuit used by the ramp generator to generate a ramp voltage;

a bandgap of a current source of the quantizer;

a resistor, capacitor and/or transistor size ratio;

a clock speed of the quantizer; and

circuit parasitics.

36. The PWM modulator of claim 34,

wherein the first of the at least two non-ideal characteristics comprises a time delay of the comparator; and wherein the second of the at least two non-ideal characteristics comprises an offset of the comparator.

37. The PWM modulator of claim 34, wherein the at least two non-ideal characteristics comprises at least three of the non-ideal characteristics;

wherein the quantizer is adjustable using the measured first, second and third values to correct for respective the first, second and a third of the at least two non-ideal characteristics; and

wherein the gain of the quantizer is calibratable while the quantizer is adjusted using the measured first, second and third measured values.

38. The PWM modulator of claim 34,

wherein the PWM modulator is on a same integrated circuit as a calibration system that measures the first and second values and adjusts the quantizer using the first and second values.

39. A method, comprising:

in a pulse width modulation (PWM) modulator including a quantizer having a ramp

generator and a comparator, the quantizer having a gain and at least two non-ideal characteristics that affect the gain:

measuring a first value of a first of the at least two non-ideal characteristics;

measuring a second value of a second of the at least two non-ideal characteristics; adjusting the quantizer using the measured first and second values to correct for the first and second of the at least two non-ideal characteristics; and

calibrating the gain of the quantizer while the quantizer is adjusted using the

measured first and second measured values.

40. The method of claim 39,

wherein the at least two non-ideal characteristics are from the set comprising:

a time delay of the comparator;

an offset of the comparator;

a process variation-dependent time constant of a resistor-capacitor (RC) circuit used by the ramp generator to generate a ramp voltage; a bandgap of a current source of the quantizer;

a resistor, capacitor and/or transistor size ratio;

a clock speed of the quantizer; and

circuit parasitics.

41. The method of claim 39,

wherein the first of the at least two non-ideal characteristics comprises a time delay of the comparator; and

wherein the second of the at least two non-ideal characteristics comprises an offset of the comparator.

42. The method of claim 39,

wherein the at least two non-ideal characteristics comprises at least three of the non-ideal characteristics;

the method further comprising: adjusting the quantizer using the measured first, second and third values to correct for the first, second and third of the at least two non-ideal characteristics; and calibrating the gain of the quantizer while the quantizer is adjusted using the

measured first, second and third measured values.

43. The method of claim 39,

wherein the PWM modulator and a calibration system that performs said measuring and adjusting are on a same integrated circuit.