Analog modulator and switching amplifier using an analog modulator
The analog modulator, employing a comparator and low-pass filter in a feedback loop with virtual ground, addresses output distortion and instability in digital amplifiers, achieving high accuracy and low distortion in audio applications.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- 佐藤 厚
- Filing Date
- 2022-01-11
- Publication Date
- 2026-07-16
AI Technical Summary
Existing digital audio amplifiers face issues with output distortion, instability, and complexity due to reliance on pulse width modulation (PWM) and delta-sigma modulation, which are susceptible to power supply fluctuations and require complex circuits for noise shaping and high oversampling rates, leading to increased cost and impedance.
An analog modulator using a comparator and a first-order low-pass filter in a negative feedback loop, with a virtual ground and level shift circuit, operates as a PWM modulator controlled by feedback, achieving stable operation and improved output accuracy with a simple structure.
The analog modulator achieves high output accuracy and resistance to power supply fluctuations, enabling low-distortion sound and high-quality audio output, while the switching amplifier exhibits excellent responsiveness and low THD, suitable for audio applications with a single power supply.
Smart Images

Figure 0007891339000002 
Figure 0007891339000003 
Figure 0007891339000004
Abstract
Description
[Technical Field]
[0001] This invention relates primarily to an analog modulator used in audio and a switching amplifier using an analog modulator. [Background technology]
[0002] The advancement of large-scale integrated circuits (LSIs) in semiconductors has led to digital amplifiers replacing traditional analog linear amplifiers due to their compact size and high sound quality. Digital amplifiers are sometimes collectively referred to as Class D amplifiers, a designation based on the operating point of the output stage elements. This means that only the saturation point is used (in other words, the intermediate linear portion of the semiconductor is not used) (i.e., it is used as a switching element), resulting in high power efficiency. The importance of this high efficiency in mobile devices has also spurred the spread of digital amplifiers.
[0003] There are two main types of modulation methods for digital audio amplifiers: pulse width modulation (PWM) and delta-sigma modulation.
[0004] The PWM method consists of a comparator that compares the input signal with a triangular wave, and the output pulse width changes accordingly (for example, Patent Document 1). Because this method inherently lacks feedback control, it is susceptible to the accuracy of the triangular wave generator and the stability of the power supply voltage, and output distortion tends to be large. Therefore, many PWM amplifiers are equipped with separate negative feedback in the audio frequency domain to improve the linearity of the amplifier and make it more resistant to power supply voltage fluctuations.
[0005] The Delta Sigma method has a quantizer and an integrator in the feedback loop (Non-Patent Document 1) and generates a bit stream with a density (frequency) corresponding to an input shaped by an external clock. Since the quantizer (comparator) literally has a resolution of only 1 bit, conversion noise is superimposed, leading to a decrease in the resolution of the time-averaged output. To push most of this conversion noise into a frequency band beyond the audible range (referred to as Noise Shaping) and increase the resolution of the output in the audible range, it is necessary to have a high oversampling rate (usually OSR = 64) and a number of integrators (usually 5th order, up to 7th order at most). As a result, the circuit becomes complex, leading to an increase in cost and a tendency to become unstable.
[0006] In addition to the impedance of the output transistor, the impedance of the subsequent inductor causes an increase in the output impedance, resulting in a poor damping factor (the ability to brake the motion of the speaker). Also, since the output voltage to the speaker depends on the product of the frequency of the bit stream and the supply voltage, there is no way for the feedback to sense fluctuations in the power supply voltage, and special attention must be paid to its stability.
Prior Art Documents
Patent Documents
[0007]
Patent Document 1
Non-Patent Documents
[0008]
Non-Patent Document 1
Non-Patent Document 2
Summary of the Invention
Problems to be Solved by the Invention
[0009] In the present invention, a new configuration of an analog modulator is proposed. Regarding analog modulation, the principle itself has already been studied in the process of devising digital modulation (Non-Patent Document 2). The block diagram thereof is shown in FIG. 14.
[0010] The analog input signal is subtracted from the signal fed back from the output via a 1-bit digital-to-analog converter (1-bit DAC), then integrated and supplied to a comparator where it is compared. The output signal from the comparator is latched by an external clock having a frequency usually 64 times that of the input signal frequency and is regularly shaped, resulting in an output time-averaged over the period of the input signal that matches the input value. When the bit stream at each element in the case of having a constant input approximately at the center of the input range is displayed for this modulator, it is as shown in FIG. 15. In the figure, two wide and strong pulses appear in the pulse train of the 1-Bit DAC (arrow portions in the figure), and it is recognized that signals in the audible band that are irrelevant to the final output remain. This is a major drawback of first-order integrated analog modulation. This can be reduced by adding one more integrator to perform second-order integration, but if analog integration is increased further, it will cause a phase transition and become unstable.
[0011] Therefore, an object of the present invention is to provide an analog modulator that solves the above problems and achieves stable operation and improved output accuracy with a simple structure composed of only a small number of analog elements, and a switching amplifier that enables low-distortion sound using the analog modulator.
Means for Solving the Problems
[0012] To achieve the above objective, the invention described in claim 1 provides an input signal and It is a feedback signal. It consists of a comparator that compares with the integrated signal and a first-order low-pass filter placed in a negative feedback loop, and the first-order low-pass filter is a resistor. and Capacitor only It consists of, The configuration includes feeding the output of the comparator back to the input side of the comparator via the negative feedback loop, The technical means employed is an analog modulator that operates as an integrator, integrating the difference between the input signal and the feedback signal from the output pulse generated by the comparator.
[0013] The invention described in claim 2 employs the technical means of connecting the input signal side to a virtual ground in the analog modulator described in claim 1.
[0014] The invention described in claim 3 employs the technical means of providing a level shift circuit adjacent to the comparator in the analog modulator described in claim 1 or claim 2 for expanding the output amplitude.
[0015] The invention described in claim 4 employs the technical means that the switching amplifier is equipped with an analog modulator described in any one of claims 1 to 3.
[0016] The invention described in claim 5 employs a technical means in which the switching amplifier described in claim 4 is equipped with one set of analog modulators described in any one of claims 1 to 3 and configured in a bridge-tied load connection. [Effects of the Invention]
[0017] The analog modulator of the present invention achieves stable operation and improved output accuracy through a simple structure composed of only a small number of analog elements. In particular, the analog modulator equipped with a virtual ground is suitable for audio applications as it is designed for use with a single power supply. Furthermore, the switching amplifier of the present invention has excellent responsiveness as an audio amplifier, enabling low-distortion sound and high-quality audio output. [Brief explanation of the drawing]
[0018] [Figure 1] This is a schematic diagram of an analog modulator. [Figure 2] This is a circuit diagram of an analog modulator. [Figure 3] This is a circuit diagram of an analog modulator with a virtual ground. [Figure 4] This is a circuit diagram of an analog modulator with a level shift circuit. [Figure 5] This is an explanatory diagram showing a simulation of a waveform from an analog modulator. [Figure 6] This is a schematic diagram showing the cut-out of a pulse waveform. [Figure 7] This is a circuit diagram of a switching amplifier using an analog modulator. [Figure 8] This is the circuit diagram of the prototype high-speed switching amplifier. [Figure 9] This is an explanatory diagram showing the frequency response of a high-speed switching amplifier. [Figure 10] This is an explanatory diagram showing the square wave response (1W output) of a high-speed switching amplifier. [Figure 11] This is an explanatory diagram showing the THD vs. audio frequency (1W output) of a high-speed switching amplifier. [Figure 12] This is an explanatory diagram showing the THD vs. output (frequency 1kHz) of a high-speed switching amplifier. [Figure 13] This is an explanatory diagram showing the THD (1kHz, 1W) vs. OSR of a high-speed switching amplifier. [Figure 14] This is a circuit diagram of a conventional first-order integrating analog delta-sigma modulator. [Figure 15] This is an explanatory diagram showing the bitstream (input=0.03*VRef) for each element. [Modes for carrying out the invention]
[0019] Embodiments of the present invention will be described below with reference to the figures.
[0020] (Modulator) The technical concept of the analog modulator of the present invention is as follows:
[0021] If we remove the external clock and latch in the block diagram of Figure 14 and leave it to self-oscillation, the frequency will be determined by the delay (response delay) of the comparator and integrator. By avoiding the constraint of the external clock, it becomes easier to maintain a high operating frequency, i.e., OSR (OSR > 200). Also, the 1-bit DAC (digital-to-analog converter) in the figure is effectively a level shifter, and when the comparator is driven by a single power supply, the output voltage level and the input level can be matched.
[0022] Figure 1 shows the principle diagram of the analog modulator of the present invention, and Figure 2 shows the circuit diagram of the simplest and most simplified configuration. Analog modulator 1 is a simple analog system consisting only of a comparator 10 and a first-order LPF (low-pass filter) in a negative feedback loop. The configuration is simplified by replacing the digital integrator with a first-order LPF represented by the transfer function H(s)=1 / (1+τs) and the quantization element with a comparator. Here, τ is the time constant of the LPF τ=R f ·C f That is the case.
[0023] In Figure 1, 11 is an integrator using an analog circuit, which is replaced by a first-order low-pass filter in this invention, hence the notation "1st-order LPF" in parentheses. 12 is the signal confluence point, representing the difference between the feedback signal W and the analog input signal X. 13 is a level shift circuit for shifting the output signal level to the input signal level.
[0024] The analog modulator 1 consists of a comparator 10 that compares the input signal with the integrated signal, and a first-order low-pass filter in a negative feedback loop.
[0025] The integrator 11 is connected to a resistor (R f )111 and capacitor (C f This is a first-order LPF consisting of )112. Although this function is an incomplete integral, it is a passive element and can therefore follow even sharp fluctuations in the output pulse.
[0026] The input signal is input to the non-inverting input terminal (+) of comparator 10. The integrator 11 is connected to the inverting input terminal (-). Comparator 10 compares the integral signal C with the input signal X, not with GND (ground).
[0027] When using analog modulator 1 for audio and driving comparator 10 with a single power supply, a configuration can be adopted in which a virtual ground 14 (VG) is connected to the input signal side, as shown in Figure 3. VG is set to be midway between the power supply voltage and GND, thereby defining the center potential of the audio input. The input signal with VG added is input to the non-inverting input terminal (+) of comparator 10. This is equivalent to the level shift circuit in Figure 1.
[0028] The circuit self-oscillates like a relaxed oscillator, and the output of comparator 10 is approximately +V from the power supply. CC + It oscillates between the 0 / 1 pulse and ground (GND). The difference between this 0 / 1 pulse and the input signal is imperfectly integrated by an LPF (low-pass filter) composed of R and C, and supplied to the inverting input pin of comparator 10, where it is compared with the audio signal on the non-inverting input pin. When the feedback signal exceeds the input, comparator 10 inverts and, after a certain response delay, drops in the negative direction. Conversely, it rises in the positive direction. Therefore, although analog modulator 1 belongs to PWM because it generates a series of variable-width pulse trains, it corresponds to a PWM modulator controlled by feedback.
[0029] As shown in Fig. 4, by inserting a level shift circuit 13 on the output side of the comparator 10, the output amplitude can also be enlarged beyond that of the comparator 10. According to this configuration, the emitter common circuit can easily increase the output level. Conversely, by adding a resistor (Re) 15, the feedback signal is leveled down. As a result, the gain A NF =(R e +R f ) / R e will be obtained.
[0030] Here, Fig. 5 shows the result of simulating the output waveform from the circuit of Fig. 2. The output (square wave) from the comparator 10 and the incomplete integration result (triangle wave) are shown. The elapsed time t is converted into dimensionless time X = (t / τ), and the peak heights V p + and V p - are normalized to + / -0.5 in the dimensionless scale. Here, τ is the time constant of the LPF (τ = R f ·C f ). The simulation was performed for an input value Vi = 0.1 and a dimensionless system delay x D =(t D / τ)=1 / 100. Here, t D is the system delay. The upward (high side) pulse is wider than the downward (low side), and its duty cycle is 60%.
[0031] Thus, although the analog modulator 1 appears to conform to delta-sigma modulation on the schematic diagram, in reality, it operates as what is so-called "PWM controlled by feedback". The integration waveform rises and falls like a triangle wave in the normal PWM method, but this is not mechanically generated. Instead, it behaves like that as a result of feedback, and its center is the input value. Therefore, the time average value within each cycle of the output pulse always coincides with the input value and is completed. Since this cycle continues for the number of times of OSR, the output accuracy is high.
[0032] By feeding back the time integral of the output pulse and constantly comparing it to the input value for control, the feedback ensures that the "area" of each pulse above or below the input line on the voltage-time diagram is equal. As a result, the time-averaged output perfectly matches the input. This is why analog delta-sigma modulators using integration can accurately reproduce audio signals and build high-quality analog switching amplifiers.
[0033] Furthermore, even if some kind of disturbance occurs midway through, the feedback immediately corrects the pulse width in the following half-cycle, making it resistant to disturbances such as fluctuations in power supply voltage. Also, since only the same pulse pattern is repeated, irrelevant residual signals, as shown in Figure 15, do not occur.
[0034] Figure 6 shows a single cycle of the waveform from Figure 5, and explains in detail how "the area of each pulse above or below the input line is precisely controlled by feedback to be equal." In the figure, the value V on the vertical axis is conveniently represented by the input value V. i We will use as the base for the transformation, and V=(VV i We will display it in lowercase as ). Therefore, the pulse heights of the high-side and low-side are v p + =V p + -V i , v p - =V p - -V i It is represented as follows.
[0035] Let's consider a complete integral. A pulse wave starting from point [A] reaches point [B] v p - After continuing as is, it reverses direction and the response delay x D After the time has elapsed, the maximum value v is reached at point [D]. p + It reaches this point. Here, the dimensionless system delay x DThe value obtained by integrating over the low-side pulse [AB] during the period is v L If we represent it as a (negative value), it must be canceled out by the integral over the subsequent high-side period [DF], because the integral over the interval [AF] is zero.
[0036]
number
[0037] In the diagram, the area of rectangle [AB] is equal to the area of the corresponding high-side [DF]. Similarly, the area of the other rectangle [FG] is equal to [IK]. Therefore, overall, the area of the low-side pulse below the input line (v L -v H ) is the area of the high-side pulse on the input line immediately preceding it (v H -v L This cancels out the other effect. Needless to say, the converse is also true. From the above, if we take the time average of both pulses, it will be exactly equal to the input value (V=Vi or v=0). This is how a delta-sigma modulator using complete integration can output a signal identical to the input.
[0038] As long as the integrator integrates the pulse intensity on both sides without omission, the pulse waveform, v p + (x) and v p - (x) does not necessarily have to be a rectangle; it can be any shape. Thus, this modulation method suggests that it has high tolerance to any fluctuations in the output pulse. Here, the modulation error (relative value) resulting from replacing the integral with an incomplete integral using an LPF is extremely small, usually less than 1 / 100, and does not fundamentally affect the distortion of the audio waveform.
[0039] The integrator is replaced with a simple first-order low-pass filter (LPF). This consists only of a resistor R and a capacitor C, and its function is called incomplete integration. However, unlike active elements such as op-amps, it is a passive element, which has the advantage of being able to follow even sharp fluctuations in the output pulse. Here, the modulation error (relative value) due to the replaced LPF is constant regardless of the input, and when the ratio of the system delay to the LPF time constant is sufficiently small compared to 1, such as 1 / 100 or less, it depends only on the ratio of the system delay to the LPF time constant.
[0040] Analog modulator 1 contains no digital elements and is composed of only a few analog elements, yet it can obtain audio output with the same high precision as analog modulation. Furthermore, unlike digital delta-sigma modulation, the feedback works more directly, making it more resistant to power supply voltage fluctuations and allowing for a larger damping factor. Within each cycle of the output pulse, the time-averaged value always matches the input value and is completed, and since this cycle continues for the number of OSRs, the output precision is high. Moreover, even if some disturbance is applied midway, the feedback immediately corrects the pulse width in the next half-cycle, making it resistant to disturbances such as power supply voltage fluctuations. In addition, since the pulse waveform is not restricted to equal intervals by an external clock and latch, and only the same pattern of pulse train is repeated, there is no residual irrelevant audible signal.
[0041] (Switch-on / switch-off amplifier) Figure 7 shows the block diagram of the audio switching amplifier 2. Two sets of these circuits are needed to create a standard 2-channel stereo system.
[0042] The switching amplifier 2 includes two analog modulators 1A and 1B connected to speaker 3, a Butterworth filter 17, and an inverting / non-inverting converter 18 connected to them, all connected in a BTL (Bridge-Tied Load) configuration to drive speaker 3. This allows it to be driven by a single power supply.
[0043] Analog modulators 1A and 1B each have a power driver 16. The power driver 16 is a collective term for the output stage and the circuit that drives it. The output stage often uses a complementary emitter follower (CEF) circuit with a pair of power transistors or a complementary common source (MOSFET) circuit with a pair of MOSFETs (Metal Oxide Semiconductor Field Effect Transistors). The Butterworth filter 17 consists of an inductor and a capacitor and is a type of LPF (low-pass filter). This extracts the time-averaged low-frequency components (audio signal) from the high-frequency pulse train.
[0044] The virtual ground 14 is connected to the inverting / non-inverting converter 18. The input audio signal is input to the analog modulator 1A as an inverted signal with inverted phase, and to the analog modulator 1B as a non-inverted signal with uninverted phase, via the inverting / non-inverting converter 18.
[0045] The comparator 10 and power driver 16 (output stage) have a certain response delay and perform self-oscillation at a constant frequency. Therefore, a crystal oscillator clock is not required. Since the switching amplifier 2 is, in principle, merely a unity-gain power amplifier, it is preferable for the inverting-to-non-inverting converter 18, which supplies inverting and non-inverting signals simultaneously, to have a constant voltage gain.
[0046] As the reference potential, the power supply voltage V CC +The audio signal is defined by a virtual ground 14 located between VG and GND. This means the audio signal is input to this VG. The time-averaged output, i.e., the audio output, is typically taken through a low-pass filter called a Butterworth filter. Since the output is generally VG-referenced to the load, another amplifier needs to be connected in a BTL (Bridge-Tied Load) configuration. This BTL connection is a common output method in Class D amplifiers, and it not only enables a single power supply but also doubles the output voltage.
[0047] Switching amplifiers, with the configuration described above, exhibit low THD (Total Harmonic Distortion), enabling them to produce high-quality sound with low distortion.
[0048] [Effects of the Embodiment] The analog modulator 1 of the present invention achieves stable operation and improved output accuracy through a simple structure composed of only a small number of analog elements. In particular, the analog modulator 1 equipped with a virtual ground 14 can be suitably used for audio applications that assume the use of a single power supply. Furthermore, the switching amplifier 2 of the present invention has excellent responsiveness as an audio amplifier, enabling low-distortion sound and obtaining high-quality audio output. [Examples]
[0049] (Frequency response characteristics of a switching amplifier) Using the analog modulator of the present invention, a switching amplifier operating at slightly less than 10MHz was constructed using a MAX913 in a BTL (Bridge Tied Load) configuration, as shown in Figure 8. This amplifier operates on a single +5V power supply and outputs complementary signals for TTL / CMOS logic. In the figure, the other identical amplifier for the non-inverting signal is omitted. Furthermore, a rapid common-emitter circuit is used to level-shift the output amplitude. The feedback signal uses a resistor divider (R e / Rf ) must be shifted down. This high-speed switching amplifier has a constant voltage gain A NF =(R e +R f ) / R e This results in the following: The total gain of this circuit, combining the gains of the converter and amplifier, was approximately 20 dB. The output signal from level shifter Q1 is simply amplified by a two-stage CEF. The idle current was only 0.16 A on the two-channel board.
[0050] Figure 9 shows the frequency response to a sine wave signal. The voltage gain was defined as 0 dB when a 1 W output was applied to an 8.2 Ω resistive load at a frequency of 1 kHz. A gain of >-1 dB was obtained from DC to 100 kHz, but the gain remained almost 0 dB in the practical range from DC to 40 kHz.
[0051] This wide power bandwidth corresponds to the excellent 10kHz square wave response shown in Figure 10. Sharp edges are observed at both the rising and falling edges of the square wave, and no overshoot or undershoot is detected, demonstrating good responsiveness.
[0052] (Distortion characteristics of switching amplifiers) Figure 11 shows the harmonic distortion (THD and THD+N) against audio frequency. The graph shows that the THD at 1W output remains around 0.01% in most frequency ranges. The fact that distortion does not increase even in the high-frequency range explains the excellent square wave response shown in Figure 10.
[0053] Figure 12 shows the change in THD and THD+N against output power. In the low output range of approximately 0.01W to 1W, the THD is well below 0.01%, demonstrating the low distortion characteristics of this unit. These distortion characteristics are comparable to those of conventional analog delta-sigma audio amplifiers.
[0054] We prototyped various switching amplifiers using analog modulators. As shown in Figure 13, the prototype operating at high speed (around 10 kHz) showed excellent square wave response at 10 kHz and 1 W output, and exhibited THD (Total Harmonic Distortion) of less than 0.01% in the output range of approximately 0.01 W to 1 W. Such good audio characteristics confirmed that switching amplifiers using analog modulators are practical. [Explanation of symbols]
[0055] 1 (1A, 1B) ... Analog modulator 10... Comparator 11...Integrator 111... Resistance 112... Capacitor 13... Level shift circuit 14…Virtual Ground 15… Resistance 16…Power Driver 17…Butterworth filter 18…Inverting / Non-inverting converter 2…Switching amplifier 3…Speaker
Claims
1. It consists of a comparator that compares the input signal with the integrated signal which is the feedback signal, and a first-order low-pass filter placed within a negative feedback loop. The first-order low-pass filter consists only of a resistor and a capacitor, and is configured to feed back the output of the comparator to the input side of the comparator via the negative feedback loop. An analog modulator characterized by operating as an integrator that integrates the difference between an input signal and a feedback signal from an output pulse generated by the comparator.
2. The analog modulator according to claim 1, characterized in that the input signal side is connected to a virtual ground.
3. The analog modulator according to claim 1 or 2, characterized by comprising a level shift circuit for increasing the output amplitude.
4. A switching amplifier comprising an analog modulator according to any one of claims 1 to 3.
5. The switching amplifier according to claim 4, characterized in that it comprises one set of analog modulators according to any one of claims 1 to 3, and is configured in a bridge-tied load connection.