Class D audio amplifiers in mobile applications
Active In SP
Joined: Feb 2011
14-02-2011, 04:28 PM
classdsem.docx (Size: 1.18 MB / Downloads: 77)
A circuit that increases the amplitude of a given input signal is called an amplifier. A small a.c. signal fed to the amplifier is obtained as a larger a.c. signal of the same frequency at the output.Amplifiers constitute an essential part of radio, television and other communication circuits. Bipolar junction transistors and Field effect transistors are commonly used as amplifying elements.
Amplifiers can be classified as follows:
1. Based on transistor configuration:
(a) Common emitter amplifier
(b) Common collector amplifier
© Common base amplifier
2. Based on active devices:
(a) BJT amplifier
(b) FET amplifier
3. Based on Q-point:
(a) Class A amplifier
(b) Class B amplifier
© Class AB amplifier
(d) Class D amplifier
4. Based on number of stages:
(a) Single stage amplifier
(b) Multi stage amplifier
5. Based on output:
(a) Voltage amplifier
(b) Power amplifier
6. Based on frequency response:
(a) AF amplifier
(b) IF amplifier
© RF amplifier
7. Based on bandwidth:
(a) Narrow band amplifier
(b) Wide band amplifier
The commonly accepted audio amplifier classes are
Where efficiency is not a consideration, most small signal linear amplifiers are designed as Class A, which means that the output devices are always in the conduction region. Class A amplifiers are typically more linear and less complex than other types, but are very inefficient. This type of amplifier is most commonly used in small-signal stages or for low-power applications (such as driving headphones).
In Class B, there are two output devices (or sets of output devices), each of which conducts alternately for exactly 180 deg (or half cycle) of the input signal. Amplifiers are commonly classified by the conduction angle (angle of flow) of the input signal through the amplifying device.
Popular for high power RF amplifiers, Class C is defined by conduction for less than 180° of the input signal. Linearity is not good, but this is of no significance for single frequency power amplifiers. The signal is restored to near sinusoidal shape by a tuned circuit, and efficiency is much higher than A, AB, or B classes of amplification.
Class AB amplifiers are a compromise between Class A and B, which improves small signal output linearity; conduction angles vary from 180 degrees upwards, selected by the amplifier designer. Usually found in low frequency amplifiers (such as audio and hi-fi) owing to their relatively high efficiency, or other designs where both linearity and efficiency are important (cell phones, cell towers, TV transmitters).
A Class D audio amplifier is basically a switching amplifier or PWM amplifier. There are a number of different classes of amplifiers. This application note takes a look at the definitions for the main classifications
A Class D amplifier or switching amplifier is an electronic amplifier where all power devices (usually MOSFETs) are operated as binary switches. They are either fully on or fully off. Ideally, zero time is spent transitioning between those two states.
Output stages such as those used in pulse generators are examples of class D amplifiers. However, the term mostly applies to power amplifiers intended to reproduce signals with a bandwidth well below the switching frequency.
Class D amplifiers work by generating a square wave of which the low- frequency portion of the spectrum is essentially the wanted output signal, and
of which the high-frequency portion serves no purpose other than to make the wave-form binary so it can be amplified by switching the power devices.
A passive low-pass filter removes the unwanted high-frequency components, i.e. smoothes the pulses out and recovers the desired low-frequency signal. To maintain high efficiency, the filter is made with purely reactive components (inductors and capacitors), which store the excess energy until it is needed instead of converting some of it into heat. The switching frequency is typically chosen to be ten or more times the highest frequency of interest in the input signal. This eases the requirements placed on the output filter.
The structure of a class D power stage is essentially identical to that of a synchronously rectified buck converter, a type of non-isolated switched-mode power supply. Whereas buck converters usually function as voltage regulators, delivering a constant DC voltage into a variable load and can only source current (one-quadrant operation), a class D amplifier delivers a constantly changing voltage into a fixed load, where current and voltage can independently change sign (four-quadrant operation). A switching amplifier must not be confused with any amplifier that uses an SMPS. A switching amplifier may use any type of power supply but the amplification process itself operates by switching.
Theoretical power efficiency of class D amplifiers is 100%. That is to say, all of the power supplied to it is delivered to the load, none is turned to heat. This is because a switch in its on state will conduct all current but has no voltage across it, hence no heat is dissipated. And when it is off, it will have the full supply voltage standing across it, but no current flows through it. Again, no heat is dissipated. Real-life power MOSFETs are not ideal switches, but practical efficiencies well over 90% are common. By contrast, linear AB-class amplifiers are always operated with both current flowing through and voltage
standing across the power devices. An ideal class B amplifier has a theoretical maximum efficiency of 78%.
The binary waveform is derived using pulse-width modulation (PWM), pulse density modulation (sometimes referred to as pulse frequency modulation), sliding mode control (more commonly called "self-oscillating modulation" in the trade) or discrete-time forms of modulation such as delta-sigma modulation.
The most basic way of creating the PWM signal is to use a high speed comparator ("C" in the block-diagram above) that compares a high frequency triangular wave with the audio input. This generates a series of pulses of which the duty cycle is directly proportional with the instantaneous value of the audio signal. The comparator then drives a MOS gate driver which in turn drives a pair of high-power switches (usually MOSFETs). This produces an amplified replica of the comparator's PWM signal. The output filter removes the high-frequency switching components of the PWM signal and recovers the audio information that the speaker can use.
Another way to create the PWM signal is adopted when a SPDIF signal or other form of digital feed is available. The digital signal is fed to a DSP that uses software to create the PWM signal. This drives the MOSFETs through a suitable gate driver chip.
Two significant design challenges for MOSFET driver circuits in class-D amplifiers are keeping dead times and linear mode operation as short as possible. "Dead time" is the period during a switching transition when both output MOSFETs are driven into Cut-Off Mode and both are "off". Dead times need to be as short as possible to maintain an accurate low-distortion output signal, but dead times that are too short cause the MOSFET that is switching on to start conducting before the MOSFET that is switching off has stopped conducting. The MOSFETs effectively short the output power supply through themselves, a condition known as "shoot-through". Meanwhile, the MOSFET drivers also need to drive the MOSFETs between switching states as fast as possible to minimize the amount of time a MOSFET is in Linear Mode, the state between Cut-Off Mode and Saturation Mode where the MOSFET is neither fully on nor fully off and conducts current with a significant resistance, creating significant heat. Driver failures that allow shoot-through and/or too much linear mode operation result in excessive losses and sometimes catastrophic failure of the MOSFETs.
The actual output of the amplifier is not just dependent on the content of the modulated PWM signal. The power supply voltage directly amplitude-modulates the output voltage, dead time errors make the output impedance non-linear and the output filter has a strongly load-dependent frequency response. An effective way to combat errors, regardless of their source, is negative feedback. A feedback loop including the output stage can be made using a simple integrator. To include the output filter, a PID controller is used, sometimes with additional integrating terms. The need to feed the actual output signal back into the modulator makes the direct generation of PWM from an SPDIF source unattractive.
Despite the complexity involved, a properly designed class-D amplifier offers the following benefits:
• Reduction in size and weight of the amplifier,
• Reduced power waste as heat dissipation and hence smaller (or no) heat sinks,
• Reduction in cost due to smaller heat sink and compact circuitry,
• Very high power conversion efficiency, usually ≥ 90%.
Boss Audio mono amp. The output stage is top left; the output chokes are the two yellow toroids underneath.
• Home Theatre systems: In particular the economical "home theatre in a box" systems are almost universally equipped with class D amplifiers. On account of modest performance requirements and straightforward design, direct conversion from digital audio to PWM without feedback is most common.
• Mobile phones: The internal loudspeaker is driven by up to 1 watt. Class D is used to preserve battery lifetime.
• Powered speakers
• High-end audio: Is generally conservative with regards to adopting new technologies but class D amplifiers have made an appearance.
• Active subwoofers
• Sound Reinforcement and Live Sound: The weight reduction makes class D amplifiers more transportable. The Crest Audio CD3000, for example, is a class D power amplifier that is rated at 1500 watts per channel, yet it weighs only 21 kg (46 lb).
• Bass amplifiers: Again, an area where portability is important. Example: Yamaha BBT500H bass amplifier which is rated at 500 watts, and yet it weighs less than 5 kg (11 lb).The Promethean P500H by Ibanez is also capable of delivering 500W intoa4Ohmload,andweighsonly2.9Kg.
HIGH-POWER class-D amplifiers have become standard in many consumer electronic applications, such as television sets and home theater systems. Currently, class-D is also making a cautious entrance into the automotive domain. The first integrated class-D audio amplifiers designed to directly operate from the car battery are now entering the market. A third domain where class-D is emerging is in mobile applications, such as cellular phones, portable navigation systems, and portable gaming devices. In these applications, the output power is low, e.g., about 1 W. Reduced heat production and extension of battery life are the drivers behind class-D. Class-D amplifiers are realized as standalone products and as part of audio codec’s that contain multiple analog microphone and line inputs, and analog speaker, headphone, and line outputs.Currently, audio codec’s containing class-D can be found in large mixed-signal silicon-on-insulators where Global System for Mobile Communications (GSM) analog baseband and audio interfaces are combined using advanced nanometer CMOS technologies. In these highly integrated systems, reduced heat production of a 1-W audio amplifier is of paramount importance, which is a good reason to use class-D. In addition, audio codec’s are integrated together with power management units (PMUs) that can be found not only in cellular phones but also in many other portable applications, such as personal digital assistants and MP3 players. Such PMU chips are often made in dedicated CMOS technologies with high voltage capabilities that facilitate the realization of switch-mode dc/dc converters\ that are akin to class-D amplifiers. These technologies are generally better suited for the implementation of class-D amplifiers.
Output power for speaker drivers in mobile application ranges from 500 mW to 1W, with a trend toward higher power. The output stage is preferably directly connected to the battery. The battery voltage ranges between 2.5 and 4.5 V, and can be as high as 5.5 V during charging. A high (80 dB) power supply rejection is essential, particularly in cell phones, because the battery voltage is polluted by the characteristic 217 Hz interval of the transmit RF power amplifier. The most important feature of class-D amplifiers is high efficiency, which is typically higher than 90% at full output power. This efficiency allows very high output power with modest heat sinking. Output powers well over 100Wper channel are no exception. The efficiency is largely determined by two factors:
1) Switch impedance and
2) (Average) switching frequency.
Finite switch impedance causes conductive losses that are proportional to the square of the output current and is usually the limiting factor for efficiency at full power. At low output power, switching losses are the dominant factor. For each output transition, the input capacitance of the switch devices needs to be charged (or discharged), causing charging losses those are proportional to input capacitance and switching frequency. In addition, finite switching speeds cause energy loss during each output transition. Switch impedance is inversely proportional to the size of the switch device, whereas input capacitance linearly scales with size. For high efficiency, the switching frequency must be as low as possible, whereas the switch device size should be optimized to trade off conductive and charging losses.Usually, some form of pulse width modulation (PWM) is used to encode the audio signal that is subsequently retrieved by means of an external LC low-pass filter connected between the class-D output stage and the load. The simplest form of PWM is the so-called natural sampling PWM (NPWM). An NPWM signal can easily be constructed by comparing the audio signal to a triangular reference VREF. The fundamental frequency of the triangular reference is usually much higher than the highest audio frequency, e.g., about 350 kHz or 8 44.1 kHz, and is called the carrier frequency. The modulation depth M is defined between +1 and -1 and is related to the duty cycle D of the PWM signal as M=2.D-1
Although the generation of NPWM involves a highly nonlinear comparator, the frequency spectrum of an ideal NPWM signal does not contain harmonics of the input signal but only intermodulation products of the carrier and the input signal, i.e., NPWM is free from harmonic distortion. The ideal frequency spectrum of an NPWM signal at near-maximum modulation is shown in the above figure. Assuming that the triangular reference has a sufficiently high frequency, the intermodulation products do not fold back to the audio frequency band and are filtered out by the LC low-pass filter. In a practical implementation, it is not possible to reproduce the PWM pulses at the output with mathematical precision, because the switching output stage introduces timing and amplitude errors that result in distortion. Integrated class-D amplifiers typically have distortion better than 70 dB and a dynamic range of 100 dB.A class-D output stage can be either single-ended (SE) or differential, yielding the so-called bridge-tied load (BTL) configuration, as shown in the above figure. In a BTL amplifier, both sides of the loudspeaker load are driven in opposite (audio) phase. This enables operation from a single supply while doubling the voltage swing across the load, yielding four times more output power than an SE amplifier. On the downside, a BTL amplifier needs twice the number of power switches and inductors, making it relatively expensive. In a BTL class-D amplifier, the phase of the carriers of both bridge halves can independently be chosen. When the carriers are in opposite phase, as shown in fig4 (a) this is called AD modulation or binary modulation. The main advantage of AD modulation is that the output signal has zero common modes since the bridge halves always switch simultaneously in opposite directions. Conversely, when the carriers are in phase, as shown in Fig. 4(b), this is called BD modulation or ternary modulation. Using BD modulation opens the possibility of filter less application, which is a must have for mobile class-D amplifiers. In BD modulation, the amount of differential-mode high-frequency energy is reduced, compared with AD modulation. In this case, the speaker itself can act as a filtering element, provided that the speaker is close to the amplifier output, which is usually the case in mobile applications.
CLASS-D AMPLIFIER ARCHITECTURE:
Many Class-D amplifier architectures exist. A coarse division can be made in open-loop and closed-loop architectures.
A. Open-Loop Architectures:
In open-loop class-D amplifiers, a one-bit signal is generated in the digital domain and directly drives a class-D output stage. In this architecture, the class-D output stage itself
Serves as D/A converter.
In principle, the one-bit signal can straightforwardly be generated with a sigma-delta modulator (SDM). A drawback of SDM bit streams is that the average switching frequency depends on the modulating signal and can be quite high, which results in poor efficiency. A typical SDM produces switching frequencies of as high as 1.2 MHz at low modulation depths. A digitally generated PWM signal on the other hand has a constant and relatively low switching frequency but requires a 4- to 16-times higher sample rate for comparable signal-to-noise ratio (SNR). Hybrid techniques can be used to reduce the switching frequency of SDM signals without increasing the sample rate. However, in modern CMOS technologies, the sample rates required for straightforward digital PWM generation are easily achieved. Because the PWM signal is generated in the digital domain, the effects of sampling and quantization have to be dealt with. The effect of sampling can be modeled by inserting a sample and- hold in series with the modulating signal, as shown in Fig. 6. As can be seen, the sampled signal crosses the reference triangle at different times than the original modulating signal, causing the edges of the PWM signal to shift.
The spectrum now fills up with harmonics of the modulating signal. Apparently, sampling the input signal causes distortion. In a digital PWM modulator, the edges are synchronized to a high-frequency bit clock, e.g., 256 fs . Consequently, the pulse widths are quantized to a limited number of discrete values. This can be modeled by inserting a sample-and-hold in series with the reference triangle operating at a multiple of the PWM frequency
Sampling and quantization effects can be separately or integrally handled. In a separated approach, as shown in Fig. 10(a), first, the distortion caused by sampling is corrected by either approximating NPWM using linear or higher order interpolation or applying pre correction based on a digital PWM distortion model. In the second stage, the quantization noise is shaped out-of-band ∑∆ by a modulator before converting to digital PWM.
In an integral approach, the digital PWM generator is used as quantizer ∑∆ in a loop yielding a PWM- ∑∆ loop, as shown in Fig. 10(b). A disadvantage of this approach is that the entire loop needs to run at the high bit-clock frequency. Essentially, with both approaches, any required SNR and total harmonic distortion (THD) can be achieved at the expense of higher clock rates as long as the signal remains in the digital domain.
Even without correction of output stage errors, open-loop class-D amplifiers can have very good performance, provided that an accurate regulated power supply is used. The main problem with open-loop class-D amplifiers is the lack of supply rejection. This problem is concealed in the BTL configuration shown earlier, because then supply variations largely cancel at small signal levels. Fortunately, BTL, particularly with BD modulation, is preferred for mobile, because it enables filter less application. Another advantage of BD modulation is that even harmonics are canceled, which means correcting for sampling distortion is not necessary.Some attempts have been made to improve power supply rejection by sensing the supply voltage with an analog-to digital converter and using feed forward correction of the digital PWM signal, but these designs only yield a moderate improvement in power supply rejection. Open-loop class-D amplifiers can be found mainly in consumer electronics, such as television sets and home theater systems, but the lack of supply rejection makes them unsuitable for mobile applications.
B. Closed-Loop Architectures:
An evident way to improve power supply rejection is to apply feedback. Since the output signal of a class-D amplifier is essentially analog, most feedback class-D amplifiers require an analog input signal. Calculation of loop gain in class-D feedback loops can be done in a straight forward manner by adding a static disturbance VN after the switching output stage of the loop, A static disturbance works fine for the delay-based self-oscillating, fixed-carrier, and direct PWM loops. For the hysteretic loop, however, the calculation is slightly more complicated. In the hysteretic loop, the steady-state average of the integrator output is always zero if the slopes of VX are assumed linear. This would yield infinite loop gain, which is obviously incorrect. A more accurate result is obtained when a dynamic time-varying disturbance, instead of a static disturbance, is used. By using a linear ramp as disturbance VN , the waveforms at the integrator output VX now become parabolic, as shown (exaggerated) in Fig. 14(b), yielding a nonzero steady-state average value. Apparently, a constant average at the input of the hysteresis yields a ramp at the output. Consequently, the hysteresis behaves as an integrator and gives the hysteresis loop a second-order loop transfer. Large signal stability of direct PWM loop requires the closed loop GCL gain to be higher than unity to guarantee that the PWM input signal always dominates the feedback by making the amplitude of the PWM input current larger than that of the feedback current. This causes the modulation depth M of the PWM input signal to be multiplied by a factor GCL at the output. Consequently, the signal level and, in case the source is a digital PWM modulator, the quantization noise are amplified with the closed-loop gain GCL. In this way, the modulation depth of the PWM source does not necessarily have to be 100% in order to drive the amplifier output into clipping, relaxing the design of the noise shaper in the digital PWM source.The hysteretic loop has a second-order loop transfer yielding huge magnitudes in the audio range. Unfortunately, the loop transfer of hysteretic loops reduces to first order at low frequencies when a delay is introduced in the loop. The loop transfer of a hysteretic loop with delay cannot exceed the loop transfer of a delay-based loop with that same delay. Still, for modest (realistic) delays, the loop transfer of hysteretic loops is always higher than that for any other feedback topology, particularly at higher audio frequencies.
The main drawback of self-oscillating class-D amplifiers is the variable switching frequency. In multichannel systems, differences in switching frequency can cause audible intermodulation products known as beat tones. In addition, when the switching frequencies of two channels are nearly equal, they tend to lock onto each other, deteriorating the performance.The variable switching frequency of self-oscillating class-D amplifiers is sometimes presented as an electromagnetic force (EMI) advantage since it distributes energy over a range of frequencies. Others prefer the predictability of fixed-carrier class-D amplifiers. The best choice usually depends on the application.Power supply rejection is proportional to loop gain. In this respect, the hysteretic loop is clearly the best. However, because the unity gain frequency of fixed-carrier and direct PWM loops is fixed, it is rather straightforward to upgrade the feedback loop to the second (or higher) order by adding integrators. With regard to efficiency, there is not much difference between amplifier architectures. Efficiency is mainly determined by conduction losses and, to a lesser extent, by the switching frequency. The latter contribution would favor self-oscillating amplifiers that have lower switching frequency at large modulation depths, but this is only a small advantage.Distortion in class-D amplifiers is, in practice, more determined by quality of implementation than choice of architecture. In general, it can be expected that higher loop gain will yield lower distortion, but all class-D architectures discussed are capable of producing THD figures of 70 dB or better, which is good enough for mobile applications.
Development of Class D Audio IC:
Nowadays, most of portable electronics, home audio/video device and the more and more car audio systems all start using class-D audio amplifier. The market of Class-D audio amplifier is growing at a rate of over 50 percent. National Semiconductor LM4670 has super power, with the lowest price, highest price performance ratio, minimum package area and good timbre. MAXIM MAX9700 has good timbre. Its S/R reaches 90 db with 94% efficiency. It is mainly applied to music mobile phones, with advantages of reasonable price and small package area. However, its output power is not high. The internal frame of MAXIM MAX9700 adopts fixed frequency FEM mode and spread spectrum mode, which can not only reduce distortion, but also reduce electromagnetic interference (EMI) ONS NCP2820 has good timbre and high output power, which is mainly applied to music mobile phone and hand free phone. As its package area is small, it is also applied to folding handset. But its price is on the high side. LG mobile phones (LG-KV1300, LG-SV130, LG-SV140 and LG-SV9140) have already adopted TI TPA2005D1 chip. TI TPA2005D1 has good timbre and high price performance ration. However, it has some disadvantages, such as low output power and large package area.
Demand for features dominates the mobile design landscape, but energy management has ultimate power of veto. This fact has been the major influence in drawing class D amplification into the mobile arena; designers are under pressure to deliver greater multimedia capability for fewer Watts. It is an effective solution, but optimal implementation demands good audio and mixed signal design skills, to create an integrated solution that will meet small footprint, low component count and low leakage demands implicit in any mobile system requirement.
This new product is targeted specifically for mobile phones and wireless PDAs where size, low standby current, and power efficiency benefits are critical. A major goal is to design the lowest cost Class-D audio IC with on-board output MOSFETs for use in these very high volume markets.For the ultimate in size and power conservation, Microsemi has created an ultra-low power Class-D amplifier with integrated output MOSFETs for hearing aid applications.
According to feedback theory, the disturbance appears at the output attenuated by the loop gain plus unity. The integrator yields a very high loop gain at low frequencies, so the disturbance VN is attenuated to practically zero. Consequently, the steady-state average of the PWM signal produced by the switching output stage is almost the exact opposite of the disturbance VN. This can be translated to a duty cycle for the PWM signal that exactly compensates the disturbance. With this duty cycle, the steady-state signal at the output of the integrator VX can now easily be constructed using simple geometry A static disturbance works fine for the delay-based self-oscillating, fixed-carrier, and direct PWM loops.