Application and advantages of GaN devices in Class D audio amplifiers
[Copy link]
This document introduces the typical design of a Class D audio amplifier, summarizes the basic applications of GaN devices in Class D audio amplifiers, and briefly introduces the advantages of GaN devices over silicon-based devices in Class D audio amplifier design.
A typical design of a Class D audio amplifier
1. What is a Class D audio power amplifier?
The Class D amplifier was first invented by British scientist Alec Reeves in 1950. In simple terms, a Class D power amplifier is an electronic amplifier, also known as a power switching amplifier, that works on pulse width modulation, which converts the input signal into a stream of pulses. The output transistor stage of a Class D power amplifier operates as an electronic switch and does not have a linear gain like other amplifiers. The Class D power amplifier starts working by taking the incoming analog input signal and generating PWM or PDM. It then converts the input signal into a stream of pulses. Therefore, it can be said that a typical Class D amplifier consists of two output MOSFETs, a pulse width modulator, and an external low-pass filter to recover the amplified audio.
Figure 1. Comparison of efficiency and distortion performance between different types of audio power amplifiers.
Unlike the linear power regulation used in Class AB power amplifiers, which results in energy losses in the power transistors, Class D amplifiers use switching transistors that only operate in two phases, "on" or "off". There is almost no energy loss in the transistors, and almost all the power is transferred to the transducer. Therefore, compared with Class A, Class B and Class AB amplifiers, D audio amplifiers can have efficiencies as high as 90-95%, while AB amplifiers have a maximum efficiency of only 60-65%.
2. Working principle of Class D audio amplifier
The Class D amplifier generates a series of fixed-amplitude rectangular pulses at the beginning of operation, which vary in area and spacing, or number per unit time. In addition, the amplitude changes of the analog audio input stream are also represented by these pulses, and it is also possible to synchronize the modulator clock with the incoming digital audio input signal, thus eliminating the need to convert the digital audio signal to analog. The output stage of the modulator controls their operation by alternately turning the output transistors on and off.
Figure 2. Schematic diagram of pulse width modulation waveform
Since transistors are either fully "on" or fully "off", the time they spend in the linear region is very short and during this time they consume very little power, which is a major factor in their high efficiency.
Advantages of GaN Devices in Class D Audio Amplifier Applications
To put it bluntly, the advantages of GaN switching devices over silicon-based transistors in Class D audio amplifiers are mainly the following three points:
- Higher overall efficiency
- Distortion index has been improved
- Switching waveform is clearer
So how do GaN switching devices bring the above three advantages to Class D audio amplifiers?
1. Higher overall efficiency
First of all, from the perspective of conduction loss, in order to achieve excellent performance of Class D audio amplifiers, it is necessary to provide the lowest possible on-resistance to minimize conduction losses. GaN switching devices provide much lower on-resistance than silicon-based transistors and achieve this on a smaller die area.
Secondly, switching loss is another important factor to consider. Class D amplifiers are extremely efficient at medium and high power output. However, when it is in a low power output state, due to the loss in the power devices, the efficiency is much lower than that at medium and high power output.
To overcome this challenge, some Class D amplifiers use two operating modes. Once the output power level reaches a predetermined threshold, the output voltage rail of the amplifier switch tube will increase, providing a full-scale voltage swing. Therefore, in order to further reduce the impact of switching losses, zero voltage switching (ZVS) technology can be used at low output power levels, and hard switching can be used at high power levels. The extremely low switching losses in the zero voltage switching (ZVS) state of GaN devices can be used to improve the overall system efficiency at low output power.
Figure 3. Efficiency curve comparison of GaN and Si switching devices in Class D audio amplifier applications (load 8 ohms)
From the above figure, we can see that GaN switching devices can provide a 3%-6% efficiency improvement in Class D audio amplifier applications compared to Si switching devices. The efficiency difference is particularly obvious when the output power is between 20W and 80W.
2. Distortion index has been improved
When a Class D audio amplifier is operated in ZVS mode, switching losses are effectively eliminated because the output transition is achieved by commutating the inductor current. However, like all other half-bridge designs, we need to consider the problem of shoot-through, which is the moment when the high-side and low-side switches are turned on at the same time. We usually insert a short delay called blanking time to ensure that one of the switches is completely turned off before the other switch turns on.
It should be noted that this delay will affect the PWM signal and cause audio output distortion, so the goal is to make it as short as possible to maintain audio fidelity. The length of this delay depends mainly on the output capacitance Coss of the power device. Although GaN transistors have not completely eliminated Coss, it is significantly lower than silicon-based switching devices. Therefore, a shorter blanking time can make Class D audio amplifiers less distorted when using GaN as a switching device. In the field of professional speakers, a subtle THD gap can bring consumers a completely different auditory experience.
Figure 4. Switching waveform comparison: GaN FET waveform (left) and Si FET waveform (right)
3. The switching waveform is clearer
As with any audio amplifier, an important indicator of any Class D audio amplifier's performance is the degree of reproduction of the audio signal. For a "switching amplifier" system, such as a Class D audio amplifier, one of the main goals is to use a "perfect" switching waveform. The closer the switching waveform is to "perfection", the closer the audio reproduction effect is to "perfection".
When silicon-based transistors are used to implement switching functions in Class D audio amplifiers, hard-switching mode causes charge to accumulate in the body diode because the voltage at the output is not zero when the power device is turned off and on, and the reverse recovery charge (Qrr) that is built up needs to be discharged, and the discharge time needs to be taken into account in the PWM control action. In the design using GaN, this is no longer a problem we need to consider, because GaN transistors do not have the body diode inherent in silicon-based transistors, so there is no reverse recovery charge Qrr, allowing us to get a clearer switching waveform.
Figure 5. Negative effects of charge-related parameters on Si FET and GaN FET
From the above figure, we can clearly see that reverse recovery charge (Qrr) as well as Cgs and Cgd have a serious negative impact on the restoration of the switching waveform of silicon-based switching devices. However, thanks to the absence of reverse recovery charge (Qrr) and very low Cgs and Cgd in GaN switching devices, the negative impact of such charge parameters is very limited.
Conclusion
Silicon-based switching devices have served designers of Class D audio amplifiers well for many years, thanks to continuous progress in optimizing their performance. However, further improvements in their characteristics are challenging. In addition, further reductions in on-resistance RDS(on) lead to larger chip sizes, making it more difficult to build compact audio amplifier designs. However, GaN switching devices break through this limitation, while also eliminating Qrr, which, combined with their reduced Coss and ability to operate at higher switching frequencies, means that smaller and more compact designs can be easily built. The resulting THD+N measurements also show that this new technology can achieve excellent audio performance.
Remy Zhang
|