Designing an LLC Resonant Converter for Audio Applications with High Peak Loads

For designers, power-conversion design in the audio field is a real technical challenge, as peak loads could be much higher than the root-mean-square (RMS) power requirement. They must strike an optimal balance between thermal requirements, size and weight of the solution, cost and, of course, efficiency.

High-power audio

Nowadays, the LLC resonant converter is a fairly common option for high-power audio applications with high peak loads. This type of converter offers a number of advantages, including higher efficiency and reliability than other solutions. It is a type of power converter composed of three reactive elements, that exploits the resonance between a coil and a capacitor and that implements efficient power conversion by oscillating at a specific frequency, called resonant frequency.

To design an LLC resonant converter for audio applications, there are some important factors to consider. First, the resonant frequency of the circuit must be chosen carefully, based on the specifications of the application. The inductor and capacitor must be selected according to the power of the application and the desired resonant frequency and must be of the highest quality. The coil must have a sufficiently high inductance to avoid the core saturation effect during peak load current, while the capacitor must be able to handle the voltage and current of the circuit. Consider the audio track in Figure 1, in which several elements of a soundtrack are highlighted. Among them are the following elements:

RMS audio level: The RMS audio level indicates the effective amplitude of an audio signal in the time domain. It is often used to represent the true level of an audio signal because it takes into account both the positive and negative amplitudes of the signal. The RMS audio level is often expressed in decibels relative to a power reference. This level is important because it represents the actual power of an audio signal.
Peak audio level: The peak audio level represents the maximum amplitude value of an audio signal in the time domain. It is the maximum instantaneous audio power and indicates the highest peak point reached by the signal during a given period of time. Peak level is usually measured in decibels relative to an amplitude reference. The peak value is an indicator of the maximum amplitude of the signal and can be useful in avoiding distortion or clipping of the audio signal. Measuring the peak audio level is important to ensure that the signal does not exceed acceptable amplitude limits and to avoid unwanted distortion or damage to speakers or playback devices.
Crest factor: The crest factor in audio is a measurement that indicates the gap between the peak level and the RMS level of an audio signal. It is calculated by dividing the peak value of the audio signal by its RMS value. For example, if the peak value of a signal is 2 V and its RMS value is 1 V, the crest factor is equal to 2. The crest factor is important because it provides information about the dynamics of the audio signal. A high crest factor indicates a large difference between peak and RMS levels, indicating that the signal has greater dynamics and may contain high-amplitude transients or peaks. In contrast, a low crest factor indicates lower dynamics and a more compressed or limited signal. A high crest factor requires a different approach than a low crest factor to manage signal dynamics and ensure that the signal itself remains within acceptable limits without unwanted distortion or clipping.

The relationship between the <a href= — Figure 1: The relationship between the peak audio level and the RMS audio level

A basic requirement for power supplies is that they must be able to momentarily support a peak load several times the rated load. Although such an occurrence might happen infrequently, it is imperative that the power supply system provides for this eventuality. Some general measures of crest factor on different types of sound are listed below:

Ambient noise: 3:1
Speech: 4:1
Music with peak level compression: 4:1 to 8:1
Music without peak level compression: 8:1 to 10:1
Movie audio: >10:1

If the output voltage of the power supply drops excessively during the peak load transient, clipping occurs (see oscillogram in Figure 2) and much sound information is cut off, resulting in total distortion of the audio. Moreover, undesirable effects involve not only the sounds but the electronic part of the system, particularly:

Subwoofers are in great danger because of their mechanical limitations. If the oscillations of the speakers exceed those expected, both the cone and the coil can be damaged.
The subwoofers or tweeters themselves may overheat due to the high energy passing through the audio line.
Sound reproduction is highly distorted, as the audio signal at the upper and lower peaks is cut off.

The basic requirement for the electrical circuit is to have a powerful power supply that must keep its output voltage within a strict limit of overshoot and undershoot. In the past, tube systems rarely distorted audio signals, either because of the high voltage involved or because of the power supplies, albeit with very low efficiency (in essence, a lot of unused heat was produced).

Figure 2: Insufficient power supply could cause clipping.

Some solutions regarding optimal audio power supply frequently used today involve adopting a larger percentage of the occupied pcb area or reducing the size of the circuits by using higher switching frequencies and smaller magnetic elements.

LLC up close

As mentioned earlier, the implementation of an LLC solution in audio requires complete collaboration and communication between power and audio engineers. Before designing power supplies, it is necessary to specify the continuous power and peak power that the audio amplifier will be able to handle. The ratio of peak power to continuous power (as seen in the table above) depends on the specific application, so it needs to be clearly defined at the beginning of the design.

The LLC series resonant converter (LLC-SRC) is a specific configuration of an LLC resonant converter and is used as an isolated DC/DC converter. In the LLC-SRC configuration, the LLC converter (see Figure 3) is connected in series with a transformer and a rectifier diode to achieve galvanic isolation between the input and output of the converter. The transformer allows the input voltage to be increased or decreased according to the needs of the application.

This configuration has several advantages. First, it offers high efficiency through the use of resonance between the coil and capacitor, which reduces switching losses. In addition, the galvanic isolation provided by the transformer allows for the input and output of the converter to be separated, protecting the control circuits and allowing adjustment of the transformer ratio. As can be seen from the graph of gain response curves, depending on the switching frequency, there are three different cases:

If the switching frequency is less than the resonant frequency:
- DCM operation on the secondary
- Soft-switching rectifier (ZCS)
- Higher RMS currents for a given power
If the switching frequency is equal to the resonance frequency:
- DCM/CCM operating limit on the secondary
- Soft-switching rectifier (ZCS)
- Optimal efficiency point
If the switching frequency is greater than the resonance frequency:
- CCM operation on the secondary
- Reverse recovery of the rectifier
- Low RMS currents for a given power

The ideal goal, therefore, is to operate close to the resonant frequency at nominal load and below resonance during peak load. Once the design specifications have been defined, the design of the power supply can proceed. Depending on the power quality standards of the region and application, a power supply with power-factor correction will likely be required for power designs. The first key design step is to select the resonant circuit components to set the resonant frequency and characterize the gain. At this stage, it is essential that the output voltage be sufficient for the system to operate during the peak power level. If the circuit is unable to achieve the required gain, the output voltage will decrease during the power peaks, reducing the audio quality or disabling the amplifier. The duration of peak power may be quite long, so the power supply must be able to continuously supply the entire peak load.

Principle diagram of <a href= — Figure 3: Principle diagram of LLC resonant converter

The critical phase of operation is precisely during peak power signals, so it is important to use first-rate components that can handle such currents, and magnets must not saturate. During continuous powers, on the other hand, components must ensure their nominal thermal performance. It is often convenient to design a pcb suitable for thermal dissipation rather than a heatsink. LLC-SRCs are often designed in burst mode to ensure efficiency under light loads and meet industry standards during standby power. This reduces standby power consumption without turning off the main output.

Conclusion

The design of resonant converters is quite challenging, especially if they are used with audio systems. They are an effective solution for high-power music applications with high peak loads. Designing an LLC resonant converter requires the judicious choice of resonant frequency, proper selection of the coil and capacitor according to the application specifications and design of phase control to handle peak loads. Proper design of the LLC resonant converter ensures efficient and reliable operation of the application, even if it is very complex. The efficiency of the audio amplifier should always be taken into consideration, as its losses result in a higher load on the power supply.

Pre：如何快速进行“批量化的PCB测试分析” Next：How to Design a PCB for the Enterprise IoT