Home > Basic Circuits >Basic Practical Circuits > Causes and solutions for noise in digital Class D amplifiers

Causes and solutions for noise in digital Class D amplifiers

Source: InternetPublisher:黄土马家 Keywords: Noise Sound System Class D Power Amplifier Updated: 2025/01/16

With the development of electric vehicles, the number of channels and output power of car audio systems are gradually increasing. In audio and video entertainment systems, audio systems with high channel counts and high output power can produce greater sound pressure and dynamic range, with a stronger sense of wrapping and space, thereby achieving 360-degree surround sound with a theater effect. In addition to in-car entertainment, car audio systems also have many functions. Electric vehicles are quieter than traditional internal combustion engine vehicles. In order to protect pedestrians and reduce accidents, all new electric vehicles need to have an acoustic vehicle alarm system (AVAS) that emits appropriate sounds. In addition, in the emergency call (Ecall) system, the audio system can trigger anti-collision prompts and vehicle deviation warnings to allow drivers to contact emergency dispatchers. The audio system contains many parts, including power amplifiers, ADCs, Codecs, etc. in addition to speakers. Among them, Class D power amplifiers have emerged in the field of car audio with their advantages of high output power, high efficiency, and small size.

When a digital Class D amplifier is powered on or the amplifier is switched to another state, the human ear may occasionally hear a "bang" sound, which we call pop noise. There are many reasons for pop noise in digital amplifiers. This article mainly analyzes the causes of pop noise and provides corresponding solutions.

1) Capacitor charging and discharging

wKgaomSiml2AatVoAAGQg6PtWaI926.png

Figure 2. Schematic diagram of single-ended amplifier structure

Figure 2 shows a single-ended input amplifier. A1 is a comparator amplifier, which is used to set the gain and enhance the load capacity of the input signal. The output of A2 is completely opposite to the output of A1. Modulator is used for signal modulation, comparing the input analog signal with the triangle wave to generate a PWM wave to drive the peripheral MOS. One end of the comparator amplifier A1 is directly connected to the reference voltage Vref, and the other end is connected to the input audio signal through RIN and CIN. When the system is powered on, Vref immediately rises to the reference voltage value, while the other end of A1 needs to charge RIN and CIN for a period of time before it can rise to the reference voltage value. After the voltage difference across A1 is amplified, the output generates pop noise. In this scenario, the pop noise can be reduced by reducing the input capacitor value, such as replacing it with 1uf or 0.47uf.

For differential amplifiers, if the output peripheral hardware circuits of the P and N terminals do not match or the input peripheral hardware circuits do not match, the input signal establishment time at both ends of the amplifier will be different, and the differential signal difference will also be input into the amplifier and form a pop sound. As shown in the figure below, if the voltage rise speed at both ends of A1 is consistent, the pop noise is 0. A signal establishment time difference of 5ns can produce a pop sound that can be heard by the human ear.

wKgZomSiml6Aef9uAAC0oVWaAjM424.png

Figure 3 POP Noise at different charging speeds

As shown in Figure 4, the voltage of VR_ANA is converted from AVDD via LDO, which causes the voltage of VR_ANA to rise slower than AVDD. Here, 1.5V is obtained by dividing the AVDD 3.3V resistor, and a pre-bias voltage is placed on VR_ANA to ensure that VR_ANA rises at the same time as 3.3V, thereby reducing pop noise. Among them, the capacitor is used to eliminate the 3.3V noise.

wKgaomSiml-AKay3AABzi32FcYk218.png

Figure 4. POP noise suppression circuit

2) PWM start and stop

When the system is powered off or on, the amplifier playback state is switched, or the input audio source is switched, PWM will start and stop, thus generating transient POP sounds. As shown in the figure below, during continuous PWM operation, the switching frequency and its nearby mirror frequency can be successfully filtered out by the LC filter. When PWM is started and stopped, the switching frequency and its odd harmonics will extend into the 20-2kHz range audible to the human ear. This switching frequency is lower than the cutoff frequency of the LC filter and cannot be filtered out, thus generating pop sounds.

wKgZomSimmCAfxgmAAC-ktUf31c893.png

Figure 5. Time-frequency plot of continuous PWM and PWM start

When the BTL structured power amplifier performs AD modulation, the PWM starts the first duty cycle. If the A-side is pulled low, the Bootstrap capacitor can be charged smoothly, but the B-side is pulled high at this time, which causes the Bootstrap capacitor to fail to charge. The Bootstrap capacitor provides the charging voltage for the N MOSFET. If the Bootstrap capacitor fails to charge, the first PWM of the B-side cannot be output normally. The unbalanced output of the A-side and B-side will produce obvious POP sound. TI has optimized this type of pop noise. In both AD and BD modulation, the first PWM is made low, thereby eliminating the Clock fault.

wKgaomSimmCAe3LDAACOrNx9IwY792.png

wKgZomSimmKABDs0AACOrNx9IwY743.png

Figure 6 AD modulation PWM start diagram

3) Wrong power on and off sequence

The audio system has a strict power-on and power-off sequence. Usually the power supply voltage of the power amplifier is higher than that of the SOC, and the voltage build-up time is earlier than that of the SOC. To avoid pop noise when the SOC and the power amplifier are powered on, the power amplifier should be kept in Hi-zi/standby state, and the PWM wave should be turned on and the audio source should be input after the power amplifier is fully charged (20ms). Similarly, when the power amplifier is powered off, in order to avoid inconsistent power-off speed, we need to Mute and pull the Standby pin low for 15ms before powering off. TI's PurePath Digital has an optimized startup sequence, which makes the pop noise in the audible audio band as small as possible.

4) PVDD voltage/Gain value rises rapidly

A sharp rise in PVDD voltage or a rapid increase in Gain value will cause pop noise. When drawing the schematic diagram, the Cstart soft-start capacitor needs to be set within a reasonable range to prevent PVDD from rising rapidly. In addition, for some amplifiers, after the first POP noise appears at power-on, a second pop noise also appears. This is because after the amplifier is powered on, the gain value will climb to the set gain in a certain step length. If the step length setting value is too large, pop noise will occur.

5) Hizi-play status switch Clock Fault

If the speaker not only produces pop noise when it is turned on or when the state changes, but also produces pop noise continuously when the amplifier switches from Hi-Zi to Play, then you should check whether a Clock Fault has occurred. Hardware engineers can disconnect the IIC control of the SOC and connect the IIC to the PPC3 through a USB adapter board to perform a Clock Fault test.

wKgaomSimmOAamnfAADKWGJFlWE769.png

If a Clock Fault occurs at this time, check whether the input audio signal I2S/TDM meets the requirements of the data sheet (see Serial Audio Port in the Electrical Characteristics of the data sheet). In addition, there are other special cases in the data sheet. Taking the TAS6424L/M-Q1 series as an example, if the customer connects SCLK and MCLK together, FSYNC needs to be more than 2 MCLK. If the SOC is the Qualcomm 8155 series, there are 3 options for FYSNC output: the first is 2MCLK, the second is 50% duty cycle, and the third is 1 slot. We can choose the latter two as FSYNC input.

Texas Instruments TAS6424E-Q1 is a four-channel digital input Class D audio amplifier with a 2.1MHz switching frequency. In terms of cost, the chip operates at a frequency of 2.1MHz, which allows the chip to use smaller and lower-cost LC filters, thereby achieving overall cost optimization. While increasing the switching frequency, the TI TAS6424E-Q1 chip has good EMI performance through spread spectrum and PWM sequence optimization. In addition, the chip integrates AC and DC fault diagnosis, which can realize fault diagnosis such as load short circuit to power supply, load short circuit to ground, load open circuit, load short circuit, and realize high-precision load impedance and phase measurement. In addition, the TAS6424E-Q1 chip integrates power-on startup sequence optimization and the first PWM is low to suppress pop sound solutions, achieving a good user listening experience. In addition to low pop noise, the chip's Burr-Brown audio architecture and increased internal audio loop bandwidth can also provide excellent sound quality and bring a good user experience.

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号