Class-D Amplifier PCB Routing: EMC, Power, and ESD Essentials

Class D audio amplifiers achieve high efficiency by switching output MOSFETs at high frequency. That switching behavior makes PCB layout and power delivery part of the EMC design, not merely a packaging task. Every PCB trace and every metal structure that can behave as an EMI radiator should be as short as practical, including the connections from the power-supply output to the Class D amplifier output and from the supply to the loudspeaker. The amplifier is also unusually sensitive to supply performance: because the output stage directly switches one of the supply rails, any supply fluctuation or ripple can appear at the output as noise or distortion. This article covers the layout and decoupling measures that keep a Class D amplifier stable and quiet, then moves to two practical EMC cases: ESD/FTB-induced lockup in a USB handheld device with a headset jack, and the ESD test practice required for an RS-485 interface in electricity meters.

Highlights:

• Shorten all EMI-exposed metal paths in a Class D board, especially supply-to-output and supply-to-speaker connections, and keep the power stage clean across the entire audio band.
• Use low-ESR reservoir capacitors with sufficient charge, placed as close as possible to the output stage, to minimize parasitic resistance; a PWM phase delay below 1 μs reduces instantaneous switching demand.
• For headset-jack ESD/FTB events, add discharge paths at the jack pins and use back-to-back capacitors on audio lines; RS-485 ESD verification requires defined air-discharge and contact-discharge procedures with the device operating normally.

D-Class Amplifier PCB Routing and EMC

In a Class D amplifier board, the PCB traces and metal parts that exhibit EMI behavior should be kept as short as possible. This applies to the metal connections from the power-supply output section to the Class D amplifier output section and from the power supply to the speaker. Another long-standing issue with Class D amplifiers is their extreme sensitivity to power-supply performance. Since the amplifier output always directly switches one of the supply rails, any change or fluctuation on the supply rail is reflected at the output signal and appears as noise or distortion. A Class D amplifier therefore needs a supply current that has good load regulation, is clean, and is low-noise, not only at DC but across the entire audio band. The operation of the power-supply transistors is equally important.

Power-Supply Sensitivity and Reservoir Capacitor Placement

In a Class D amplifier, the high-frequency pulse output section is powered by the supply current. To produce accurate square-wave pulses at the amplifier output, the supply voltage must remain stable; ripple and noise on that voltage are strictly undesirable. The storage capacitor is a key component here. First, it must hold enough charge to keep the supply voltage stable. Second, any parasitic resistance or interference path from the supply capacitor can transfer quickly to the output, so a low-ESR (effective series resistance) capacitor is required. Parasitic resistance in PCB metal traces is detrimental to supply stability. The storage capacitor should therefore be placed as close as possible to the output section to minimize parasitic resistance.

Supply current demand can be eased by introducing a short delay of less than 1 μs. This delay is set between the individual outputs in a stereo amplifier or between channels in a multichannel system. The delay is so short that the human ear cannot perceive it. Because each output's MOSFETs switch at different times, the number of switching transistors active at any one instant is effectively reduced. This technique is commonly called PWM phase technology and is used in many Class D IC designs.

USB Handheld Device Headset-Jack Lockup

Q1: A USB handheld device locks up when the headset is plugged in or removed. Oscilloscope measurements at the headset-jack pins show transient impulse voltages, suspected to be ESD or FTB interference. The problem disappears when a shielded USB cable is used, and it also disappears when the PC grounding is intact. What can be done when an unshielded cable must be used under all possible conditions? Adding an inductor to the ground line noticeably reduces ground-line interference, but what should be added to the audio lines to prevent lockup without affecting the audio signal?

A1: The transient at the headset-jack pins should be suppressed with a capacitor from each pin to ground; this is a practical way to remove the spike. The mechanism is straightforward: when analog information suddenly disappears, interference is generated. Without a good ground, this phenomenon is very likely. The solution is to provide a circuit that absorbs and discharges the transient. The simplest method is to add capacitance to ground, but this can affect audio quality. When selecting the capacitors, use two capacitors connected back-to-back rather than a single capacitor. Adding an inductor in the ground line can reduce ground-line interference, but the audio lines still need a discharge path; the back-to-back capacitor arrangement provides that path while limiting the effect on the audio signal.

RS-485 Interface ESD Test Practice

Q2: DL/T645-1997, the multifunction watt-hour meter communication protocol, specifies RS-485 standard electrical interface performance and requires the driver and receiver to withstand electrostatic discharge (ESD) of ±15 kV under the human body model. What is the human body model test method, and how does the human body model differ from air discharge?

A2: In one human body model test setup, the equipment is placed on an insulating wooden board approximately 10 cm thick. An electrostatic gun is used to discharge at 6 kV to a metal plate whose plane is parallel to the display and control section of the equipment under test. The gun is also used at 8 kV against metal parts of the enclosure, with one discharge per second. The electrostatic gun has both a pointed tip and a round, finger-simulating tip. Air discharge uses a blunt discharge tip at 8 kV, held about 1 cm from the equipment under test while searching for a discharge point. For a mixed metal and plastic enclosure, if the enclosure is plastic, bring the tip close to the surface to find the point. If a discharge point is found, apply one discharge per second, 20 discharges per polarity, for a total of 40 discharges per test point. Contact discharge uses a pointed tip and contacts metal bodies on the surface of the equipment under test. If the metal enclosure has a large area, select multiple evenly distributed points and test each one separately, again one discharge per second, 20 per polarity, 40 per test point. In both tests, the equipment must be operating normally. If a fault occurs during discharge, faults are classified into three levels: (1) the equipment stops but recovers automatically; (2) the equipment stops but recovers with manual intervention; (3) permanent damage. For general commercial standards, level (1) is acceptable.