Look Who’s Talking Now–Give Your IoT Device a Voice Worth Hearing

In the not too distant past, if you were seen holding a conversation with a household appliance, you might have run the risk of being committed. However, the reality of the current age of the Internet of Things or IoT (Figure 1) means it is now possible for your oven to ask you how you would like your meat to be cooked, your gaming module to ask you what game you wish to play, or your internet-connected personal assistant to remind you to take an umbrella with you because it’s raining outside. In short, sound is fast becoming the interface of choice for communicating with the next generation of consumer electronics that proliferate within our homes and businesses.

Figure 1

Internet of Things

Internet of Things

Formerly the sole preserve of top-end music systems, high-quality audio is becoming “de-rigueur” in an increasing number of applications. Accordingly, electronic equipment manufacturers are looking for the easiest way possible to integrate high-quality audio into their devices, without tearing their hair out or breaking the bank. Due to its high efficiency and PSRR, the filterless Class D amplifier has become the accepted norm for driving speakers in consumer electronic equipment. In this design solution, we will review the evolution of the use of the Class D amplifier in audio systems.

After considering the limitations of the analog and traditional digital approaches, we will introduce a new alternative that enables quick and easy integration of high-quality audio into all types of IoT devices.

Analog Class D Amplifiers

Analog-input Class D amplifiers normally require a DAC and line driver on the application processor (Figure 2). This increases both die size and power consumption while adding noise to the speaker output. These Class D amplifiers also require careful board design to avoid degradation from signals coupling onto the analog board routes. To minimize noise coupling, a stereo analog-input Class D amplifier will also normally require two differential input signals (four wires) to be routed to the speakers, along with AC-coupling capacitors, requiring more board traces with greater layout complexity. Two full output filters are also required, a total of four large inductors, which further increases design space and cost.

Figure 2

Analog Class D Amplifier

Analog Class D Amplifier

Digital Class D Amplifiers

Unlike their analog counterparts, digital-input Class D audio amplifiers are immune to most board design issues. Single-channel Class D amplifiers can be placed at remote locations on a board to minimize the routing of high-current battery and speaker load connections. These amplifiers do not need the DAC and line driver required by analog-input Class D designs. Thus, space and system costs drop, and designs become simpler. The most common type of digital input for an amplifier is pulse-density modulation (PDM) which requires only two wires: PDM_CLK and PDM_DATA (Figure 3). Single-bit PDM data is created with an oversampled sigma-delta modulator on the application processor. In comparison to the analog approach, only PDM_CLK and PDM_DATA lines are needed for PDM inputs to provide stereo data to the two Class D amplifiers.

Figure 3

2-Wire PDM-Input Class D Amplifier

2-Wire PDM-Input Class D Amplifier

Some amplifiers accept pulse-code modulated (PCM) or I2S data which requires three wires: BCLK, LRCLK, and DIN (Figure 4). The PCM data format does not require a modulator or upsampling of the data on the application processor. The BCLK, LRCLK, and DIN lines are needed for PCM inputs to provide stereo data.

Figure 4

PCM-Input Class D Using Three Wires

PCM-Input Class D Using Three Wires

However, there are limitations associated with some traditional implementations of digital-input amplifiers. Some of these require a clean master clock (MCLK) to derive a jitter-free sampling clock. Others offer adjustable sample rate and/or bit depth, but this can require complex programming of the amplifier. Also, most digital-input amplifiers require both a low digital-supply voltage (1.8V) and a high speaker-supply voltage (2.5V to 5.5V).

A Simpler Alternative

Integrating high-quality audio into any type of electronic equipment can now be done even more simply with a new type of digital-input Class D amplifier. These devices accept PCM input (I2S or left justified, as required).

Figure 5

MAX98357 Digital Class D Amplifier

MAX98357 Digital Class D Amplifier

Unlike older Class D amplifiers, a Pulse Code Modulation (PCM)-input audio amplifier uses automatic sample rate and bit depth auto configuration to eliminate the need for complicated I2C programming. By supporting clocking rates from 8kHz to 96kHz, it has the advantage of auto-detecting 21 PCM + 14 8-channel Time Division Multiplexing (TDM) clocking configurations. Other benefits include the fact that an MCLK is not required and the device can operate off a single supply voltage (reducing pin count and complexity). The added option for a tiny 1.9mm2 Wafer Level Package (WLP) eliminates the need for expensive board vias. Connecting the GAIN_SLOT pin according to Table 1 allows a choice of fixed gain settings.

Table 1

Gain Settings for MAX98357A/MAX98357B

Gain Settings for MAX98357A/MAX98357B

However, for applications that require variable gain settings, the GAIN_SLOT pin can easily be routed using only a tiny 1.75mil trace with 1.75mil clearance, further simplifying board design and cost. For applications that require multiple speakers, a further benefit of these devices is they allow up to eight speakers to be daisy-chained (with a single TDM input) compared to the more complex, conventional approach of using the Audio Processing Unit (APU) to drive four speaker pairs with separate I2S inputs (Figure 6).

Figure 6

Simpler Design Using an 8-Channel TDM

Simpler Design Using an 8-Channel TDM


With the audio interface fast becoming a ubiquitous feature of IoT devices and other types of portable electronic equipment, designers are looking for simpler and more cost-effective ways to add high-quality audio to their devices. In this design solution, we reviewed the shortcomings of using analog and traditional digital Class D amplifiers for this purpose. We can conclude that the new breed of digital-input Class D audio amplifiers bring “Plug’n’Play” simplicity to the task of integrating audio into any type of electronic equipment. As well as being suitable for use in laptops, notebooks, and tablets, these audio amplifiers are also suitable for use in gaming devices and internet-connected smart-home devices.


Michael Jackson has over 20 years’ professional experience as an Analog IC Design Engineer and holds the position of Senior Technical Writer at Maxim Integrated. He has a MSEE from Dublin City University.

Gregory Mow is Business Manager for Audio Solutions at Maxim Integrated. He has a BS in Electrical Engineering at the University of California, San Diego

1 comment on “Look Who’s Talking Now–Give Your IoT Device a Voice Worth Hearing

  1. DavidHamilton
    January 16, 2019

    I personally find it a little creepy to think that there's a voice coming out of my inanimate objects, and even more so when it knows more about what's going on in my house than I do. But obviously there are benefits to that. I would just prefer everything be presented to me on a screen rather than audibly, thank you very much!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.