An audio signal chain begins with a microphone and ends with a speaker. In between those two components, you can have very simple circuitry or very complicated circuitry. It all depends on what you’re trying to accomplish and what your design constraints are.
Two fundamental aspects of an audio circuit are processing and amplifying. Processing can refer to something as basic as a low-pass filter or as complex as a compression IC. Amplifying (I’m referring to the final amplification stage here) is what turns your processed audio signal into something that can drive a speaker. The amplifier might increase the voltage level of your signal, or it might simply buffer the signal in such a way as to allow the existing voltage to supply more current to the load (i.e., the speaker coil).
Amplifiers are grouped into different “classes”—class A, class B, etc. Nowadays, the class D amp is quite popular. This topology can achieve very low power dissipation (or, in other words, very high efficiency) because it employs transistors that are fully on or fully off. Very-high-efficiency audio amps are particularly important these days because they help to extend the battery life of portable audio systems (such as Bluetooth speakers).
Here, we'll go over three ICs: two class D amps used for amplifying audio signals and one codec used for processing them.
The TPA3128D2 is a stereo class D amplifier that incorporates a variety of useful features, including integrated protection circuitry, high PSRR, and six switching-frequency options.
Image courtesy of Texas Instruments
The low on-resistance of the integrated driver FETs allows the part to provide high output current (7.5 A) without burning itself up. But, as always, thermal issues depend on various factors, so you’ll need to look at the detailed specs to determine if you need a heatsink. At the very least, connect the thermal pad to plenty of copper and/or vias!
Another new class D option is the BM28723MUV from ROHM. This amplifier doesn’t provide as much output power as the TPA3128D2, but it incorporates a digital-signal-processing functionality described as “audio signal processing for TVs”. As with the TPA3128D2 and so many other ICs these days, it incorporates various useful features that are conveniently summarized on page 1 of the datasheet.
Image courtesy of ROHM Semiconductor
As you can see in the application circuit, audio comes to the BM28723MUV as digital data, not as analog signals. This is fine if you have some other reason to digitize your audio, and it’s especially handy when you are starting with digital data instead of analog signals (such as with playback of digital music files). Just pass the data to the BM28723MUV and it can worry about the analog business.
On the other hand, sometimes we prefer to handle everything in the analog realm. In such cases, you’ll want to focus on analog-input amplifiers such as the TPA3128D2.
For those of us who are more concerned about the signal-processing portion of the design, let’s take a look at the ADAU1777.
Image courtesy of Analog Devices. Click to enlarge
This rather complicated chip performs digital signal processing, but it has analog input and output. In other words, you get the benefit of DSP without worrying about custom data-converter circuitry; the ADAU1777 handles the analog-to-digital and digital-to-analog conversion.
The signal-processing capabilities are described as “filtering, level control, signal level monitoring, and mixing”. However, the datasheet makes it very clear that this part is optimized for a particular task—namely, noise cancellation.
Also, in contrast to the two previous ICs, this part is not intended for high-volume applications. (I mean volume as in loudness . . . not as in production volume.) However, this doesn’t mean that you will always need an additional amplifier: the ADAU1777’s analog output signals can drive headphone speakers with impedance of at least 16 Ω.
Are there any new audio products that look promising to you? Feel free to mention them in the comments.