Now that Class-D power amplifiers are common in professional audio, much attention has been directed toward their performance compared to conventional analog amplifiers.
Let’s briefly address the question of what Class-D means in order to clear up any confusion, because marketing people have conjured up several fictitious designations.
An amplifier that uses its output devices as switches is Class-D. It’s as simple as that. Everything else is analog, so “analog amp” means “not a switching amp.” The phrase “conventional amp” equates to “analog amp.” It is understood that the nomenclature can be confusing.
Class-D is the latest “wave” in topology because it offers high efficiency at reasonable cost. Early Class-D designs suffered from poor noise and distortion performance, but as the years rolled on and the technology continues to be developed, this has been addressed, and in some cases, cured.
Some of the “specsmanship” games pertain to consumer electronics more than pro audio, although the same pattern generally applies to all power amplifiers, regardless of intended application. There are several key amplifier specifications that get most of the attention when the “what can this amp do?” question is asked.
KEYS TO THE CASTLE
One key specification is noise, measured as signal-to-noise ratio (SNR). Analog amps are very refined in that they have been around since the beginning of audio, so noise is easily kept low. Noise in an analog amp has much to do with circuit design and layout, just as in a Class-D amp.
However, with Class-D, an additional factor contributes to noise performance. That factor is timing. In a switching amp, the timing of the switching transitions translates to the output voltage, so timing errors wind up being voltage errors, and timing “jitter” winds up being output voltage noise. Thus, designing Class-D for low noise is significantly more difficult than with analog.
Another key figure is distortion, measured as total harmonic distortion (THD). Actually, the measurement used in typical spec sheets is THD+N, where the N is noise. THD+N can be a better indication of audible performance because it takes into account the total error of the output.
This measurement is taken over a specified bandwidth, 10 Hz to 30 kHz, for example, where output error outside this bandwidth is disregarded. This works in favor of Class-D amps because many of them use noise shaping to push errors out of the audible spectrum. Analog amplifiers have been able to produce very low distortion for many years, but Class-D had a hard time catching up due to circuit design complexity and availability of key components.
Most amplifier design engineers consider THD+N of 1 percent to be the “onset of clipping” where output power should be measured as “maximum power.” The input test frequency can be fixed at 1 kHz, or range from 20 Hz to 20 kHz (and sometimes beyond), but this should be specified per measurement. The load resistance should also be stated.
In some cases, THD+N is measured at its lowest point along the THD+N versus power curve. And sometimes, THD+N is measured at a low power level such as 1 watt or 10 watts, even for higher-powered amps.
A more useful but less common measurement is THD+N at half power, which indicates output quality before distortion starts ramping up to the maximum power point.
Typical 1kW amplifier THD+N versus output power curve. (click to enlarge)
A graph representing THD+N versus output power (usually measured at 1 kHz) is particularly useful. For a broad power range, starting at milliwatts, THD+N usually decreases as power is increased until distortion overcomes noise. This part of the curve is the “noise dominates” section.
As power continues to increase, THD+N usually stays at its minimum until power approaches maximum. At that point, THD+N will rise until clipping is reached. The shape of the THD+N curve can tell you a lot about the amp.
The third and most important performance figure is power output, measured in watts. “How much power?” is the question most frequently asked, because it is important in determining just how much audio can be pumped into a given venue with a given loudspeaker system.
Class-D made its way into audio with the promise of big power in small packages due to high efficiency. One of the first applications for Class-D was driving car audio subwoofers, where noise and distortion are not nearly as important as raw output drive. In analog amps, power comes with the expense of heat dissipation, and the primary purpose of Class-D is providing more power with less heat.