A Class D amp works by taking the analogue input signal and creating a PWM (pulse width modulation) replica of it. This is essentially a train of pulses, which correspond to the amplitude and frequency of the input signal. In its most basic form, a comparator circuit is used to match the input signal with the PWM signal. The PWM signal is then amplified by an output stage operating in switch mode. So there are two states, on or off, at very high speed, corresponding to the PWM pulses. A linear amplifier’s output stages, by comparison, see a continuous waveform and, to avoid distortion, are on for more than half the waveform (Class A/B) or for the complete waveform (Class A), greatly reducing efficiency and generating heat.
The amplified PWM waveform is low pass filtered to recover the audio waveform and eliminate spurious ultrasonic noise before outputting it to the speakers. This process seems digital but is in fact analogue in nature. The signal is not “digitised”, or assigned a numerical value; the PWM pulse train is an “analogue” of the input audio signal. What distinguishes Rotel Class D amps from other designs on the market are innovations in the area of generating a highly accurate PWM signal (COM, which stands for Controlled Oscillation Modulation) and in the feedback circuits (MECC, Multivariable Enhanced Cascade Control) to provide a stable filter characteristic in spite of variable loudspeaker impedances. In simple terms, this means that our Class D amps offer full bandwidth performance at very low distortion in “real world” applications—just like our linear amps, while being smaller, cooler and much more energy efficient.
Why aren’t there more Class D amps on the market? For starters, creating stable, full bandwidth Class D circuits while controlling RF/EMI byproducts isn’t easy. Few companies have the technological know-how to do it. It also requires extensive use of Surface Mount Devices (SMDs), again putting it out of technological grasp of most audio manufacturers. We've put in a lot of work with technology partners to realise these designs.
Here’s another key detail that often causes confusion. The Switch Mode Power Supply (SMPS) is not what makes these “switching” amplifiers. As just described, the amplification stage is a high speed switching circuit and what defines this design as Class D. A Class D amp could, in fact, use a conventional power supply; and a linear amp could use an SMPS. A traditional power supply stores large amounts of energy, but wastes “excess” energy not demanded by the load in the process. An SMPS matches output to real-time requirements, supplying only the power required by the load, as a result operating very efficiently. The analogy is a water tank (linear supply), which is always being refilled and spills over if demand is insufficient; compared to an endless series of buckets (SMPS), which can be slowed down or sped up as required. The SMPS in our Class D amps reflects the fact that the Class D amplification circuit does not require the massive energy storage of a linear power amp, so the more efficient/compact SMPS is a better choice.