What is a Class T Amplifier?


A Class T amplifier is an audio amplifier IC design. Rather than being a separate “class” of the amplifier, Class T is a registered trademark for Tripath’s amplifier technologies. The control signals in Class T amplifiers may be computed using digital signal processing or fully analog techniques.

The covered products use a class-D amplifier combined with proprietary techniques to control the pulse width modulation. In this way, they produce better performance than other class-D amplifier designs. Among the publicly disclosed differences is real-time control of the switching frequency depending on the input signal and amplified output. One of the amplifiers, the TA2020, was named one of the twenty-five chips that shook the world by the IEEE Spectrum magazine.


Digital or Analog


You can compute the control signals in Class T amplifiers using digital signal processing or fully analog techniques. Currently, available implementations use a loop similar to a higher-order Delta-Sigma (ΔΣ) (or sigma-delta) modulator, with an internal digital clock to control the sample comparator. The two critical aspects of this topology are that (1) it takes the feedback directly from the switching node rather than the filtered output, and (2) the higher-order loop provides much higher loop gain at high audio frequencies than would be possible in a conventional single-pole amplifier.

Financial difficulties caused Tripath to file for Chapter 11 bankruptcy protection on 8 February 2007. Cirrus Logic purchased Tripath’s stock and intellectual property later that year.


Signal Processing Circuit


A signal processing circuit includes a frequency selective network in a feedback loop for noise shaping purposes. Moreover, the sampling analog-to-digital converter in the feedback loop operates at a sampling frequency substantially above the Nyquist frequency. A sampling analog-to-digital converter drives the switching device. It produces a continuous-time output signal which feeds back to the frequency selective network for noise and distortion correction in the feedback loop. This is in contrast to traditional techniques that employ only state feedback. You can also employ the state feedback (i.e., digital or sampled) of the output of the analog-to-digital converter in combination with the continuous-time feedback of the switching device output.


Source Text