A power amplifier uses energy from a power supply to deliver a controlled signal to a load. An audio amplifier drives a loudspeaker; an RF power amplifier drives a transmitter output network. The input controls the output, while the supply provides the additional energy. Choosing a device therefore requires a load specification, a signal specification and a thermal budget, rather than an isolated wattage figure.
How a power amplifier works
In a linear output stage, transistors vary their conduction as the signal changes. A switching stage instead uses a modulated pulse waveform and a suitable output path to deliver the desired signal to the load. Class D describes this switching operation; it does not specify whether the input is analog or digital. A speaker amplifier with an analog input can be Class D.
A practical design may contain input conditioning, gain control, a driver, an output stage and protection. These functions can be integrated into one IC or distributed among components. There is no universal three-stage circuit, mandatory input impedance or requirement for a series output capacitor. Single-ended and bridge outputs have different connections, so the manufacturer's application circuit must control the wiring.
Classes A, B, AB, C and D
For Classes A through C, conduction angle describes how much of a signal cycle an output device conducts. The following distinctions follow the Analog Devices power-amplifier glossary. They describe operating principles, not a sound-quality ranking for complete products.
| Class | Operating principle and selection concern |
|---|---|
| A | Conduction throughout the cycle. Standing bias creates heat even with little useful output. |
| B | Each device conducts for half a cycle in a complementary output stage. The transition between devices can introduce crossover distortion. |
| AB | Conduction extends beyond half a cycle. Bias reduces the crossover discontinuity but introduces quiescent dissipation. |
| C | Conduction occupies less than half a cycle. A tuned RF circuit can select the wanted frequency; this is not a general linear audio output stage. |
| D | The output devices switch. Modulation, filtering, switching noise and layout become important design constraints. |
A topology name does not establish a guaranteed efficiency. The examples in ADI's Class D explanation compare particular output stages and operating points. At low output, control and idle losses can become significant. At higher output, conduction loss and thermal limits matter. Read efficiency curves across the intended operating range instead of applying a single peak percentage to every load.
Switches also have finite transition times and parasitic capacitance. Their losses are not negligible simply because the stage is Class D. An external LC filter, cable arrangement and load can affect emissions and output behavior. Follow the device's filter guidance; the phrase “filter-free” does not establish that every speaker cable and enclosure will meet the required emissions limits.
Reading output-power specifications
For an ideal resistive load, average signal power is P = VRMS2/R. As a calculated example, a 10 V RMS sine wave across an 8-ohm resistor delivers 12.5 W, with 1.25 A RMS current. This assumes the voltage is measured across the load and the load remains resistive. It is not a measured amplifier result or a prediction for a loudspeaker whose impedance changes with frequency.
RMS describes the voltage or current used in this calculation. Calling a specification “RMS power” does not supply the missing test conditions. Record supply voltage, load, frequency, channel count, distortion limit and test duration. A burst rating and a sustained rating answer different questions.
TI's audio measurement guide, SLOA068A, covers THD+N, noise, gain and power measurements. THD+N includes both harmonic distortion and noise within the measurement bandwidth. An output-power value at 10% THD+N cannot be compared directly with one at 1%. Noise and SNR also require a reference signal and bandwidth; no single SNR value guarantees that a finished system will be inaudible at idle.
Example: checking TPA3116D2 ordering and conditions
The TI TPA3116D2 family datasheet, SLOS708G, illustrates why the conditions matter. For TPA3116D2, the electrical-characteristics table lists typical continuous output power of 50 W per channel in BTL at 21 V, a 4-ohm load, 1 kHz, 10% THD+N and 25 degrees Celsius ambient. This is not a 50 W claim at an arbitrarily lower distortion threshold. The datasheet also provides separate 1% and 10% output-power curves.
The TPA3116D2DAD ordering page identifies a catalog-grade, 32-pin DAD HTSSOP device with a specified ambient operating range of -40 to 85 degrees Celsius. The DAD thermal pad is on the top side. Confirm the complete order code, carrier and thermal arrangement before purchasing or laying out a board; a related family member is not automatically a substitution. These identifiers describe the manufacturer offering and do not certify an offered production lot.
The datasheet's main revision is December 2017, while its package addendum carries a later date. Those are document dates and do not establish the device's commercial launch date. An application listing for automotive audio also does not turn this catalog order code into an automotive-qualified variant.
Audio and RF applications need different comparisons
For audio, begin with the loudspeaker's impedance behavior, available supply, required acoustic output and enclosure cooling. Check gain, input level, channel configuration and startup or shutdown behavior. Leave enough design margin for supply tolerance and the actual program signal.
For RF, specify the operating band, modulation, output network and acceptable spectral distortion. Nonlinearity creates harmonics and intermodulation products, as explained in ADI's IP3 guide. An intercept point is a linearity comparison derived from signal behavior; it is not an available output-power rating. A tuned amplifier that works for one waveform or band requires separate review before use with a wider-band or amplitude-varying signal.
A practical selection and procurement checklist
- Write down the required load, signal bandwidth, channel count and allowable distortion before comparing devices.
- Ask engineering to identify the relevant datasheet rows and curves, including whether values are typical or guaranteed limits.
- Confirm the exact package drawing, thermal-pad position, land pattern and assembly requirements against the offered order code.
- Include the power supply, output network, heat removal and protection behavior in the design review.
- Keep a proposed replacement separate from an approved part until pinout, electrical behavior, thermal performance and application testing have been reviewed.
- Request current lifecycle, packaging and lot documentation separately when sourcing. An explanatory article is not a stock, delivery or certification statement.
Common questions
Does a higher wattage number mean a better amplifier?
It only helps when its conditions match the requirement. A part rated at greater distortion or a different load may offer less usable output in the intended design. Compare the complete conditions first.
Does Class D remove the need for cooling?
No. Use the chosen part's loss curves and thermal design instructions at the intended supply and load. Confirm behavior in the completed enclosure, including idle operation and sustained output.
Can a power amplifier be selected by class alone?
No. Class explains part of the operating method. The selection still depends on the exact device, load, signal, supply, package and verification requirements.