The Verdict: Which Amplifier Class Wins Your Build?
There is no single "best" amplifier class; the winner depends entirely on your power budget, thermal constraints, and fidelity requirements. For high-power home audio, active studio monitors, and battery-powered portable rigs, Class D is the undisputed winner due to its 90%+ efficiency and minimal heat output. For ultra-low-noise instrument preamps, boutique headphone amps, and RF transmission where switching noise is fatal, Class A or Class AB wins. If you are building a standard DIY desktop amplifier or subwoofer plate amp in 2026, buy a Class D module based on the TI TPA3255 or Infineon IRS2092; you will save money, board space, and heatsink weight without sacrificing audible fidelity.
The Single Physical Difference: Conduction Angle
Every difference in amplifier classes—efficiency, heat, distortion, and topology—stems from one single physical parameter: the conduction angle. This is the portion of the 360° input waveform cycle during which the output transistors are actively conducting current.
Think of an amplifier like a water valve controlling flow to a speaker:
- Class A (360° conduction): The valve is always wide open. To lower the flow, it bleeds water out a waste pipe. It responds instantly (zero crossover distortion) but wastes massive amounts of energy as heat.
- Class AB (180° to 360° conduction): Uses two valves (push-pull). One handles the positive wave, the other the negative. To prevent a "dead zone" when switching between valves, both are kept slightly open (biased) at idle.
- Class D (Switching/PWM): The valve doesn't throttle; it slams fully open or fully shut hundreds of thousands of times per second. The average flow is smoothed out by an LC filter (inductor and capacitor) before hitting the speaker.
Because transistors dissipate the most heat when they are partially on (acting as resistors) and the least heat when fully on or fully off, the conduction angle directly dictates your thermal management requirements.
Head-to-Head: Class A, Class AB, and Class D Compared
Here is how the big three audio topologies stack up when you look past the marketing and examine the datasheet realities.
| Criteria | Class A | Class AB | Class D |
|---|---|---|---|
| Conduction Angle | 360° (Always on) | >180° to 360° | N/A (PWM Switching) |
| Theoretical Max Efficiency | 25% (resistive load) to 50% (inductive) | 78.5% | 100% (Practical: 85-95%) |
| Typical THD+N at 1kHz (1W) | < 0.01% | 0.01% - 0.05% | 0.02% - 0.08% (Depends on LC filter) |
| Heat Dissipation (for 50W RMS out) | ~150W - 200W | ~30W - 50W | ~3W - 8W |
| Typical IC / Module Cost | Discrete only ($50+ in heatsinks) | $4 - $12 (e.g., ST TDA7293) | $2 - $15 (e.g., TI TPA3116D2) |
Numeric Worked Example: The 50W Heat Problem
Suppose you need to deliver 50W RMS into an 8-ohm speaker.
If you use a Class A design, the amp draws roughly 200W from the wall continuously, even at idle. You must dissipate 150W of heat, requiring a massive extruded aluminum heatsink weighing upwards of 4 kg per channel.
If you use a Class AB IC like the ST TDA7293, it draws about 85W to deliver 50W. You need to dissipate ~35W, requiring a moderate finned heatsink and perhaps a small 40mm cooling fan.
If you use a Class D module based on the TI TPA3255, it draws about 53W to deliver 50W. You dissipate just 3W. The IC package itself, mounted to a standard 2-layer PCB with thermal vias, is sufficient. No external heatsink is required.
Where They Are NOT Interchangeable (And Cost Realities)
While Class D has largely conquered consumer audio, you cannot blindly swap it into every application. The physical differences create hard boundaries.
Class D Pros & Cons
Pros: Runs cool, allows battery-powered high-wattage rigs, extremely cheap at scale.
Cons: Generates high-frequency EMI (electromagnetic interference) from the switching node. Requires careful PCB layout and output LC filtering. Can interact poorly with complex reactive speaker loads without proper Zobel networks.
Class AB Pros & Cons
Pros: Inherently stable into difficult loads, zero high-frequency switching noise, simple power supply requirements.
Cons: Heavy, runs hot, requires large transformers or high-current linear power supplies, lower efficiency drains batteries quickly.
Where they are NOT interchangeable:
- RF Power Amplification: You generally cannot use Class D for unfiltered RF transmission. The high-frequency PWM switching harmonics will bleed into adjacent radio bands, violating FCC/CE spectral masks. Class A or AB (or specific RF switching classes like Class E/F) are required.
- Ultra-Low Noise Sensor Preamps: In medical or precision piezoelectric sensor amplification, the nanosecond switching edges of Class D introduce broadband noise floors that swamp microvolt signals. Class A remains the standard here.
- High-Power Continuous PA Systems: You cannot use Class A for a 1000W stadium PA. The sheer physical size and cost of the copper and aluminum required to dissipate 3000W of heat makes it economically and physically impossible.
Regarding cost and availability, Class D has completely commoditized high power. A 2x50W Class D board using the TPA3116D2 can be sourced for under $10. A comparable Class AB build using LM3886TF chips (if you can find them in the current market) will cost $15 for the ICs, plus $30 for the linear power supply transformer, and $25 for the heatsinks. Class D wins the bill of materials every time.
Decision Framework: Choose Class AB When vs. Choose Class D When
Stop agonizing over "which sounds better." Modern Class D ICs easily achieve <0.05% THD+N, which is below the threshold of human hearing. Base your decision on physics and mechanics.
Choose Class AB when:
- You are building a guitar pedal or instrument preamp where even-order harmonic distortion (clipping behavior) is musically desirable.
- You are driving highly reactive, exotic electrostatic speakers that cause Class D output filters to ring or trigger over-current protection.
- You are designing a circuit where EMI shielding is impossible, and the PWM switching noise of Class D would interfere with nearby high-gain analog circuitry.
Choose Class D when:
- You are building a portable, battery-powered Bluetooth speaker (the 90% efficiency doubles your battery life compared to Class AB).
- You are designing an active subwoofer where the amp is mounted inside a sealed enclosure with zero airflow.
- You need to output more than 100W per channel without bolting a 2-pound block of aluminum to the back panel.
- You are integrating audio into a tight embedded system (like an ESP32 smart home hub) where board space is at a premium.
Frequently Asked Questions
What is the actual audible difference in amplifier classes for music playback?
In a properly engineered modern circuit, there is no audible difference. A well-designed Class D amplifier using post-filter feedback (like the Hypex NC400 or TI TPA3255) measures with less distortion and lower output impedance than most Class AB designs. The historical reputation of Class D sounding "harsh" or "brittle" stems from early 2000s designs that lacked adequate LC output filtering and suffered from dead-time distortion. If a Class D amp sounds bad today, it is a bad implementation, not a flaw in the topology.
Why do audiophiles still pay thousands for Class A amplifiers?
Class A amplifiers never suffer from crossover distortion because the output transistors never turn off. When a Class AB amp crosses the zero-voltage line, there is a microsecond where one transistor turns off and the other turns on, creating a tiny non-linearity. While negative feedback largely corrects this, some audiophiles prefer the absence of this mechanism entirely. Furthermore, when a Class A amp clips (overloads), it does so symmetrically and gradually, which is less fatiguing to the ear than the hard clipping of underpowered Class AB or D amps. You are paying for massive power supplies, exotic heatsinks, and boutique engineering, not necessarily measurable superiority.
Can I use a Class D amplifier IC without an output filter?
Only in very specific "filterless" architectures, and only with inductive loads like speakers. Some low-power ICs (like the TI TPA2005D1) use a specific modulation scheme that allows you to wire the speaker directly between two out-of-phase outputs (BTL), canceling the high-frequency switching carrier. However, for high-power Class D ICs, omitting the LC output filter will radiate massive EMI, potentially destroying the speaker's voice coil with high-frequency energy and causing the output transistors to overheat due to reactive current spikes. Always check the datasheet's recommended LC values.
What is Class G or Class H, and where do they fit?
Class G and Class H are essentially clever variations of Class AB designed to fix its efficiency problem. They use multiple power supply rails. At low volumes, the amp runs off a low-voltage rail (acting like a low-power Class AB, generating little heat). When the music peaks, the circuit automatically switches to a higher-voltage rail to deliver the required power. They are common in professional PA amplifiers (like those from Crest or Crown) where high fidelity and high power are needed, but the weight and heat of pure Class A or standard Class AB are unacceptable. They are largely being replaced by high-power Class D in modern touring rigs.
