A Class A power amp is an amplifier topology where the output transistors conduct current continuously for the full 360 degrees of the input signal cycle, eliminating crossover distortion but generating massive amounts of heat. What this changes in your physical installation is the thermal and power infrastructure: you must design for continuous heat dissipation equal to the amplifier's maximum idle draw, requiring massive extruded aluminum heatsinks, over-specced toroidal transformers, and careful PCB layout to manage thermal drift.
The Core Mechanics: 360-Degree Conduction and Crossover Distortion
In Class B or Class AB amplifiers, the positive and negative halves of an audio waveform are handled by separate transistors. As the signal crosses the zero-volt line, one transistor turns off and the other turns on. Because transistors have a turn-on threshold (roughly 0.6V for BJTs, higher for MOSFETs), there is a microscopic dead zone where neither device is fully conducting. This creates crossover distortion, which introduces high-order harmonics that are particularly fatiguing to the human ear.
A Class A power amp solves this by biasing the output devices so they are always fully on. Think of a Class A output stage like a water valve that is permanently wedged half-open; water (current) flows constantly whether you are drawing from the pipe or not, and the excess pressure (voltage) is burned off as heat. Because the transistors never switch off, there is no transition dead zone, resulting in a perfectly continuous waveform and a harmonic profile dominated by even-order harmonics, which sound musically natural.
Worked Numeric Example: Sizing a 15W Class A Power Amp
Let's design the thermal and power requirements for a push-pull Class A amplifier delivering 15W RMS into an 8Ω speaker load. This is a common target for desktop or near-field audiophile monitors.
- Calculate Peak Voltage and Current:
Vpeak = √(2 × Power × Resistance) = √(2 × 15 × 8) = 15.49V.
Ipeak = Vpeak / Resistance = 15.49 / 8 = 1.93A. - Determine Quiescent Bias Current (Iq):
In a push-pull Class A design, the quiescent current per rail must be at least half the peak output current to maintain Class A operation across the full swing.
Iq = 1.93A / 2 ≈ 1.0A per device. - Size the Power Supply:
We need rails higher than the peak voltage to account for transistor saturation and ripple. Let's use ±20V DC rails.
Total continuous DC power drawn from the supply = (20V × 1.0A) + (20V × 1.0A) = 40W per channel. - Calculate Heatsink Thermal Resistance (θSA):
At idle (zero audio output), the full 40W is dissipated as heat in the output transistors. We must size the heatsink for this worst-case continuous scenario.
Assume maximum ambient room temperature = 35°C.
Target maximum transistor case temperature = 75°C (safe for silicon, hot to touch).
Allowable temperature rise (ΔT) = 75°C - 35°C = 40°C.
Required Heatsink θSA = ΔT / Power = 40°C / 40W = 1.0°C/W.
Where You Meet Class A in Practice
You will rarely find true Class A topology in high-power home theater receivers or portable Bluetooth speakers. Instead, it dominates specific niches where linearity trumps efficiency:
- Headphone Amplifiers: Headphones require less than 1W of power. A Class A headphone amp dissipates only a few watts of heat, making it easy to cool while delivering pristine, low-impedance drive for planar magnetic headphones.
- RF Power Amplifiers: In amateur radio and commercial broadcasting, single-ended Class A is used for low-level driver stages where preserving the exact amplitude and phase of AM or SSB signals is critical, and crossover distortion would cause catastrophic spectral splatter.
- High-End Audiophile Stereo: Brands like Pass Labs and Sugden build massive, heavy Class A amplifiers for purists who prioritize the complete absence of switching artifacts and the specific harmonic signature of continuously biased MOSFETs or BJTs.
Common Confusions: True Class A vs. Marketing Hype
The most frequent point of confusion on the bench and in retail showrooms is the term 'Class A/B with Class A bias'. Many manufacturers advertise amplifiers that operate in Class A for the 'first 5 watts' or 'first 10 watts' before switching to Class B.
This is not a true Class A power amp; it is a high-bias Class AB design. While it successfully eliminates crossover distortion at low listening volumes (where human hearing is most sensitive to it), the moment the signal exceeds that 5W threshold, the output stage switches to Class B operation. True Class A amplifiers never switch off their output devices, regardless of whether you are playing a 1mW test tone or driving the amp into hard clipping.
Another confusion is single-ended vs. push-pull Class A. Single-ended Class A (using one transistor per channel) is limited to 25% efficiency and requires output transformers or massive DC-blocking capacitors. Push-pull Class A (using complementary N/P pairs) cancels even-order harmonics at the output, doubles the efficiency to roughly 50% theoretical, and allows direct coupling to the speaker. Most modern DIY and commercial Class A amps are push-pull.
Decision Tree: Should You Build or Buy a Class A Power Amp?
Use this framework to determine if Class A is the right topology for your next project, or if you should pivot to Class AB or Class D.
| If your project requirement is... | Then choose this topology... | Recommended Part / Platform |
|---|---|---|
| >100W per channel, subwoofer drive, or minimal heatsink space | Class D (Switching) | Hypex NC400 or Purifi Eigentakt modules |
| 50W-80W, general home stereo, reliable daily driver | Class AB (Linear) | LM3886 IC or standard discrete BJT designs |
| <20W, near-field monitors, zero crossover distortion mandate | Push-Pull Class A | DIYAudio ACA or Pass Labs XA series |
| Headphone drive (<2W), ultra-low noise floor | Single-Ended Class A | Discrete JFET designs or Burr-Brown BUF634 |
FAQ: Thermal Runaway and Biasing Edge Cases
Q: What is thermal runaway in a Class A amp, and how do I prevent it?
A: As bipolar junction transistors (BJTs) heat up, their base-emitter voltage (Vbe) drops, causing them to draw more bias current. More current creates more heat, which drops Vbe further, leading to a destructive feedback loop that melts the silicon. To prevent this, you must use a Vbe multiplier bias spreader thermally coupled directly to the output device heatsink. As the heatsink warms, the spreader reduces the bias voltage, stabilizing the current. MOSFETs are inherently safer here, as their positive temperature coefficient at high currents naturally limits thermal runaway, though low-gate-threshold MOSFETs still require thermal tracking.
Q: Why does my Class A amp have 50mV of DC offset at the speaker terminals?
A: Because Class A amps are often direct-coupled (no output capacitors) to preserve low-frequency phase response, minor mismatches in the input differential pair or thermal gradients across the PCB can cause DC drift. 50mV is generally safe for 8Ω woofers, but if it exceeds 100mV, you need to implement a DC servo loop (an op-amp integrator in the feedback path) to force the output to absolute zero, or ensure your input JFETs/BJTs are tightly matched using a curve tracer.
Q: Can I use a switching power supply (SMPS) instead of a toroidal transformer?
A: Yes, but with caveats. A Class A amp draws a constant, unvarying current from the rails regardless of the audio signal. This is actually the ideal load for an SMPS, as it avoids the dynamic current spikes that cause transformer sag in Class AB amps. However, you must ensure the SMPS has exceptionally low high-frequency ripple, as the power supply rejection ratio (PSRR) of simple Class A topologies is often poor, and switching noise will inject directly into your audio band.
