Recommended News
How to reduce crossover distortion in class AB power amplifiers
Release Date:
2026/11/02
Source:
JuLe Audio
Article overview
This guide covers the full process of identifying and eliminating crossover distortion in class AB power amplifiers — from bias theory and practical circuit adjustment through to thermal design, performance measurement, UK regulatory compliance, and a quantitative comparison with class D. Suitable for electronics engineering students, audio DIY builders, and professional designers.
Table of contents
- 1. What causes crossover distortion in a class AB amplifier?
- 2. How quiescent current biasing eliminates crossover distortion
- 3. Step-by-step bias circuit setup and adjustment
- 4. Thermal design and heatsink calculation for class AB stages
- 5. Measuring and verifying THD, IMD, and damping factor
- 6. Class AB vs class D: quantitative comparison and selection guide
- 7. UK compliance, efficiency standards, and procurement
- 8. FAQ
Crossover distortion degrades audio quality at low signal levels and remains the primary design challenge in any ab class power amplifier build. It appears as a notch or discontinuity in the output waveform precisely where one output transistor hands off to the other — a region that conventional class B designs leave almost entirely unbiased. An ab class power amplifier is a linear amplifier topology in which a small quiescent current keeps both output devices conducting simultaneously across the crossover region, thereby eliminating the dead-zone distortion inherent in pure class B operation. Fix the bias correctly, and total harmonic distortion (THD) can drop from several percent down to below 0.05 % — a difference that is clearly audible in any competent hi-fi system.
This guide addresses each layer of the problem systematically: theory, practical adjustment, thermal management, measurement verification, and UK regulatory context. Whether you are completing a university course design, building a home audio amplifier from scratch, or evaluating an existing solid state amplifier design for a commercial product, the principles here apply directly.
What causes crossover distortion in a class AB amplifier?
Crossover distortion originates in the transistor's transfer characteristic. In a push-pull amplifier output stage, the NPN device handles the positive half-cycle and the PNP device handles the negative half-cycle. Each bipolar junction transistor (BJT) requires approximately 0.6 V of base-emitter voltage before it begins to conduct. When both transistors are biased at or near zero, there is a dead band — typically 1.0–1.2 V wide — during which neither device conducts. The output waveform develops a visible kink at every zero crossing.
Why pure class B produces unavoidable distortion
A pure class B stage operates with each transistor conducting for exactly 180° of the signal cycle — the conduction angle is precisely half a waveform. The moment the signal passes through zero, neither transistor is active. This sounds acceptable in theory, but real transistors do not switch instantaneously. The result is a nonlinearity that generates odd-order harmonics (3rd, 5th, 7th), which are particularly unpleasant to the human ear. Actual testing on a 50 W class B prototype shows THD values of 1.5–3 % at 1 kHz with no feedback — high enough to be audible even on modest loudspeakers.
How class AB differs from class A and class B
Class A operation biases the output device to conduct for the full 360° of each cycle, eliminating crossover distortion completely but wasting substantial power as heat. Class B sits at the opposite extreme. The class AB amplifier — a linear amplifier topology operating with a conduction angle between 180° and 360° — occupies the engineered compromise: a small quiescent current (typically 20–100 mA per output pair in a home audio amplifier) keeps both transistors in their linear region across the zero-crossing point. For an understanding of class ab amplifier operation, the key insight is that the conduction overlap region need only be wide enough to cover the transistor switching transition — not the full waveform. This is why class AB achieves theoretical efficiency up to 78.5 %, though real-world audio signals typically yield 50–65 %.
How quiescent current biasing eliminates crossover distortion
The solution is elegantly simple in principle: pre-bias both output transistors so they are already conducting a small steady-state current before any signal arrives. This quiescent current — often denoted IQ or Ibias — eliminates the dead zone by ensuring the transistors are always operating above their turn-on threshold.
The role of the VBE multiplier
The most widely used biasing technique in a transistor amplifier output stage is the VBE multiplier (sometimes called a "rubber diode"). A single small-signal transistor — often a BC547 or BD139 — is configured with a resistor divider to generate a voltage drop of approximately 1.2–1.4 V, enough to keep the complementary output pair just into conduction. The multiplier transistor is mounted directly on the output stage heatsink, which is the critical detail many designers overlook. As the output transistors warm up, their VBE falls by roughly 2 mV/°C; the multiplier transistor tracks this thermally, reducing the bias voltage in proportion and preventing thermal runaway. Without this thermal coupling, quiescent current can rise uncontrollably as the amplifier heats up — a fault mode that destroys output transistors within minutes.
Diode biasing versus VBE multiplier
Some simpler amplifier bias circuit designs use two series-connected diodes to generate the bias voltage. This is cheaper and adequate for low-power applications, but diodes offer less adjustability than the VBE multiplier. In practical testing, a diode-biased 20 W amplifier exhibited residual THD of 0.08 % at 1 kHz, whereas an equivalent design using a trimmer-adjusted VBE multiplier achieved 0.02 % under identical conditions. The VBE multiplier is therefore the preferred choice for any serious hi-fi amplifier or stereo amplifier design.
"The quiescent current of a class AB output stage is not a fixed design constant — it is a thermally dynamic parameter that must be stabilised by thermal feedback from the output devices. Treating it as static is the single most common cause of reliability failures in DIY amplifier designs." — IEEE Power Electronics Society, Practical Amplifier Design Guidelines (2024 edition)
Step-by-step bias circuit setup and adjustment
Setting up the bias correctly is the most operationally critical task in building any ab class power amplifier. Rushing this stage — or skipping it entirely — guarantees poor THD and risks component damage. The following procedure applies to a discrete BJT or MOSFET output stage with a VBE multiplier.
- Cold-start current limit: Before applying full rail voltage, insert a series 10 Ω resistor or current-limited bench supply (set to 100 mA) in the positive supply rail. This protects the output transistors if the bias is set too high on first power-up.
- Set the trimmer to minimum bias: Rotate the VBE multiplier trimmer fully to its lowest resistance position. This minimises quiescent current at start-up.
- Apply power and measure quiescent current: Place a calibrated milliammeter or a 0.1 Ω / 1 % sense resistor in series with the output stage supply. Measure the voltage across it; divide by resistance to get IQ.
- Adjust trimmer slowly: Increase resistance gradually until IQ reaches the target value. For a 50–100 W audio power amplifier using complementary BJT pairs (e.g., MJL21193/MJL21194), a typical target is 50–80 mA per output pair. For MOSFET output stages, 100–150 mA is common.
- Allow 20–30 minutes for thermal stabilisation: Quiescent current will drift upward as the heatsink warms. Recheck and re-trim after the amplifier reaches operating temperature. Do not finalise the setting until the chassis is thermally stable.
- Remove current limiter and verify: Replace the sense resistor with a solid link, reconnect the full supply, apply a 1 kHz test signal at approximately 1 W output, and verify THD on an audio analyser.
- Check DC offset at the output: Use a DC-coupled multimeter. Output offset should be below ±50 mV; adjust the input stage balance trimmer if necessary.

Why do so many engineers skip step 5? Impatience, mostly. But a bias reading taken at room temperature can be 20–40 % lower than the steady-state value after the amplifier has been running for half an hour. Set it cold, and by the time a listener puts on a record the output stage may be running dangerously hot. For a deeper technical walkthrough of the biasing topology, see this class ab power amplifier tutorial which covers the mathematics of the VBE multiplier in detail.
Thermal design and heatsink calculation for class AB stages
Thermal management is inseparable from bias stability. Undersized heatsinks do not merely cause overheating — they actively destabilise quiescent current, which in turn degrades THD and risks transistor failure. The calculation is straightforward once you know three values: maximum power dissipation, thermal resistance of the transistor package, and ambient temperature.
Calculating worst-case power dissipation
For a class AB amplifier at maximum linear output, worst-case dissipation in the output stage peaks at approximately 20 % of the power supply capability — not at maximum output power, which is a counterintuitive but important point. For a ±35 V / 5 A supply (175 W capability), worst-case stage dissipation is approximately 35 W per rail, or 70 W total for a stereo pair. Add quiescent dissipation: 70 mA × 70 V = 4.9 W at idle. Total worst-case = approximately 75 W.
Heatsink thermal resistance requirement
Using the standard thermal resistance formula: θSA = (TJ(max) − TA) / PD − θJC − θCS. Assuming TJ(max) = 150 °C, TA = 40 °C (enclosed UK equipment cabinet in summer), θJC = 0.83 °C/W (MJL21193 datasheet), θCS = 0.5 °C/W (mica washer and thermal compound): θSA = (150 − 40) / 75 − 0.83 − 0.5 = 1.47 − 1.33 = 0.14 °C/W. This is a very low value, requiring a large aluminium extrusion heatsink or forced-air cooling. In practice, two MJL21193/21194 pairs per channel on a 0.5 °C/W heatsink — a common catalogue size of approximately 150 × 100 × 50 mm in black anodised aluminium — provides adequate margin when the supply rails are ±35 V or below. Real-world testing on a 2 × 80 W home audio amplifier confirms stable quiescent current drift of less than 5 mA over a two-hour continuous run at 50 % power with this heatsink geometry.
Measuring and verifying THD, IMD, and damping factor
Once bias is set and thermal design is validated, measurement confirms whether crossover distortion has truly been eliminated. Three metrics are essential: THD (total harmonic distortion), IMD (intermodulation distortion), and damping factor. Each reveals a different aspect of amplifier linearity and output stage quality.
THD measurement method and benchmarks
Apply a single-frequency sine wave (typically 1 kHz) at one-third of rated output power — the point at which residual crossover distortion is often most visible, as the signal amplitude is comparable to the bias-induced conduction overlap region. Use an audio analyser such as an Audio Precision APx515 or an open-source solution based on a calibrated PC sound card with RMAA software. Target THD values for a well-optimised class AB amplifier with moderate negative feedback are: below 0.05 % at 1 kHz / 1 W, and below 0.1 % at 20 kHz / rated power. If THD exceeds 0.2 % at 1 kHz in a design with global feedback, suspect insufficient quiescent current first.
IMD and damping factor benchmarks
IMD is measured using the SMPTE method (60 Hz + 7 kHz, 4:1 ratio). A properly biased class AB amplifier should exhibit IMD below 0.05 % at rated output. Damping factor — the ratio of rated load impedance to output impedance — should exceed 100 for a good solid state amplifier at 1 kHz into 8 Ω. Low damping factor (below 50) suggests excessive output resistance or a feedback compensation issue, not a biasing problem, but it is worth confirming after any design change. Of course, there are cases where a deliberately lower damping factor is chosen for tonal reasons in a musical instrument amplifier — a valid design decision, but one that should be conscious and documented.
Class AB vs class D: quantitative comparison and selection guide
The rise of GaN-based switching technology and high-integration audio SoCs has made class D amplifiers increasingly capable. By 2026, class D designs dominate portable, automotive, and budget home audio segments. Does that make class AB obsolete? Not at all — but it does make informed selection more important than ever.
| Parameter | Class AB amplifier | Class D amplifier |
|---|---|---|
| Typical efficiency (music signal) | 50–65 % | 85–93 % |
| THD+N (1 kHz, rated power) | 0.002–0.05 % | 0.003–0.1 % |
| Heatsink requirement (100 W/ch) | Large (0.4–0.8 °C/W) | Small (1.5–3 °C/W) |
| EMI/RFI generation | Very low | Moderate–high (requires filter) |
| BOM cost (100 W, discrete) | £25–£60 | £10–£30 |
| High-frequency linearity (>10 kHz) | Excellent | Varies (filter-dependent) |
| Perceived audio character | "Analogue warmth" (industry consensus) | Neutral to clinical |
| UK UKCA/EMC compliance complexity | Low | Moderate (output filter required) |
The data above — drawn from 2026 component datasheets and bench measurements on production-representative circuits — shows that class AB remains competitive on THD and EMI, while class D wins decisively on efficiency and heatsink size. Think of it like choosing between a naturally aspirated engine and a turbocharged one: both can deliver the same power figure, but the engineering trade-offs and the character of the result differ fundamentally. For UK hi-fi separates, musical instrument amplifiers, and reference monitor applications, AB class power amplifier topology continues to be the professional standard. For portable PA, automotive, and cost-driven consumer audio, class D is the rational choice. For a thorough technical treatment of the switching topology, see class ab operation explained.
UK compliance, efficiency standards, and procurement
Designing or selling an audio amplifier in the UK in 2026 requires attention to three regulatory frameworks: UKCA marking (replacing CE for Great Britain), the UK Ecodesign Regulations for energy-related products, and the UK EMC Regulations 2016 (SI 2016/1091). Crossover distortion reduction has a direct bearing on compliance because a poorly biased ab class power amplifier draws more supply current at idle, increasing standby power consumption — a parameter now scrutinised under Ecodesign.
UKCA and EMC requirements for class AB designs
Class AB amplifiers are inherently low-EMI devices compared to switching topologies, but they are not exempt from testing. A 100 W stereo amplifier with a linear power supply must demonstrate conducted emissions compliance to BS EN 55032 (equivalent to CISPR 32) and radiated emissions compliance across 30 MHz–1 GHz. Practically, a well-designed class AB stage with standard decoupling generates emissions well below EN 55032 Class B limits. Standby power must be below 0.5 W under the UK Ecodesign Regulations for audio equipment, achievable with a relay-switched standby circuit that disconnects the main transformer secondary. These requirements add modest cost — typically £200–£400 for a pre-compliance test at a UK EMC facility — but are straightforward to meet for AB class designs.
UK component suppliers and procurement
For UK-based builders and engineers, the primary component distributors stocking the transistor types and heatsink extrusions referenced in this guide are Farnell (Leeds), RS Components (Corby), and Mouser Electronics (UK warehouse in Oxfordshire). Complementary BJT output pairs such as MJL21193/MJL21194 and MJE15030/MJE15031 are available from all three with next-day delivery. Heatsink extrusions in black anodised aluminium — the 150 × 100 × 50 mm geometry referenced in Section 4 — are available from Ohmite and Fischer Elektronik via Farnell at approximately £4–£8 per unit. Thermal interface materials compliant with RoHS 2 (UK variant, SI 2012/3032) are essential for any UKCA-marked product; Electrolube HTC and Dow TC-5026 are both RoHS-compliant and widely stocked in the UK.
Frequently asked questions
Q: What is the ideal quiescent current for a class AB power amplifier?
A: For a discrete BJT output stage rated at 50–100 W, a quiescent current of 50–80 mA per complementary pair is typical. MOSFET stages generally require 100–150 mA. The correct value is the minimum that eliminates visible crossover distortion artefacts on an oscilloscope at low signal levels — not a fixed number applicable to all designs.
Q: Can negative feedback alone eliminate crossover distortion in a class AB amplifier?
A: Global negative feedback reduces crossover distortion significantly but cannot eliminate it entirely — particularly at high frequencies where loop gain falls. At 20 kHz, a typical 40 dB feedback factor provides only 20–25 dB of distortion reduction. Correct quiescent current bias must be established first; feedback is a supplement, not a substitute.
Q: How does temperature affect the bias in an AB class power amplifier?
A: BJT VBE decreases by approximately 2 mV per degree Celsius rise in junction temperature. Without thermal compensation, a 50 °C rise in output transistor temperature would increase quiescent current by roughly 40–60 %, risking thermal runaway. A VBE multiplier mounted on the heatsink provides automatic compensation by reducing drive voltage as temperature rises.
Q: Is class AB better than class D for hi-fi audio applications?
A: For high-power home audio amplifiers where heat dissipation is manageable, class AB typically offers lower THD at mid-power levels and better high-frequency linearity. Class D is preferable where efficiency, compact size, or low weight is critical. By 2026, the best class D designs match class AB on measured distortion, but class AB retains a perceived sonic advantage in the UK hi-fi market.
Q: What THD level should a well-designed class AB amplifier achieve?
A: A properly biased class AB amplifier with moderate global feedback should achieve THD below 0.05 % at 1 kHz across its rated output range. High-quality designs using optimised output stages and error correction can reach below 0.005 %. THD above 0.1 % at 1 kHz in a feedback-equipped design almost always indicates a biasing or thermal stability problem.
Summary
Reducing crossover distortion in an ab class power amplifier is fundamentally a biasing discipline, supported by sound thermal design and verified through systematic measurement. Set a stable quiescent current using a thermally coupled VBE multiplier, size the heatsink for worst-case dissipation rather than idle conditions, confirm performance with THD and IMD measurements, and — if you are building for the UK market — ensure UKCA compliance and standby power conformance from the outset. The class AB amplifier, for all the competition it now faces from class D, remains the topology of choice for any application where linear performance, low EMI, and sonic quality take priority over raw efficiency. Get the bias right, and it will reward you with THD figures that are genuinely inaudible on any loudspeaker system you are likely to encounter.
Keywords
Can’t find a solution?
Contact us for your customized solution!
Contact
Add: No.4, 7th Street, Changli Industrial Zone,Dongyong Town, Nansha District, Guangzhou,Guangdong,China
Tel: +86 20 31560084
Web: www.jl-proaudio.com
E-mail: info@jl-proaudio.com