Chapter 15 — Audio Applications and Amplifiers

Chapter Overview (click to expand)

Audio electronics brings together nearly every circuit concept from the course — gain stages, filters, impedance matching, and noise analysis — into a practical system that faithfully amplifies sound from microphone to loudspeaker. This chapter traces the audio signal chain from preamplifier through power amplifier, covering the quality metrics (SNR, THD, power factor) that distinguish professional-grade designs from mediocre ones. **Key Takeaways** 1. The audio signal chain stages each small-signal gain stage, tone control filter, and power stage in a sequence designed to maximize signal quality while minimizing noise and distortion. 2. Preamplifiers provide voltage gain with very low noise, while power amplifiers deliver current to low-impedance speaker loads — the two stages have fundamentally different design priorities. 3. Total harmonic distortion (THD) and signal-to-noise ratio (SNR) are the primary specifications that quantify audio amplifier fidelity, and circuit topology choices directly control both.

15.1 Introduction

When you press play on your favorite song, a series of carefully engineered circuits invisibly transforms microscopic electrical voltages into the sound pressure waves that reach your ears. A dynamic microphone generates a signal as small as 1 millivolt. By the time that signal drives a loudspeaker, it may have been amplified a million-fold in power — yet the music you hear should sound exactly like the original performance, only louder. That requirement of faithful amplification drives every design decision in audio electronics.

Audio engineering is where nearly every topic from this course converges. Filters shape the frequency response. Op-amp gain stages boost signals. Fourier analysis reveals what distortion actually adds to a waveform. Impedance matching ensures maximum power transfer. Noise analysis determines how quiet a system can be. The skills you have built across Chapters 1–14 are the tools; this chapter shows you the job they are built for.

This chapter develops two parallel threads: the signal chain — the functional blocks audio passes through — and the quality metrics — the specifications that tell you whether each block is doing its job. Together, they give you the language, the mathematics, and the intuition to design audio circuits that both measure well and sound great.

15.2 The Audio Signal Chain

An audio amplifier is any circuit that increases the power of an audio-frequency signal. In practice, no single stage handles the entire voltage and power gain required — the job is divided into specialized stages, each optimized for its particular role.

The Standard Signal Chain

Stage	Function	Typical Output Level
Source (mic, instrument, playback)	Convert sound or stored audio to electrical signal	−60 dBV to 0 dBV
Preamplifier	Raise low-level signal to line level	0 dBV (1 V RMS)
Signal processing (EQ, effects, mixing)	Shape frequency content and blend sources	0 dBV
Power amplifier	Deliver high current to drive speakers	10 W to 1000 W+
Loudspeaker	Convert electrical power to acoustic energy	Sound pressure level

The concept of line level (nominally 0 dBV or about 1 V RMS in consumer equipment; +4 dBu or 1.23 V RMS in professional equipment) is the universal handoff point between stages. A signal at line level can drive signal-processing equipment and power amplifiers directly. The preamplifier's job is to reach that level; the power amplifier's job is to go far beyond it in power — but not in voltage.

Typical Source Signal Levels

Source	Level (dBV)	Approximate Voltage	Required Preamp Gain
Dynamic microphone	−60 dBV	1 mV	40–60 dB
Condenser microphone	−40 dBV	10 mV	20–40 dB
Electric guitar (passive)	−20 dBV	100 mV	10–20 dB
CD player / line out	0 dBV	1 V	0 dB (already at line level)

Notice that a dynamic microphone generates a signal 1000 times smaller in voltage than a CD player. That enormous range — 60 dB — is why preamplifier design is one of the most demanding disciplines in analog electronics.

Diagram: Audio Signal Chain — Interactive Walkthrough

The Noise Cascade Principle

Noise added by the first stage is amplified by every stage that follows. A 2 μV noise voltage at the preamplifier input, after 60 dB of gain, becomes 2 mV at the output. The same 2 μV added by the last stage arrives at the output as 2 μV. This is why the preamplifier's noise performance dominates the entire system's SNR — and why preamplifier design demands the lowest-noise components and topologies.

15.3 Preamplifiers

A preamplifier (preamp) takes a weak, low-impedance or high-impedance source signal and raises it to line level with as little added noise and distortion as possible. It is the most noise-sensitive stage in the audio chain.

Key Preamplifier Requirements

• High input impedance: Avoids loading the source and robbing it of signal. Microphone preamps typically present 1 kΩ–10 kΩ; instrument preamps 1 MΩ or higher.
• Low noise figure: The noise figure (NF) in dB describes how much noise the stage adds above the theoretical thermal noise minimum. A professional mic preamp achieves NF < 1 dB.
• Variable gain: Microphone sensitivity varies widely; a gain range of 20–60 dB (±20 dB adjustable) is standard.
• Flat frequency response: ±0.5 dB from 20 Hz to 20 kHz is a typical specification.
• Low output impedance: Drives cables and downstream equipment without signal loss.

Preamplifier Topologies

Most modern preamplifiers are built around low-noise op-amps in non-inverting or instrumentation amplifier configurations. The non-inverting topology is preferred because its input impedance is set by the op-amp's differential input resistance (typically very high) rather than by feedback resistors.

**Non-inverting preamplifier gain:** \[A_V = 1 + \frac{R_f}{R_i}\] For a gain of 100 (40 dB): choose \(R_i = 1\ \text{k}\Omega\), \(R_f = 99\ \text{k}\Omega\).

The instrumentation amplifier (in-amp) is the preferred topology for balanced microphone inputs. It provides extremely high CMRR (common-mode rejection ratio), rejecting the noise and hum induced equally on both conductors of a balanced cable while amplifying only the differential signal.

Recommended low-noise op-amps for preamplifier stages include the NE5532 (general purpose, very low noise), OPA2134 (audio grade, extremely low THD), and the INA217 (dedicated microphone preamplifier IC with variable gain and 130 dB CMRR).

15.4 Power Amplifiers

A power amplifier takes a line-level signal (around 1 V RMS) and delivers high current to drive a loudspeaker load — typically 4 Ω or 8 Ω — at voltage swings of tens of volts. The fundamental challenge is doing so efficiently while minimizing distortion. The two goals are in direct conflict, and different amplifier classes resolve that tension in different ways.

Class A

In a Class A amplifier, the output transistor conducts for the full 360° of the input waveform cycle. The bias current is set high enough that even at peak output swing, the transistor never cuts off.

• Advantage: The output transistor always operates in its linear region → extremely low crossover distortion → cleanest sound.
• Disadvantage: The transistor dissipates maximum power when no signal is present. Efficiency is at best ~25% theoretically; practical Class A amplifiers often achieve 15–20%. All the wasted power becomes heat.
• Typical use: High-end audiophile amplifiers and headphone amplifiers where sound quality justifies the thermal penalty.

Class B

In a Class B amplifier, two complementary transistors share the load: one handles positive half-cycles, the other handles negative half-cycles. Each transistor conducts for exactly 180° and is completely off for the other half.

• Advantage: Much higher efficiency (~70% theoretical) because transistors dissipate power only while conducting.
• Disadvantage: Near the zero crossing, neither transistor is conducting — producing a dead zone called crossover distortion. This harsh glitch is readily audible and makes pure Class B unacceptable for audio.

Class AB

Class AB is the dominant choice in real-world power amplifiers. Both transistors are biased with a small quiescent current (typically a few milliamps) so that each conducts slightly beyond its 180° half-cycle. This eliminates the dead zone around the zero crossing while keeping efficiency high.

• Efficiency: ~50–65% in practice — a reasonable compromise.
• Distortion: Very low, approaching Class A performance with good negative-feedback design.
• Heat: Moderate; heat sinking is required but manageable.
• Typical use: Home stereo receivers, car audio amplifiers, professional PA amplifiers.

Class D (Switching Amplifier)

Class D amplifiers do not linearly amplify the signal. Instead, they convert the audio signal to a high-frequency pulse-width modulated (PWM) waveform, switch transistors fully on or fully off at that PWM frequency (typically 300 kHz–1 MHz), and then pass the output through a low-pass filter to recover the audio.

• Efficiency: >90% — transistors spend almost no time in the linear (power-dissipating) region.
• Heat: Minimal; Class D amplifiers can often be fanless even at hundreds of watts.
• Complexity: Requires precise PWM generation and a high-quality output filter.
• Distortion: Depends heavily on implementation; modern Class D designs achieve THD comparable to Class AB.
• Typical use: Subwoofer amplifiers, portable speakers (Bluetooth/battery-powered), automotive amplifiers, hearing aids.

Power Amplifier Class Comparison

Class	Conduction Angle	Efficiency	Distortion	Heat	Typical Application
A	360°	~25%	Lowest	Very high	High-end audiophile
B	180°	~70%	High (crossover)	Moderate	Rarely used alone
AB	180°–360°	~60%	Low	Moderate	Most audio amplifiers
D	Switching	>90%	Low (modern)	Low	Portable, subwoofers

Diagram: Audio Amplifier Chain — Stages and Gain Budgeting

15.5 Signal-to-Noise Ratio and Noise Floor

Every real amplifier adds some noise to the signal. The signal-to-noise ratio (SNR) quantifies how much louder the desired signal is compared to the unwanted noise — the higher the SNR, the cleaner the audio.

\[SNR = 20\log_{10}\!\left(\frac{V_{signal}}{V_{noise}}\right) \text{ dB}\]

The noise floor is the RMS noise voltage present at a system's output when no signal is applied. Expressed in dBV (decibels relative to 1 V RMS), it tells you the system's quietest possible output level.

SNR Quality Benchmarks

SNR	Quality Category	Typical Application
>110 dB	Excellent	Studio-grade converters and preamps
90–110 dB	Professional	Mixing consoles, audiophile equipment
70–90 dB	High consumer	Home stereo receivers, quality headphone amps
<70 dB	Consumer / budget	Inexpensive portable devices

Worked Example — SNR Calculation

A microphone preamplifier outputs a maximum signal of 2 V RMS. Its noise floor is measured at 20 μV RMS. What is the SNR?

\[SNR = 20\log_{10}\!\left(\frac{2\ \text{V}}{20 \times 10^{-6}\ \text{V}}\right) = 20\log_{10}(100{,}000) = 20 \times 5 = 100\ \text{dB}\]

A 100 dB SNR is professional quality. At maximum volume, the noise is 100,000 times smaller in voltage than the signal — completely inaudible.

Dynamic Range

Dynamic range is closely related to SNR — it is the ratio between the maximum undistorted output level and the noise floor, measured in dB. For an amplifier that clips at 10 V RMS with a 10 μV noise floor:

\[DR = 20\log_{10}\!\left(\frac{10\ \text{V}}{10 \times 10^{-6}\ \text{V}}\right) = 20\log_{10}(10^6) = 120\ \text{dB}\]

▷ MicroSim — SNR and Noise Floor

Set signal level and noise floor to see SNR. Add harmonics to simulate distortion. The green annotation shows the SNR gap between signal and noise.

15.6 Thermal Noise

Thermal noise (also called Johnson-Nyquist noise) is the fundamental noise generated by the random thermal motion of electrons in any resistor at any temperature above absolute zero. It is not a design flaw — it is a consequence of thermodynamics. No resistor, no matter how well manufactured, can produce less than this noise.

\[V_{noise} = \sqrt{4 k_B T R \,\Delta f}\] where: - \(k_B = 1.38 \times 10^{-23}\ \text{J/K}\) (Boltzmann's constant) - \(T\) = absolute temperature in Kelvin - \(R\) = resistance in ohms - \(\Delta f\) = noise bandwidth in Hz

Three factors control thermal noise:

• Temperature: Lowering \(T\) reduces noise. This is why some radio-astronomy receivers are cooled to 4 K with liquid helium.
• Resistance: Lower resistance means less noise. Use the smallest practical resistor values in signal-path positions.
• Bandwidth: A narrower bandwidth filter captures less noise. Audio systems naturally benefit from limiting bandwidth to 20 kHz.

Worked Calculation — Thermal Noise of a 10 kΩ Source

A microphone has a source resistance of 10 kΩ. At room temperature (300 K) over the full audio bandwidth (20 Hz to 20 kHz, Δf = 20,000 Hz), what is the minimum possible noise voltage?

**Step 1:** Substitute values. \[V_{noise} = \sqrt{4 \times (1.38 \times 10^{-23}) \times 300 \times 10{,}000 \times 20{,}000}\] **Step 2:** Compute the product under the radical. \[= \sqrt{4 \times 1.38 \times 10^{-23} \times 6 \times 10^{7}}\] \[= \sqrt{4 \times 8.28 \times 10^{-16}}\] \[= \sqrt{3.31 \times 10^{-15}}\] **Step 3:** Take the square root. \[V_{noise} \approx 1.82 \times 10^{-6}\ \text{V} = 1.82\ \mu\text{V RMS}\]

This 1.82 μV is an irreducible physical lower bound. Any preamplifier connected to a 10 kΩ microphone will see at least this noise voltage at its input terminal, regardless of how perfectly the amplifier itself is designed.

Noise Sources Add as RMS

Multiple independent noise sources (thermal noise, op-amp voltage noise, op-amp current noise) add in quadrature — root sum of squares — not linearly. If the resistor contributes 1.82 μV and the op-amp contributes 2 μV, the total noise is \(\sqrt{1.82^2 + 2^2} = \sqrt{3.31 + 4} = \sqrt{7.31} \approx 2.7\ \mu\text{V}\), not 3.82 μV.

15.7 Total Harmonic Distortion (THD)

Harmonic distortion occurs when a non-linearity in an amplifier causes the output to contain frequency components that are integer multiples of the input frequency. If you feed a pure 1 kHz sine wave into a distorting amplifier, the output will contain 1 kHz (the fundamental), 2 kHz (2nd harmonic), 3 kHz (3rd harmonic), and so on.

The origin is the non-linear term in the amplifier's transfer function. A linear amplifier has \(V_{out} = A \cdot V_{in}\). A real amplifier has additional terms:

\[V_{out} = a_1 V_{in} + a_2 V_{in}^2 + a_3 V_{in}^3 + \cdots\] The \(a_2\) term produces 2nd-harmonic distortion; the \(a_3\) term produces 3rd-harmonic distortion.

Even vs. Odd Harmonics

Harmonics have distinct perceptual characters that explain why some distorting amplifiers sound more pleasant than others:

• Even harmonics (2nd, 4th, 6th…): Correspond to octave relationships (2nd harmonic = one octave above fundamental). These are musically consonant and are perceived as adding warmth or richness. Vacuum tube amplifiers naturally emphasize even harmonics.
• Odd harmonics (3rd, 5th, 7th…): These intervals are more dissonant. Strong odd harmonics produce the harsh, edgy character associated with transistor clipping or square-wave-like distortion.

The THD Formula

Total Harmonic Distortion (THD) expresses the combined harmonic content as a percentage of the fundamental amplitude:

\[THD = \frac{\sqrt{V_2^2 + V_3^2 + V_4^2 + \cdots}}{V_1} \times 100\%\] where \(V_1\) is the RMS amplitude of the fundamental and \(V_n\) is the RMS amplitude of the \(n\)-th harmonic.

Worked Example — THD Calculation

A spectrum analyzer measures the following at the output of a power amplifier (1 kHz input, 10 V RMS output):

Harmonic	Frequency	Measured RMS Voltage
Fundamental (\(V_1\))	1 kHz	10.000 V
2nd (\(V_2\))	2 kHz	15.0 mV
3rd (\(V_3\))	3 kHz	8.0 mV
4th (\(V_4\))	4 kHz	3.0 mV

\[THD = \frac{\sqrt{(0.015)^2 + (0.008)^2 + (0.003)^2}}{10} \times 100\%\] \[= \frac{\sqrt{2.25 \times 10^{-4} + 6.4 \times 10^{-5} + 9 \times 10^{-6}}}{10} \times 100\%\] \[= \frac{\sqrt{2.98 \times 10^{-4}}}{10} \times 100\% = \frac{0.01726}{10} \times 100\% \approx 0.173\%\]

THD Quality Table

THD	Audio Quality	Assessment
< 0.01%	Excellent	Below audibility threshold for most listeners
0.01% – 0.1%	High quality	Professional and audiophile equipment
0.1% – 1%	Acceptable	Good consumer equipment
1% – 5%	Noticeable	Low-cost consumer gear; audible coloration
> 5%	Significant	Obvious distortion; objectionable in most contexts

The 0.173% result from the example would be considered acceptable for a mid-tier consumer amplifier. Reducing it to <0.1% would require either operating at lower output power (more headroom), adding more negative feedback, or selecting transistors with better linearity characteristics.

15.8 Intermodulation Distortion (IMD)

Real audio contains many simultaneous frequencies — instruments, voices, overtones. When two or more frequencies pass through a non-linear system together, the non-linearity mixes them, producing new frequencies that are sums and differences of the originals and their harmonics. This is intermodulation distortion (IMD).

How IMD Products Are Generated

Consider two input tones at frequencies \(f_1\) and \(f_2\). The \(a_2\) (quadratic) non-linearity produces second-order IMD products:

\[\text{2nd-order IMD: } f_1 + f_2 \quad \text{and} \quad |f_1 - f_2|\] \[\text{3rd-order IMD: } 2f_1 - f_2 \quad \text{and} \quad 2f_2 - f_1 \quad \text{(most problematic)}\]

The 3rd-order products \(2f_1 - f_2\) and \(2f_2 - f_1\) fall close to the original frequencies when \(f_1\) and \(f_2\) are close together. Unlike harmonics, these products are not harmonically related to the original tones. They produce dissonant, musically unrelated frequencies that listeners find objectionable — even at levels lower than equivalent THD.

Example — IMD Product Frequencies

Input tones: \(f_1 = 1000\ \text{Hz}\), \(f_2 = 1500\ \text{Hz}\).

Product Type	Calculation	Resulting Frequency
2nd-order sum	\(f_1 + f_2\)	2500 Hz
2nd-order difference	\(f_2 - f_1\)	500 Hz
3rd-order (lower)	\(2f_1 - f_2\)	500 Hz (coincides)
3rd-order (upper)	\(2f_2 - f_1\)	2000 Hz

None of these new frequencies (500 Hz, 2000 Hz, 2500 Hz) is a harmonic of 1 kHz or 1.5 kHz. They are musically alien interlopers.

Standard IMD Test Methods

• SMPTE/DIN method: \(f_1 = 60\ \text{Hz}\) (4:1 amplitude ratio), \(f_2 = 7\ \text{kHz}\). Widely used for amplifier qualification.
• CCIF method (difference tone): Two closely spaced high-frequency tones, e.g., 19 kHz and 20 kHz. The 1 kHz difference product is easily measured.

15.9 Clipping

Clipping is the most severe form of audio distortion. It occurs when an input signal demands an output voltage or current that exceeds what the amplifier's power supply can deliver. The waveform is literally clipped — its peaks are flattened.

Hard Clipping

Hard clipping occurs abruptly when a transistor saturates or a rail is reached. The top and bottom of the waveform are replaced by flat horizontal sections. In the frequency domain, hard clipping of a sine wave generates very strong odd harmonics (3rd, 5th, 7th…) — a severely hard-clipped sine wave approaches a square wave, which is rich in odd harmonics. THD under hard clipping can easily exceed 10%.

Hard clipping is the characteristic sound of overdriven transistor guitar amplifiers and is the cause of tweeter damage in PA systems driven into overdrive — the high harmonic content represents substantial high-frequency power that tweeters are not designed to handle.

Soft Clipping

Soft clipping is a more gradual onset of limiting. Instead of the sharp corner of transistor saturation, the gain compresses progressively as the signal approaches the limit. Vacuum tubes exhibit this behavior naturally because their transconductance curves roll off smoothly rather than cutting off abruptly.

Soft clipping emphasizes even harmonics (2nd, 4th) more than odd harmonics, which is part of the reason overdriven tube amplifiers are considered more musical sounding than overdriven transistor amplifiers. Guitar effects pedals deliberately introduce soft clipping to simulate tube overdrive.

Headroom

Headroom is the margin between the normal operating signal level and the clipping threshold, measured in dB. Professional equipment maintains at least 20 dB of headroom above nominal operating level to handle transient peaks without clipping. Digital audio systems express this as the margin below 0 dBFS (digital full scale).

\[\text{Headroom} = 20\log_{10}\!\left(\frac{V_{clip}}{V_{nominal}}\right) \text{ dB}\]

Diagram: Distortion Types — Harmonic Distortion, Clipping, and Nonlinear Effects

15.10 Practical Design: Grounding, Shielding, and Balanced Connections

A circuit that measures beautifully on a bench can produce hum, buzz, and noise when installed in a real system. Electromagnetic interference (EMI), ground loops, and capacitively coupled noise are the practical enemies of audio quality. Professional audio design employs three systematic countermeasures.

Star Grounding

In a star ground topology, all circuit ground connections return to a single physical point rather than being daisy-chained along a shared ground bus. When multiple ground currents flow through a shared impedance, the voltage drop across that impedance appears as a noise signal — a phenomenon called a ground loop.

Star grounding ensures that high-current return paths (power supply, output stage) share no impedance with sensitive low-level signal returns (preamplifier input, A/D converter reference). The classic symptom of a ground loop is 50 Hz or 60 Hz hum in the audio — the AC mains frequency.

Shielded Cables

A shielded cable surrounds the signal conductor with a grounded conductive braid or foil. Radio-frequency interference, hum from AC power wiring, and noise from other equipment capacitively couples onto the shield rather than onto the signal conductor. The shield is connected to ground at one end only (typically the source end) to avoid creating a ground loop through the cable shield.

Balanced Connections

Balanced connections transmit the audio signal on two conductors simultaneously, with the signal on one conductor being the polarity-inverted copy of the signal on the other (differential transmission). Any noise induced along the cable — mains hum, RF interference — appears equally on both conductors (common-mode noise). The balanced receiver, implemented as a differential amplifier or instrumentation amplifier with high CMRR, subtracts the two signals, canceling the common-mode noise while doubling the desired differential signal.

\[\text{Noise rejection (dB)} = \text{CMRR of receiver (dB)}\] A balanced receiver with CMRR = 80 dB rejects 80 dB of common-mode noise — a noise voltage reduction factor of 10,000.

Balanced connections use XLR connectors (3-pin: pin 1 = shield, pin 2 = hot, pin 3 = cold) in professional audio and are the reason studio equipment connected by long cable runs produces clean audio while consumer equipment on the same circuit would buzz and hum.

Component Selection for Low Noise

Component	Recommendation	Reason
Op-amp (preamp stage)	NE5532, OPA2134, AD797	Ultra-low voltage and current noise
Resistors	1% metal-film, lowest practical value	Lower value → less thermal noise
Capacitors (signal path)	Film (polypropylene or polyester)	Low dielectric absorption, low noise
Capacitors (bypass)	Ceramic X7R in parallel with film	Wide-frequency power supply decoupling
PCB layout	Short signal traces, guard rings around sensitive nodes	Minimize parasitic coupling

15.11 Chapter Summary

This chapter developed the theory and practice of audio amplifier systems:

1. The audio signal chain connects source → preamplifier → processing → power amplifier → speaker. Line level (0 dBV / +4 dBu) is the universal handoff between stages.

2. Preamplifiers must achieve high input impedance, very low noise figure, and 40–60 dB of gain to bring microphone-level signals to line level. Their noise dominates system SNR.

3. Power amplifier classes trade efficiency against distortion: Class A (~25%) for maximum linearity, Class AB (~60%) for the best practical balance, Class D (>90%) for maximum efficiency.

4. SNR (\(20\log_{10}(V_s/V_n)\)) measures signal quality. Professional audio targets SNR > 90 dB.

5. Thermal noise (\(V_n = \sqrt{4k_BTR\Delta f}\)) is the irreducible physical noise floor. A 10 kΩ source at 300 K over 20 kHz bandwidth generates 1.82 μV RMS.

6. THD (\(\sqrt{V_2^2+V_3^2+\cdots}/V_1 \times 100\%\)) measures harmonic distortion. Below 0.01% is excellent; above 1% is noticeable.

7. IMD generates sum and difference frequencies that are not harmonically related to the input — more perceptually objectionable than equivalent THD.

8. Hard clipping (abrupt) generates strong odd harmonics and can damage tweeters. Soft clipping (gradual) generates predominantly even harmonics and is perceived as less harsh.

9. Star grounding, shielded cables, and balanced connections are the three professional techniques for controlling noise and interference in real installations.

✓ Ready to test your knowledge? Go to Chapter 15 Quiz →