The linearity of an amplifier is the measure of how precisely its output signal replicates the exact shape and proportional amplitude of its input signal without adding distortion or altering the waveform. In a real circuit, this single characteristic dictates your Total Harmonic Distortion (THD), Intermodulation Distortion (IMD), and spectral purity, ultimately determining whether your audio sounds muddy, your RF transmitter bleeds into adjacent channels, or your precision sensor readings drift under load.
The Transfer Curve and the Math of Distortion
Every amplifier has a transfer function—a graph of output voltage versus input voltage. In a perfectly linear amplifier, this graph is a straight line passing through the origin (or the bias point). The slope of that line is your gain. When the transfer curve bends, flattens, or kinks, the amplifier introduces non-linear distortion.
Let's look at a concrete numeric example to see how this manifests on the bench. Suppose you are designing an audio preamp stage. You feed a 100 mV RMS pure 1 kHz sine wave into an amplifier configured for a gain of 20 (26 dB).
- Ideal Linear Output: 2,000 mV (2.0 V) RMS at exactly 1 kHz.
- Real-World Non-Linear Output (e.g., a basic LM358 pushed near its rails): The fundamental 1 kHz output drops slightly to 1,950 mV RMS due to gain compression. Worse, the bending transfer curve generates harmonics: 15 mV RMS at 2 kHz (2nd harmonic) and 8 mV RMS at 3 kHz (3rd harmonic).
We calculate the Total Harmonic Distortion (THD) using the root-sum-square of the harmonics divided by the fundamental:
THD = (√(15² + 8²) / 1950) × 100
THD = (√(225 + 64) / 1950) × 100
THD = (17 / 1950) × 100 = 0.87%
A THD of 0.87% is easily audible as harshness in a hi-fi system. By swapping that generic op-amp for a high-linearity part like the TI OPA1612, the THD drops to 0.00003%, rendering the harmonics entirely unmeasurable on standard bench gear.
The Great Confusion: Linearity vs. Gain vs. Efficiency
On the jobsite and in hobbyist forums, three terms get tangled constantly. Here is how to separate them:
| Metric | What It Actually Means | The Trap |
|---|---|---|
| Gain | The simple multiplier (Vout / Vin). Usually expressed in dB. | Assuming high gain means high fidelity. You can have 100 dB of gain with terrible distortion. |
| Efficiency | The ratio of RF/Audio power delivered to the load versus DC power drawn from the supply. | Chasing efficiency in Class D or Class C amps while ignoring the massive non-linear switching noise they generate. |
| Linearity | The straightness of the transfer curve; absence of harmonic generation. | Assuming negative feedback fixes everything. Feedback improves linearity, but at the cost of bandwidth and phase margin. |
Where You Meet Amplifier Linearity in Practice
You don't just 'meet' linearity in textbooks; it dictates pass/fail on real-world projects across three main domains:
1. High-Fidelity Audio and Studio Gear
In audio, non-linearity causes intermodulation distortion (IMD). If you feed a 1 kHz and a 10 kHz tone into a non-linear amp, the circuit will multiply them, creating phantom sum and difference frequencies (9 kHz and 11 kHz) that were never in the original recording. This is what makes cheap Bluetooth speakers sound 'muddy' or 'fatiguing' at high volumes.
2. RF Transmitters and Software Defined Radio (SDR)
In RF, linearity is measured by the 1 dB Compression Point (P1dB) and the Third-Order Intercept Point (IP3). If your RF power amplifier is pushed past its linear region, the signal 'splatters' into adjacent frequency bands. This spectral regrowth will cause you to fail FCC/CE emissions testing and can result in the FCC shutting down your transmitter for jamming neighboring channels. For deep theory on RF distortion, Analog Devices' MT-043 tutorial is the definitive reference.
3. Precision Instrumentation and Sensor Conditioning
When amplifying a 5 mV signal from a strain gauge or thermocouple, any non-linearity in the instrumentation amplifier translates directly into measurement error. A 0.1% linearity error on a 1000-pound load cell means your scale could be off by a full pound, which is unacceptable for industrial batching systems.
Decision Tree: Picking the Right Topology and Part Number
Stop guessing based on generic 'op-amp' searches. Use this decision matrix to terminate your design phase with a concrete, proven part number.
| Your Primary Constraint | Required Topology / Architecture | Concrete Part Recommendation | Key Datasheet Spec to Verify |
|---|---|---|---|
| Ultra-Low Noise DC / Strain Gauges | Chopper-Stabilized (Zero-Drift) Instrumentation Amp | TI ADA4522 or ADI AD8421 | Offset voltage drift (< 0.05 µV/°C) |
| Hi-Fi Audio Line Drivers / DAC Buffers | Bipolar Input, High Open-Loop Gain Class AB | TI OPA1612 | THD+N at 1 kHz (< -120 dB) |
| High-Speed ADC Drivers (100MHz+) | Fully Differential Current Feedback | TI THS4551 | Slew Rate (> 300 V/µs) and HD3 |
| RF / IF Amplification (up to 3 GHz) | MMIC Darlington or pHEMT (Class A biased) | Mini-Circuits GALI-74+ | IP3 (> +40 dBm) and P1dB |
| Battery-Powered IoT (Audio/Voice) | Class D Digital Input with Integrated DSP | TI TAS5805M | Efficiency (> 90%) and I2C EQ tuning |
Frequently Asked Questions
Does adding more negative feedback always improve linearity?
Yes, but with a strict penalty. Negative feedback forces the output to correct its own errors, effectively dividing the inherent distortion by the loop gain. However, as you increase feedback, you reduce the amplifier's bandwidth and degrade its phase margin. Push it too far, and your highly linear amplifier turns into a high-frequency oscillator. Always check the datasheet's 'Gain vs. Phase' plot before closing the loop.
Why does my amplifier's linearity degrade at high frequencies?
As frequency increases, the amplifier's internal open-loop gain rolls off (typically at -20 dB/decade). Because negative feedback relies on excess open-loop gain to correct non-linear errors, a drop in open-loop gain means a drop in loop gain. Less loop gain means less error correction, resulting in higher THD at 20 kHz compared to 1 kHz.
Can layout and bypass capacitors affect linearity?
Absolutely. Linearity isn't just about the silicon; it's about the power delivery. If your bypass capacitors are too small or placed too far from the IC pins, the power supply rails will 'sag' microscopically during high-current output transients. This power supply rejection ratio (PSRR) limitation introduces low-frequency modulation into the high-frequency signal, destroying your linearity. Always use a parallel combination of a 100 nF X7R ceramic and a 10 µF tantalum/polymer capacitor placed within 2 mm of the VCC/GND pins.
How do I measure the 1 dB compression point on the bench?
Set up a signal generator and a spectrum analyzer (or a high-precision RF power meter). Feed a single CW tone into the amplifier. Increase the input power in 1 dB steps and record the output power. Initially, output power will rise exactly 1 dB for every 1 dB of input. The exact point where the output power falls 1 dB short of the expected theoretical value is your P1dB. Never exceed this point if signal fidelity is required.
