A non-inverting amplifier is an operational amplifier (op-amp) circuit configuration where the input signal is applied to the non-inverting (+) terminal, producing an output voltage that is scaled up but maintains the exact same phase and polarity as the input. What it changes in a real circuit is the signal’s voltage amplitude without flipping the waveform or loading down the source, thanks to the op-amp’s naturally high input impedance. Unlike the inverting configuration, which forces the input signal through a resistor to a virtual ground, the non-inverting topology preserves the original signal’s phase relationship, making it the default choice for high-impedance sensor buffering and audio preamplification.
The Core Formula and Standard Resistor Ratios
The closed-loop voltage gain ($A_v$) of a non-inverting amplifier is determined entirely by the external feedback resistor network. The input signal enters the high-impedance non-inverting pin, while a fraction of the output voltage is fed back to the inverting (-) pin via a voltage divider. The op-amp’s internal high open-loop gain forces the voltage at the inverting pin to match the non-inverting pin (the virtual short concept).
The governing equation is:
$A_v = 1 + (R_f / R_i)$
Where $R_f$ is the feedback resistor (from output to inverting input) and $R_i$ is the ground resistor (from inverting input to ground).
On the bench, the biggest hurdle isn’t the math—it’s finding standard resistor values that yield your exact target gain without resorting to custom trimming. The table below provides exact 1% tolerance E96 series resistor pairs for common gain targets. Using 1% resistors is mandatory for precision DC and audio applications; standard 5% E24 carbon film resistors will introduce unacceptable gain errors and channel-to-channel mismatch in stereo audio builds.
| Target Gain ($A_v$) | Required Ratio ($R_f / R_i$) | Standard $R_i$ (1% E96) | Standard $R_f$ (1% E96) | Actual Achieved Gain | Gain Error |
|---|---|---|---|---|---|
| 2x (+6 dB) | 1.000 | 10.0 kΩ | 10.0 kΩ | 2.000 | 0.00% |
| 5x (+14 dB) | 4.000 | 11.0 kΩ | 44.2 kΩ | 5.018 | +0.36% |
| 10x (+20 dB) | 9.000 | 12.1 kΩ | 110 kΩ | 10.09 | +0.90% |
| 15x (+23.5 dB) | 14.00 | 10.0 kΩ | 140 kΩ | 15.00 | 0.00% |
| 50x (+34 dB) | 49.00 | 2.00 kΩ | 97.6 kΩ | 49.80 | -0.40% |
| 100x (+40 dB) | 99.00 | 1.10 kΩ | 110 kΩ | 101.0 | +1.00% |
Worked Numeric Example: Designing a 15x Sensor Preamplifier
Let’s design a real-world front-end for a high-impedance piezoelectric vibration sensor. The sensor outputs a maximum of 0.2V peak, and we need to amplify this to 3.0V peak to maximize the resolution of a 3.3V ADC on an ESP32 microcontroller. Our target non inverting amp gain is exactly 15.
Step 1: Select the Op-Amp
We need a modern, rail-to-rail input/output (RRIO) op-amp that operates cleanly on a single 3.3V supply. The Microchip MCP6002 is a bench staple for this. It features a 1 MHz Gain-Bandwidth Product (GBWP) and rail-to-rail swing, meaning our 3.0V peak output won’t clip against the 3.3V supply rail.
Step 2: Calculate the Feedback Network
Using the formula $15 = 1 + (R_f / R_i)$, we need a resistor ratio of exactly 14:1. Looking at our E96 table above, we select $R_i = 10.0\text{ k}\Omega$ and $R_f = 140\text{ k}\Omega$. Both are standard 1% values.
Step 3: Verify Bandwidth and Slew Rate
The MCP6002 has a GBWP of 1 MHz. At a closed-loop gain of 15, the small-signal bandwidth is $1,000,000 / 15 = 66.6\text{ kHz}$. Since our vibration sensor only outputs frequencies up to 5 kHz, we have more than enough bandwidth.
Next, we check the slew rate, which is 0.6 V/µs for the MCP6002. The maximum frequency before slew-rate limiting distorts our 3.0V peak sine wave is calculated as:
$f_{max} = \frac{Slew Rate}{2 \pi \times V_{peak}} = \frac{0.6 \times 10^6}{2 \pi \times 3.0} \approx 31.8\text{ kHz}$
Since 31.8 kHz is well above our 5 kHz signal, the amplifier will reproduce the waveform without triangular distortion.
Step 4: Physical Layout and Decoupling
On the breadboard or PCB, place a 100 nF (0.1 µF) X7R MLCC ceramic capacitor physically adjacent to the VCC and GND pins of the MCP6002. Single-supply op-amps are highly susceptible to high-frequency power rail noise coupling into the non-inverting input via parasitic capacitance. According to Analog Devices’ Linear Circuit Design Handbook, failing to decouple the supply pins within 2 mm of the IC package is the leading cause of unexplained high-frequency oscillation in high-gain non-inverting circuits.
Where You Meet This in Practice (and Common Confusions)
You will encounter the non-inverting topology almost anywhere a weak, high-impedance signal must interface with a low-impedance load. Common applications include:
- Audio DI Boxes and Preamplifiers: Buffering electric guitar pickups (which have source impedances of 100 kΩ to 1 MΩ) without loading them down and killing the high-frequency harmonics.
- Electrochemical Sensors: pH probes and ion-selective electrodes output millivolt signals with source impedances in the hundreds of megaohms. The non-inverting configuration’s ultra-high input impedance prevents the sensor from collapsing under load.
- Current Shunt Monitoring: Amplifying the small differential voltage across a low-side current sense resistor before feeding it to a microcontroller ADC.
What People Commonly Confuse It With
The most frequent error among junior engineers and hobbyists is confusing the non-inverting amplifier with the voltage follower (unity gain buffer). A voltage follower is simply a special case of the non-inverting amplifier where $R_f = 0\Omega$ (a direct wire from output to inverting pin) and $R_i = \infty$ (inverting pin grounded directly or left floating if bias currents allow). The gain is exactly 1. While the voltage follower provides impedance transformation, it provides zero voltage amplification.
The second major confusion is assuming the inverting amplifier is interchangeable. In an inverting configuration, the input impedance is strictly equal to the input resistor ($Z_{in} = R_i$). If you use an inverting amp with a 10 kΩ input resistor to buffer a 500 kΩ guitar pickup, you create a massive voltage divider that attenuates the signal before it even gets amplified. The non-inverting topology avoids this because its input impedance is determined by the op-amp’s internal common-mode impedance (often $>10^{12} \Omega$ for FET-input op-amps like the TL072), effectively isolating the source from the load.
Real-World Limits: Input Common-Mode and DC Errors
While the ideal op-amp equations assume infinite performance, real-world silicon imposes strict boundaries that dictate component selection. Texas Instruments’ SLOA011A application report details how ignoring these limits leads to catastrophic circuit failure.
If you are using a classic LM358 on a single 5V supply, the inputs cannot safely read voltages near the positive rail. The LM358’s ICMR typically maxes out at $V_{CC} - 1.5V$ (3.5V). If your non-inverting input receives a 4.0V signal, the internal PNP differential pair starves for current, and the output will unpredictably phase-invert or latch up. For single-supply designs where the signal approaches the positive rail, you must specify a Rail-to-Rail Input (RRI) op-amp.
Input Bias Current and Offset Voltage:
In high-gain configurations (e.g., 100x), the DC errors of the op-amp are amplified right alongside your signal. An input offset voltage ($V_{os}$) of just 2 mV, when multiplied by a non inverting amp gain of 100, results in a 200 mV DC error at the output. If you are amplifying a thermocouple signal where 200 mV represents a 5°C temperature error, your measurement is useless. For high-gain DC applications, swap standard bipolar op-amps for precision chopper-stabilized or zero-drift op-amps like the TI OPA2188, which boast a maximum $V_{os}$ of just 25 µV.
FAQ: Edge Cases on the Bench
Can I use a non-inverting amp for a gain of less than 1 (attenuation)?
No. The mathematical floor of the non-inverting formula $1 + (R_f/R_i)$ is 1 (when $R_f = 0$). You cannot achieve a gain of 0.5 with this topology. If you need attenuation followed by buffering, use a passive resistor voltage divider followed by a unity-gain voltage follower.
Why is my high-gain circuit oscillating at 5 MHz?
Stray capacitance across the feedback resistor ($R_f$) creates an unintended low-pass filter in the feedback path, introducing phase shift that destroys the op-amp’s phase margin. To fix this, add a small compensation capacitor (typically 10 pF to 47 pF) in parallel with $R_f$ to intentionally roll off the high-frequency gain, or use a physically smaller surface-mount resistor to reduce parasitic pad capacitance.
