An inverting amplifier flips the phase of the input signal by 180 degrees while scaling its voltage, whereas a non-inverting amplifier scales the voltage while preserving the original phase, with both configurations using negative feedback to dictate precise closed-loop gain. While the underlying operational amplifier (op-amp) silicon remains the same, the external resistor network you choose fundamentally changes the circuit's input impedance, phase relationship, and noise profile.
The Core Difference: Phase, Impedance, and Common Confusion
The most critical distinction between these two topologies is how they interact with the signal source. An inverting amplifier presents an input impedance roughly equal to its input resistor ($R_{in}$). If you use a 10kΩ input resistor, your signal source sees a 10kΩ load. Conversely, a non-inverting amplifier routes the signal directly into the op-amp's high-impedance non-inverting pin, meaning the source sees the op-amp's native input impedance—often exceeding $10^{12}$ ohms for JFET or CMOS input stages.
To visualize the inverting configuration, picture a mechanical seesaw. The inverting input pin acts as the fulcrum (a 'virtual ground' held at 0V by the non-inverting pin being tied to ground). When you push the input side of the seesaw down (positive voltage), the feedback side must pull up (negative output voltage) to keep the fulcrum balanced. This mechanical leverage is why the output phase is inverted.
For a deeper mathematical breakdown of these feedback loops, the All About Circuits semiconductor textbook provides excellent foundational derivations of the virtual short concept.
Worked Numeric Example: Designing a 10x Gain Stage
Let's design a voltage amplification stage with a closed-loop gain of 10 (20 dB) using a low-noise OPA2134 audio op-amp. We will calculate the exact resistor values required for both topologies.
The Inverting Approach
The gain formula is $A_v = -R_f / R_{in}$. To achieve a magnitude of 10, the ratio of the feedback resistor ($R_f$) to the input resistor ($R_{in}$) must be 10:1.
- Choose $R_{in} = 10.0 k\Omega$ (Standard E96 1% value).
- Calculate $R_f = 10.0 k\Omega \times 10 = 100 k\Omega$ (Standard E96 1% value).
- Result: Gain is exactly -10. Input impedance is $10 k\Omega$.
The Non-Inverting Approach
The gain formula is $A_v = 1 + (R_f / R_g)$, where $R_g$ is the resistor tying the inverting pin to ground. To achieve a gain of 10, the ratio $R_f / R_g$ must be 9.
- Choose $R_g = 10.0 k\Omega$.
- Calculate $R_f = 10.0 k\Omega \times 9 = 90 k\Omega$. Since 90kΩ is not a standard 1% E96 value, we select the closest standard value: 90.9 kΩ.
- Result: Actual gain is $1 + (90.9 / 10.0) = 10.09$. Input impedance is $>10^{12} \Omega$ (FET input).
If your application demands exact gain in the non-inverting topology, you would place a 91kΩ resistor in parallel with a high-value trimmer potentiometer, or use a precision resistor network. For most audio and sensor applications, a 0.9% gain error is entirely negligible.
Where You Meet This in Practice
The choice between these configurations is rarely arbitrary; it is dictated by the source impedance and the signal processing requirements of the system.
| Application | Preferred Topology | Why It Wins Here |
|---|---|---|
| Electret Microphone Preamp | Non-Inverting | Microphones require high input impedance to prevent signal attenuation and low-frequency roll-off. |
| Photodiode Current-to-Voltage | Inverting (Transimpedance) | The virtual ground at the inverting pin keeps the photodiode bias voltage at exactly 0V, eliminating junction capacitance modulation. |
| Audio Mixer / Summing Console | Inverting | The virtual ground prevents crosstalk between multiple input channels; each source only 'sees' its own input resistor. |
| DAC Output Buffering | Non-Inverting | Provides unity gain (or slight boost) without loading the delicate resistor-ladder or delta-sigma output stage of the DAC. |
Additional practical design notes, including stability compensation for high-impedance non-inverting sensor buffers, are well documented in the Electronics Tutorials op-amp guides.
Component Selection and Real-World Gotchas
Textbook formulas assume an ideal op-amp with infinite bandwidth and zero offset. On the bench, silicon physics takes over. Here is what you must account for when moving from schematic to PCB.
Gain Bandwidth Product (GBW): An op-amp's GBW is a constant. If you select a classic TL072 with a GBW of 3 MHz and configure it for a non-inverting gain of 10, your usable bandwidth drops to 300 kHz. If you are amplifying a 100 kHz ultrasonic sensor signal, a gain of 10 will severely attenuate your output. You must select an op-amp with a GBW at least 10 to 20 times higher than your target signal frequency multiplied by your closed-loop gain.
The LM358 Crossover Distortion Trap: The LM358 is ubiquitous and cheap, but it uses a push-pull output stage that suffers from severe crossover distortion when the signal crosses 0V. If you are building an inverting audio amplifier with an LM358, you will hear harsh harmonic distortion. Always pull the output slightly to ground with a 10kΩ resistor to force the output transistors into Class-A operation, or upgrade to an audio-grade part like the OPA1612 or NE5532.
Input Bias Current and Resistor Noise: In the non-inverting configuration, the input bias current flows through the source impedance. If your source is a 1 MΩ piezo sensor and you use a bipolar op-amp like the NE5532 (which has high input bias current), the voltage drop across the source impedance will create a massive DC offset at the output. Always pair high-impedance non-inverting sources with JFET or CMOS op-amps (like the TL072 or LMC6482) that feature picoamp-level bias currents.
Frequently Asked Questions
Why does my non-inverting amplifier oscillate at high frequencies?
Oscillation in a non-inverting configuration is almost always caused by parasitic capacitance at the inverting input node interacting with the feedback resistors, creating an unintended low-pass filter in the feedback path. This adds phase shift, degrading the phase margin until the negative feedback becomes positive feedback. To fix this, place a small compensation capacitor (typically 10pF to 47pF) in parallel with your feedback resistor ($R_f$) to introduce a zero that cancels the parasitic pole. Additionally, ensure you have 100nF ceramic decoupling capacitors placed within 2mm of the op-amp's VCC and GND pins.
Can I use an inverting amplifier configuration to sum multiple audio signals?
Yes, this is the standard topology for an analog audio mixer, known as an inverting summing amplifier. Because the inverting input pin sits at a virtual ground (0V), multiple input signals can be fed through their own individual input resistors into this node without interacting with one another. The output voltage is the inverted sum of the inputs, scaled by the ratio of the single feedback resistor to each input resistor. This isolation prevents 'crosstalk' where adjusting the volume of a guitar channel would inadvertently alter the voltage level of a vocal channel.
What happens if I swap the inverting and non-inverting input pins on the physical op-amp chip?
If you accidentally route your input signal to the inverting pin and your feedback network to the non-inverting pin, you have created a positive feedback loop. Instead of amplifying the signal linearly, the op-amp will act as a Schmitt trigger or a latch. The output will instantly slam into the positive or negative supply rail (saturation) and stay there, completely destroying the linear amplification you intended. Always verify pinouts against the specific manufacturer's datasheet, as pin 2 and pin 3 assignments can occasionally vary between single, dual, and quad package variants.
