A non-inverting summing amplifier is an operational amplifier circuit that combines multiple input voltages into a single output signal while maintaining the original phase polarity. In a real circuit, it changes how you merge sensor readings or inject DC bias voltages without flipping the waveform upside down, though it introduces input crosstalk that designers must manage. Makers commonly confuse it with the inverting summing amplifier (which uses a virtual ground to perfectly isolate inputs) or mistakenly assume it outputs the raw mathematical sum without requiring a specific feedback gain network to compensate for the input voltage divider.
The Core Math and a Worked Numeric Example
To understand this circuit, you have to break it into two distinct stages: the passive summing network at the non-inverting (+) input, and the active non-inverting gain stage. Unlike the inverting configuration, the non-inverting pin does not sit at a virtual ground. Instead, the input resistors form a passive voltage averager. The op-amp then buffers and amplifies that averaged voltage.
The voltage at the non-inverting pin ($V_+$) is calculated using Millman’s theorem. Assuming equal input resistors ($R_1 = R_2 = R_3 = R$) and no additional pull-down resistor to ground at the non-inverting node, the formula simplifies to the average of the input voltages:
$V_+ = (V_1 + V_2 + V_3) / N$ (where N is the number of inputs).
Because the op-amp only sees the average voltage, the feedback network must apply a gain equal to N to output the true mathematical sum. The standard non-inverting gain equation is $A = 1 + (R_f / R_g)$.
Let’s design a 3-input summing amplifier. We want the output to be the exact sum of three DC sensor voltages: V1 = 1.5V, V2 = 2.0V, and V3 = 2.5V. The target output is 6.0V.
1. Input Network: We choose 10kΩ for R1, R2, and R3.
$V_+ = (1.5V + 2.0V + 2.5V) / 3 = 6.0V / 3 = 2.0V$.
2. Gain Stage: To get from 2.0V to 6.0V, we need a gain of 3.
$A = 1 + (R_f / R_g) = 3$, which means $R_f / R_g = 2$.
If we set $R_g = 10k\Omega$, then $R_f$ must be 20k\Omega.
Result: The op-amp multiplies the 2.0V at the non-inverting pin by 3, yielding exactly 6.0V at the output. The signals are summed, and the phase remains positive.
Where You Meet This in Practice
While the inverting summing amplifier is the default choice for audio mixing due to its virtual ground isolation, the non-inverting summing amplifier solves specific problems in instrumentation, power supply design, and single-supply embedded systems.
| Feature | Non-Inverting Summing Amplifier | Inverting Summing Amplifier |
|---|---|---|
| Phase Output | Same as input (0° shift) | Inverted (180° shift) |
| Input Isolation | Poor (inputs interact via resistor network) | Excellent (virtual ground isolates inputs) |
| Input Impedance | Relatively low (set by input resistors) | Low (set by input resistors) |
| Common Use Case | DC bias injection, sensor averaging | Audio mixing, active DAC summing |
Sensor Averaging: In thermal management systems, you might place three NTC thermistors at different points on a heat sink. By feeding their buffered voltage dividers into a non-inverting summing amplifier configured for averaging (gain of 1, achieved by setting $R_f = 0$ and $R_g = \infty$), you get a single DC voltage representing the average temperature without needing a microcontroller ADC for each sensor.
DC Bias Injection: When working with single-supply microcontrollers like the ESP32 or Arduino, the ADC can only read positive voltages (0V to 3.3V). If you are measuring an AC signal (like a current transformer output swinging ±1V), you must inject a 1.65V DC offset. A non-inverting summing configuration allows you to sum the AC signal with a precise 1.65V reference from a voltage divider, shifting the entire waveform into the readable positive range.
Design Pitfalls and Component Selection
The biggest trap when building this circuit on the bench is ignoring the input crosstalk and the op-amp’s output voltage swing limitations. Because there is no virtual ground at the non-inverting pin, a signal applied to V1 will partially bleed back out through the V2 and V3 input resistors. If your source impedances are high, this crosstalk will corrupt your readings. Always buffer individual signals with unity-gain op-amps before feeding them into the summing network if source isolation is critical.
Component selection dictates whether your circuit will actually work on a single 5V or 3.3V supply. According to Texas Instruments’ op-amp design guidelines, standard legacy op-amps cannot swing their output all the way to the positive supply rail.
The ubiquitous LM358 is cheap and works fine for low-speed DC summing on a dual supply (e.g., ±12V). However, if you power an LM358 from a single 5V supply and attempt to output a summed voltage of 4.5V, the circuit will fail. The LM358’s output stage requires about 1.5V of headroom below the positive rail. Your 4.5V signal will hard-clip at roughly 3.5V. For single-supply 3.3V or 5V non-inverting summing circuits, you must use a true Rail-to-Rail Input/Output (RRIO) op-amp like the Microchip MCP6002 or the TI TLV2372.
For high-fidelity audio or high-speed sensor summing, the semiconductor characteristics of JFET or CMOS input stages are preferred. The TL072 offers vastly lower input bias currents and noise, making it the standard for audio-rate summing, provided you are using a dual-rail power supply (e.g., ±9V or ±15V).
Frequently Asked Questions
Can I use a non-inverting summing amplifier for audio mixing?
You can, but it is generally a bad idea for professional or high-fidelity audio. Because the non-inverting pin lacks a virtual ground, the audio signal from Channel 1 will bleed into the Channel 2 and Channel 3 input lines through the summing resistors. This crosstalk causes phase cancellation and channel interference. Audio mixers almost exclusively use the inverting summing amplifier topology, where the virtual ground at the inverting pin holds the summing node at 0V, perfectly isolating the input channels from one another.
Why is my non-inverting summing amplifier output lower than expected?
If your math is correct but the output voltage is stubbornly low, you are likely hitting the op-amp’s output swing limit or loading the output too heavily. First, check your power supply rails. If you are using a standard LM358 or LM741 on a single supply, the output cannot reach the positive rail (it typically maxes out at VCC - 1.5V). Second, check your load resistance. If your feedback resistors are too low (e.g., 100Ω), the op-amp may not be able to source enough current to maintain the voltage. Keep feedback and input resistors in the 10kΩ to 100kΩ range to minimize output current draw.
How do I calculate resistor values for an unequal weighted sum?
If you need to apply different weights to your inputs (for example, building a simple digital-to-analog converter where V1 is the MSB and V3 is the LSB), you abandon equal resistor values. You use the full Millman’s theorem equation: $V_+ = (V_1/R_1 + V_2/R_2 + V_3/R_3) / (1/R_1 + 1/R_2 + 1/R_3)$. To give V1 twice the weight of V2, you simply make R1 half the value of R2. After calculating the resulting $V_+$, you adjust the $R_f$ and $R_g$ feedback network to scale the final output to your desired maximum voltage range.
