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 of the inputs. Unlike the more common inverting topology, which relies on a virtual ground to isolate inputs, this configuration sums voltages passively at the high-impedance non-inverting pin before applying gain. What it changes in a real circuit is the ability to mix DC offsets or AC signals without introducing a 180-degree phase shift, which is critical in audio processing, single-supply sensor conditioning, and DAC level shifting. However, because it lacks a virtual ground, designers frequently run into hidden impedance issues that ruin their expected output.
The Core Math and a Worked Numeric Example
To understand the math, you have to split the circuit into two distinct stages: the passive input summer and the non-inverting gain stage. The inputs ($V_1$, $V_2$, etc.) are fed through individual resistors ($R_1$, $R_2$) into the non-inverting (+) pin of the op-amp. This creates a passive resistive averager. The op-amp then amplifies that summed voltage based on the feedback network ($R_f$ and $R_g$) connected to the inverting (-) pin.
The voltage at the non-inverting pin ($V_+$) is calculated using the superposition theorem. For two equal input resistors, it is simply the average of the input voltages. The final output is:
V_out = V_+ * (1 + (R_f / R_g))
Worked Numeric Example
Let's put real numbers on the bench. You are using a TL072 dual op-amp to mix two DC control voltages.
- Inputs: $V_1$ = 1.5V (Sensor baseline), $V_2$ = 2.5V (Manual offset)
- Input Resistors: $R_1$ = 10kΩ, $R_2$ = 10kΩ
- Feedback Network: $R_f$ = 10kΩ, $R_g$ = 10kΩ
Step 1: Find $V_+$
Because $R_1$ and $R_2$ are equal, $V_+$ is the average of the inputs: (1.5V + 2.5V) / 2 = 2.0V.
Step 2: Calculate Gain
Gain = 1 + ($R_f$ / $R_g$) = 1 + (10k / 10k) = 2.
Step 3: Calculate $V_{out}$
$V_{out}$ = 2.0V * 2 = 4.0V. The circuit successfully outputs the true sum (1.5V + 2.5V) while keeping the positive polarity intact.
Where You Meet This in Practice
You will rarely see a non inverting summing amp in professional audio mixing consoles; engineers prefer inverting summers there to prevent channel crosstalk. Instead, you meet this topology in embedded systems and instrumentation where phase preservation and single-supply operation matter.
- Single-Supply ADC Level Shifting: Microcontrollers like the ESP32-WROOM-32 have ADCs that only read 0V to 3.3V. If you have a bipolar audio signal swinging from -1.65V to +1.65V, you use a non-inverting summer to add a +1.65V DC offset, shifting the signal to a 0V–3.3V range without flipping the waveform upside down.
- Sensor Signal Conditioning: When a thermocouple or strain gauge outputs a tiny millivolt signal that rides on top of a common-mode voltage, a non-inverting configuration allows you to sum a trimming voltage to zero out the baseline offset before sending it to a microcontroller.
- Programmable Gain/Offset DACs: In function generators, summing a DC offset voltage with an AC waveform allows the user to shift the entire waveform up or down the Y-axis on an oscilloscope display.
Real-World Scenario Walkthrough: The Sensor Offset Failure
Theory assumes ideal voltage sources. The bench does not. Here is a scenario that burns hobbyists and junior engineers alike when building the ESP32 ADC level-shifter mentioned above.
The Setup
You need to shift a ±1V AC audio signal to a 0–2V range for an ESP32 ADC. You wire up an LM358 op-amp. $V_1$ is your AC audio (0V DC center). $V_2$ is your DC offset. You need $V_2$ to be exactly 1.0V. To get 1.0V from your 5V USB rail, you build a simple voltage divider using two 100kΩ resistors. You feed this 1.0V into the summing network via a 10kΩ resistor ($R_2$).
The Expected Numbers
With $V_1$ (DC) = 0V and $V_2$ = 1.0V, and equal 10kΩ input resistors, $V_+$ should be 0.5V. With a gain of 2, your DC output bias should sit perfectly at 1.0V, allowing the ±1V AC signal to swing from 0V to 2V.
The Outcome
You power it up. The ESP32 reads erratic, clipped values. You put your Fluke multimeter on the output pin and read a DC bias of 0.82V instead of 1.0V. The AC signal is riding too low, clipping the negative half-cycles against the 0V rail.
What Went Wrong
You forgot about source impedance. The 1.0V reference was generated by a high-impedance 100kΩ divider. When you connected the 10kΩ summing resistor ($R_2$) to it, the 10kΩ resistor acted in parallel with the bottom 100kΩ resistor of your divider. This loaded the divider, dragging the actual voltage at $V_2$ down from 1.0V to roughly 0.82V before the op-amp even saw it.
Non-Inverting vs. Inverting Summing Amps
People commonly confuse the non inverting summing amp with its inverting counterpart, assuming the math is just mirrored. It is not. The inverting summer relies on a virtual ground at the inverting input, which completely isolates the input channels from one another. The non-inverting summer lacks this virtual ground, meaning the input channels interact.
| Feature | Inverting Summing Amp | Non-Inverting Summing Amp |
|---|---|---|
| Phase Output | Inverted (180° shift) | Same phase as input (0° shift) |
| Input Isolation | High (Virtual ground prevents crosstalk) | Low (Inputs interact via passive network) |
| Input Impedance | Relatively low (Set by input resistors) | High (Depends on source and resistor network) |
| Math Complexity | Simple ($V_{out} = - (V_1 + V_2...)$) | Complex (Requires calculating $V_+$ node first) |
| Best Use Case | Audio mixing, DAC current summing | Single-supply level shifting, DC offset addition |
If you are mixing three separate audio tracks and want to ensure that adjusting the volume on Track 1 doesn't alter the voltage of Track 2, you must use an inverting summer. If you are simply adding a fixed DC bias to a single AC signal for a microcontroller ADC, the non-inverting topology saves you from having to invert the signal back later in software.
FAQ: Common Bench Questions
Can I use a non-inverting summing amp to add three or more voltages?
Yes, but the math gets messy quickly. For three inputs ($V_1$, $V_2$, $V_3$) with equal resistors $R$, the voltage at the non-inverting pin is the average: $(V_1 + V_2 + V_3) / 3$. To get the true sum at the output, your non-inverting gain stage must be set to exactly 3. If you change one input resistor to weight it differently, you must recalculate the entire parallel resistance network. For more than two inputs, most engineers switch to an inverting topology or use dedicated analog multiplier/mixer ICs.
Why is my output voltage clipping at the top?
Check your op-amp's power rails and output swing specifications. If you are running an LM358 on a single 5V supply, the output cannot physically reach 5V; it typically maxes out around 3.5V to 4V depending on the load. If your math says the output should be 4.5V, the op-amp will hard-clip. For rail-to-rail output on a 5V supply, swap the LM358 for a modern CMOS op-amp like the MCP6002 or OPA2333.
Do I need a dual power supply for this circuit?
No. The non-inverting summing amp is actually the preferred topology for single-supply designs. By summing a DC offset equal to half your supply voltage (e.g., 1.65V on a 3.3V rail), you bias the AC signal right in the middle of the op-amp's linear range, allowing it to swing both positive and negative relative to that new 'virtual ground' without requiring a negative voltage rail. See the Texas Instruments single-supply design guides for detailed biasing networks.
