The Non-Inverting Summing Amplifier: Why It’s Tricky (And How to Do It Right)

If you need to combine multiple analog signals without flipping their phase, you are looking for an op amp non inverting summing amplifier. The direct answer to building one is to use a passive resistor voltage-divider network feeding the non-inverting (+) input, followed by a standard non-inverting gain stage. However, unlike the classic inverting summing amplifier—where the virtual ground at the inverting pin perfectly isolates each input from the others—the non-inverting topology suffers from a notorious bench headache: input crosstalk.

Because there is no virtual ground at the non-inverting pin, signals can travel backward through the input resistors into adjacent channels. To build this circuit reliably, you must either buffer each input with a unity-gain follower first, or use precise superposition math and switching jacks to manage floating inputs.

Pinout and Symbol Orientation

For this guide, we will use the ubiquitous 8-pin DIP dual op-amp package. When the notch is facing left, the pinout is:

  • Pin 1: Output A (Vout)
  • Pin 2: Inverting Input A (V-)
  • Pin 3: Non-Inverting Input A (V+) — This is our summing node
  • Pin 4: V- (Negative Supply or GND)
  • Pin 5: Non-Inverting Input B
  • Pin 6: Inverting Input B
  • Pin 7: Output B
  • Pin 8: V+ (Positive Supply)

In our circuit, the input signals feed into Pin 3 through individual resistors. The feedback network connects between Pin 1 (Output) and Pin 2 (Inverting Input), while Pin 2 also connects to ground (or a bias voltage) through a gain-setting resistor.

Safe Default Part Numbers and Operating Regions

Selecting the right IC depends entirely on your supply voltage and signal type. Audio signals require low noise and dual supplies (or careful single-supply biasing), while DC sensor summing can tolerate cheaper, single-supply parts. Here are the safe defaults for the workbench:

Part NumberSupply Voltage RangeInput Bias CurrentSlew RateBest Application
LM3583V to 32V (Single) / ±1.5V to ±16V (Dual)20 nA (typ)0.3 V/µsDC sensor summing, low-cost single-supply
TL072±5V to ±18V (Dual only)5 pA (typ)13 V/µsAudio mixing, high-impedance sources
MCP60021.8V to 6.0V (Single/Dual)1 pA (typ)0.6 V/µsBattery-powered 3.3V microcontroller interfaces
OPA2134±2.5V to ±18V (Dual)5 pA (typ)20 V/µsPremium audio, studio-grade summing
Biasing for Single Supply: If you are using an LM358 on a single 9V battery to sum AC audio signals, you cannot ground the non-inverting network. You must create a virtual ground at 4.5V (using a buffered voltage divider) and tie the bottom of your summing network and your feedback ground resistor to that 4.5V rail, not 0V. Otherwise, the negative half of your AC waveform will be hard-clipped at 0V.

The Application Circuit: 2-Channel DC Sensor Summer

Let’s build a practical circuit: summing two 0–2V DC temperature sensor outputs (V1 and V2) to produce a single 0–4V output. We will use an LM358 powered by a single 5V supply (since 4V is well within the LM358’s output swing limits on a 5V rail).

Component Values and Resistor Math

We need three resistors for the summing node at Pin 3: R1 (from V1), R2 (from V2), and R3 (from Pin 3 to Ground). To keep the math clean, we set R1 = R2 = R3 = 10 kΩ.

Using the superposition theorem, the voltage at the non-inverting pin (V+) is:

V+ = V1 × [ (R2 || R3) / (R1 + (R2 || R3)) ] + V2 × [ (R1 || R3) / (R2 + (R1 || R3)) ]

Since all resistors are 10 kΩ, the parallel combinations (10k || 10k) equal 5 kΩ. The equation simplifies to:

V+ = V1 × (5k / 15k) + V2 × (5k / 15k) = (V1 + V2) / 3

Now, we need the output to be V1 + V2. Since the voltage at Pin 3 is one-third of our target, we need a non-inverting gain of exactly 3. The gain formula is Gain = 1 + (Rf / Ri).

To get a gain of 3, Rf / Ri must equal 2. We will use Ri = 10 kΩ (from Pin 2 to Ground) and Rf = 20 kΩ (from Pin 1 to Pin 2). If V1 is 1.5V and V2 is 2.0V, V+ becomes 1.166V. Multiplied by the gain of 3, the output at Pin 1 is exactly 3.5V.

Bench Scenario: The Crosstalk Trap

Theory is clean; the workbench is not. Here is a real-world scenario demonstrating why the non-inverting summer fails if you ignore source impedance.

The Setup: You are building a 2-channel guitar and microphone audio mixer using a TL072 on a ±12V dual supply. You use the 10k/10k/10k input network described above, followed by a gain stage. You plug the guitar into V1, but leave the V2 microphone jack unplugged.

The Numbers: The guitar outputs a 1V peak-to-peak (Vpp) signal. Because the mic jack is unplugged, the V2 input is floating (infinite impedance). Effectively, R2 is removed from the circuit.

The Outcome: Instead of just hearing the guitar, you notice two things. First, the guitar volume drops by 25%. Second, if you touch the exposed metal of the unplugged mic cable, you hear a massive 60Hz hum blast through the speaker.

What Went Wrong: In an inverting summer, the virtual ground at the summing node would have absorbed the guitar signal, preventing it from traveling anywhere else. In our non-inverting circuit, with V2 floating, the guitar signal travels from V1, through R1, into the Pin 3 node, and then backward through R2 into the unplugged mic jack. The floating mic cable acts as an antenna, picking up mains hum, which is then summed directly into your audio path. Furthermore, with R2 out of the parallel network, the voltage divider ratio changes, altering your gain and dropping the volume.

The Fix: Never leave inputs floating in a non-inverting summer. Use switching TRS jacks that mechanically short the signal line to ground when the plug is removed. Alternatively, place a unity-gain buffer op-amp on every input before the summing resistors to provide high input impedance and low output impedance, completely isolating the channels.

Testing and Troubleshooting with a Multimeter

When your summing amplifier output is pegged to the supply rail, dead, or noisy, use your digital multimeter (DMM) to isolate the fault. Follow these numbered steps to diagnose the IC and surrounding network.

  1. Verify Power Rails: Set your DMM to DC Volts. Measure between Pin 8 (V+) and Pin 4 (V-). For a ±12V supply, you should read exactly 24V across these pins. If you read 0V, check your breadboard power rails. If you read a few millivolts, your IC is likely shorted internally and pulling the supply down.
  2. Check the Bias / Ground Reference: Measure the voltage at the bottom of your Ri resistor (the one tied to ground or virtual ground). It should read 0.00V (or exactly half your single supply, e.g., 4.50V). If this node is floating, your op-amp has no reference and the output will immediately rail.
  3. Measure the Non-Inverting Node (Pin 3): With inputs grounded, Pin 3 should read 0V (or your bias voltage). If you apply a known DC voltage to V1, Pin 3 should rise according to your voltage divider math. If Pin 3 reads identically to V1, your R1 resistor is likely shorted or missing.
  4. Check the Feedback Network: Power down the circuit. Switch your DMM to resistance (Ohms). Measure across Rf (Pin 1 to Pin 2). It should read your resistor value (e.g., 20 kΩ). If it reads OL (open), your feedback loop is broken, and the op-amp is running open-loop, which guarantees the output will slam into the positive or negative rail.
  5. Measure DC Offset at Output (Pin 1): Power back on. With all inputs grounded, measure Pin 1 relative to system ground. A healthy TL072 will read less than 5mV. An LM358 might read up to 20mV. If you read several volts, your input bias currents are mismatched, or the IC is damaged.

Inverting vs. Non-Inverting Summing: When to Switch

If you are fighting crosstalk and complex math, you might wonder why you shouldn't just use an inverting summing amplifier. Here is a practical decision framework for choosing between the two topologies.

Choose the Inverting Topology When:

  • You are summing multiple AC audio channels and cannot guarantee all jacks will be grounded when unplugged.
  • You want simple math: the gain for each input is strictly -Rf / Rin, completely independent of the other inputs.
  • Phase inversion is acceptable, or you have a spare op-amp half to act as a unity-gain phase inverter at the output.

Choose the Non-Inverting Topology When:

  • Phase preservation is critical (e.g., summing DC sensor offsets where a negative voltage would break the microcontroller's ADC).
  • You are summing signals from high-impedance sources that cannot tolerate the low input impedance of an inverting summer.
  • You are willing to add input buffers to eliminate the crosstalk penalty.

For deeper reading on op-amp fundamentals and stability, refer to the Texas Instruments LM358 product documentation for single-supply limitations, and the Summing Amplifier theory overview for the foundational superposition mathematics used in these networks. Always verify your specific op-amp's output swing limits in its datasheet; an LM358 on a 5V supply will typically max out around 3.5V to 3.8V, not the full 5V rail, which will clip your summed signal if you aren't accounting for headroom.