An inverting averaging amplifier is an operational amplifier circuit that outputs the inverted mathematical mean of multiple input voltages applied to its inverting terminal through equal-value input resistors. While a standard summing amplifier simply adds voltages together—which quickly drives the op-amp into rail saturation as you add more inputs—the averaging variant scales the feedback network to divide the sum by the number of inputs. This fundamental change keeps the output signal within the linear operating range of the chip, making it indispensable for multi-sensor signal conditioning, battery management systems, and audio sub-mixing where signal headroom is critical.
The Core Math and Virtual Ground Mechanics
To understand the circuit, you have to look at the inverting input (pin 2 on a standard 8-pin DIP op-amp). Because the non-inverting input is tied to ground, negative feedback forces the inverting input to sit at a virtual ground (0V). It doesn't sink current; it just acts as a summing node.
When you apply voltages $V_1, V_2, ... V_n$ through identical input resistors ($R_{in}$), each generates a current based on Ohm's Law ($I = V/R_{in}$). These currents merge at the summing node and are forced through the feedback resistor ($R_f$) to the output. The output voltage equation is:
V_out = -(R_f / R_in) * (V_1 + V_2 + ... + V_n)
To convert this from a pure summer to an averager, you must force the circuit to divide the sum by the number of inputs ($n$). You achieve this by setting the feedback resistor to exactly 1/n of the input resistor value.
If you are building a 3-input averager, you need $R_f$ to be exactly 33.33% of $R_{in}$. Using standard 5% E24 resistors (e.g., 33kΩ and 10kΩ) introduces a ~1% gain error. For precision DC applications like thermocouple averaging, use 1% or 0.1% tolerance metal film resistors, or use a 10kΩ feedback resistor with three 30kΩ input resistors to hit the exact 1/3 ratio using standard E96 values.
Worked Numeric Example: 3-Channel LiFePO4 Temperature Monitoring
Let's look at a real-world scenario: averaging the voltage from three separate temperature sensors monitoring a 48V LiFePO4 battery bank. We will use a precision OPA2134 dual op-amp running on a ±12V dual supply.
- Input Resistors (R1, R2, R3): 30 kΩ each
- Feedback Resistor (Rf): 10 kΩ
- Input Voltages: $V_1$ = 1.2V, $V_2$ = 1.5V, $V_3$ = 1.8V (representing 25°C, 30°C, and 35°C)
Step 1: Calculate the input currents.
$I_1 = 1.2V / 30kΩ = 40 µA$
$I_2 = 1.5V / 30kΩ = 50 µA$
$I_3 = 1.8V / 30kΩ = 60 µA$
Step 2: Sum the currents at the virtual ground.
$I_{total} = 40 + 50 + 60 = 150 µA$
Step 3: Calculate the output voltage across Rf.
$V_{out} = - (150 µA * 10 kΩ) = -1.5V$
The mathematical average of 1.2V, 1.5V, and 1.8V is exactly 1.5V. The circuit outputs -1.5V, successfully averaging the inputs while inverting the polarity. If your downstream microcontroller ADC cannot read negative voltages, you must either add a second inverting unity-gain stage to flip the polarity back to positive, or bias the non-inverting pin to a mid-rail reference voltage.
Where You Meet This in Practice
You won't often see this circuit labeled explicitly as an "averager" in commercial schematics, but the topology is everywhere in analog front-ends.
- Sensor Array Noise Reduction: When measuring the temperature of a large motor stator or a battery pack, a single thermistor might catch a localized hot spot. Averaging three or four spatially distributed sensors gives the microcontroller a true bulk-temperature reading while inherently rejecting localized thermal noise.
- Audio Sub-Mixing: In analog mixing consoles, combining multiple microphone channels into a single subgroup bus uses this exact topology. Audio engineers use averaging (rather than pure summing) to prevent the combined audio peaks from clipping the op-amp's output stage. Chips like the TL072 or NE5532 are standard here for their low noise floor.
- DAC Ripple Smoothing: Some high-resolution measurement systems use multiple lower-resolution DACs and average their outputs through an op-amp to cancel out uncorrelated quantization noise and thermal noise, effectively increasing the bit-depth of the analog output.
Common Confusions: Summing vs. Averaging vs. Non-Inverting
It is easy to misidentify or misdesign this circuit if you confuse it with its close relatives. Here is how they differ on the bench.
| Topology | Output Formula | Primary Use Case | Main Pitfall |
|---|---|---|---|
| Inverting Summing | $-(V_1 + V_2 + ...)$ | Audio mixing, DAC current summing | Output clips/saturates as inputs increase |
| Inverting Averaging | $-\frac{1}{n}(V_1 + V_2 + ...)$ | Multi-sensor DC averaging | Requires negative supply or virtual ground for positive inputs |
| Non-Inverting Averaging | $+\frac{1}{n}(V_1 + V_2 + ...)$ | Single-supply battery monitoring | Input resistors form a passive network; requires high-impedance buffer |
A massive point of failure for hobbyists is trying to build an inverting averaging amplifier using a single-supply op-amp like the LM358 on a 5V rail, with the non-inverting pin tied to 0V. If your inputs are positive (e.g., 2V and 4V), the math demands a negative output (-3V). The LM358 cannot output below ground; it will simply rail out at ~20mV, and your microcontroller will read a flat zero. For single-supply DC averaging, either use a non-inverting passive averager followed by a voltage follower, or bias your op-amp's non-inverting pin to Vcc/2 (2.5V) to create an artificial virtual ground.
For deeper reading on op-amp summing and averaging topologies, the All About Circuits semiconductor textbook provides excellent foundational schematics, while Electronics Tutorials offers a rigorous breakdown of the virtual earth node mechanics that govern these feedback networks.
Inverting Averaging Amplifier FAQ
Can I use an inverting averaging amplifier for high-frequency RF signals?
No, this topology is generally restricted to DC and low-frequency audio/sensor applications (under 100 kHz). At RF frequencies, the parasitic capacitance of the feedback resistor and the op-amp's input capacitance create a low-pass filter, causing phase shift and potential oscillation. Furthermore, the virtual ground node becomes highly susceptible to stray capacitive coupling. For RF signal combining, engineers use passive resistive combiners, Wilkinson power dividers, or specialized RF mixer ICs designed for 50-ohm impedance matching.
Why does my averaging amplifier output clip at the negative rail?
If your output is pinned to the negative supply rail (e.g., -15V or 0V on a single supply), you have exceeded the op-amp's output voltage swing limit. This happens when the sum of your input currents multiplied by the feedback resistor demands a voltage higher than the chip can physically source. Check your input voltages: if you have four 5V inputs and a 10k feedback resistor with 10k input resistors, the circuit is trying to output -20V. If your op-amp is powered by ±15V, it will hard-clip at roughly -13.5V. Reduce your input voltages, increase the input resistor values, or switch to an op-amp with rail-to-rail output capabilities.
How do I add a fourth input to an existing 3-input averaging circuit?
You cannot simply solder a fourth input resistor to the summing node without changing the math. If you add a fourth 30kΩ input to a circuit designed with a 10kΩ feedback resistor, the scaling factor changes from 1/3 to 1/4, but your feedback resistor is still sized for 1/3. Your output will now be $V_{out} = -(10k/30k) * (V_1+V_2+V_3+V_4)$, which is no longer a true mathematical average. To fix this, you must replace the 10kΩ feedback resistor with a 7.5kΩ resistor (30kΩ / 4), or scale all four input resistors up to 40kΩ to keep the 10kΩ feedback resistor.
