An integrator is an electronic circuit that produces an output voltage proportional to the accumulated sum (integral) of its input voltage over time. In a real circuit or installation, an integrator fundamentally changes signal geometry: it converts a square wave into a triangle wave, transforms a steady DC voltage into a linearly ramping voltage, and acts as a specialized low-pass filter that accumulates charge rather than merely attenuating high frequencies.

To visualize this, think of a bucket catching water from a hose. The water flow rate (gallons per minute) is your input signal. The total volume of water in the bucket at any given second is the integral of that flow over time. If you turn the hose to a constant trickle (a DC input), the water level (your output) rises in a straight, linear ramp. If you pulse the water on and off rapidly (a square wave), the water level zig-zags up and down (a triangle wave).

The Core Math and a Worked Numeric Example

The most common implementation in modern electronics is the active op-amp integrator. By placing a capacitor in the feedback loop and a resistor at the inverting input, the op-amp forces the capacitor to charge at a rate strictly dictated by the input voltage. The governing equation for the output voltage is:

Vout(t) = - (1 / RC) ∫ Vin(t) dt + Vinitial

Let us run a worked numeric example using a standard Texas Instruments TL072 JFET-input op-amp. We will design the circuit with the following real-world values:

  • Input Resistor (R): 10 kΩ
  • Feedback Capacitor (C): 100 nF (0.1 μF)
  • Input Voltage (Vin): +2V DC steady state
  • Initial State: Capacitor fully discharged (Vinitial = 0V)
  • Power Rails: ±12V DC

First, calculate the RC time constant (τ):
τ = 10,000 Ω × 0.0000001 F = 0.001 seconds (1 ms).

Because the input is a constant DC voltage, the integral simplifies to multiplication. The rate of change (slope) of the output ramp is:
Slope = -Vin / RC = -2V / 0.001s = -2,000 V/s (or -2V per millisecond).

If we apply this +2V input for exactly 3 milliseconds, the output voltage will be:
Vout = -2,000 V/s × 0.003 s = -6.0V.

The output ramps linearly downward at 2V/ms. If we leave the +2V input connected for 7ms, the output would theoretically hit -14V. However, because our op-amp is powered by ±12V rails, the output will saturate (flatline) at roughly -11V (accounting for typical op-amp headroom loss) after about 5.5ms. This saturation is a primary failure mode in poorly designed integrators.

Component Selection: Why Op-Amp Bias Current Causes Drift

The most common reason an integrator fails on the bench is output drift. Even when Vin is exactly 0V, the output slowly ramps up or down until it hits the supply rail. This happens because of input bias current (Ib). The internal transistors of the op-amp require a tiny amount of current to operate. This current is pulled through the feedback capacitor, slowly charging it like a phantom input signal.

According to Electronics Tutorials, selecting an op-amp with ultra-low input bias current is non-negotiable for precision integrators. Bipolar op-amps (like the classic LM741) have high bias currents in the nanoamp range, making them terrible integrators without constant reset switches. JFET and CMOS op-amps operate in the picoamp or femtoamp range, drastically reducing drift.

Op-Amp ModelInput TypeTypical IbSlew RateIntegrator Suitability
LM741Bipolar80 nA0.5 V/μsPoor: High drift, requires parallel reset FET.
TL072JFET5 pA13 V/μsGood: Standard audio and general-purpose wave shaping.
OPA140Precision JFET0.5 pA20 V/μsExcellent: Precision PID loops and instrumentation.
LMC6001CMOS0.1 pA (100 fA)1.5 V/μsUltimate: Ultra-low drift charge amplifiers and electrometers.
Bench Tip: If you must use a higher-bias op-amp, or if you are integrating very slow signals, place a large 'bleeder' resistor (e.g., 1 MΩ) in parallel with the feedback capacitor. This provides DC negative feedback to stabilize the operating point, though it turns the pure integrator into a low-pass filter at very low frequencies.

Where You Meet Integrators in Practice

You will rarely see a standalone integrator just sitting on a PCB for fun; they are usually hidden inside larger control or measurement systems.

  • PID Controllers: In industrial automation and drone flight controllers, the 'I' (Integral) term of a PID loop is implemented via an integrator circuit (or its digital equivalent). It accumulates past errors to eliminate steady-state offset, ensuring a motor eventually reaches its exact target speed.
  • Dual-Slope Analog-to-Digital Converters (ADCs): High-precision digital multimeters rely on integrating ADCs like the classic ICL7106. As detailed in technical literature on dual-slope conversion, the unknown input voltage is integrated for a fixed time, and then a known reference voltage is integrated in the opposite direction until the output hits zero. The time it takes to return to zero is directly proportional to the input voltage, offering incredible noise rejection.
  • Piezoelectric Charge Amplifiers: Piezo sensors (like accelerometers and knock sensors) output a burst of charge, not a steady voltage. An integrator circuit acts as a charge amplifier, accumulating the tiny pico-coulombs of charge onto the feedback capacitor to produce a readable, proportional voltage output.

Common Confusions: Integrator vs. Low-Pass Filter vs. Differentiator

Because an integrator uses a resistor and a capacitor, it is frequently confused with a simple passive RC low-pass filter. Furthermore, students often mix up the integrator with its mathematical opposite, the differentiator. Here is how to tell them apart on the bench.

Circuit TypeCore FunctionSquare Wave Input ResultFrequency Domain Behavior
Active IntegratorAccumulates input over timeLinear Triangle WaveGain drops at -20dB/decade; phase shift is exactly -90°.
Passive RC Low-PassAttenuates high frequenciesExponential 'Shark Fin' CurvesGain drops at -20dB/decade only after cutoff; phase shifts from 0° to -90°.
Active DifferentiatorOutputs rate of changeSharp Positive/Negative SpikesGain increases at +20dB/decade; phase shift is +90°.

The Golden Rule: A passive RC circuit only acts as a true mathematical integrator if the output is taken across the capacitor and the input frequency is significantly higher than the circuit's cutoff frequency (f >> fc). An active op-amp integrator, however, forces integration across a much wider, usable bandwidth by holding the inverting input at a virtual ground.

Frequently Asked Questions

Why does my integrator output drift to the supply rail when the input is grounded?
This is caused by the op-amp's input bias current and input offset voltage. Both act as tiny DC inputs that the integrator faithfully accumulates over time. To fix this, use a JFET or CMOS op-amp (like the OPA140), ensure your non-inverting input is tied to a clean, low-impedance ground, and add a large feedback resistor (10x to 100x the input resistor) in parallel with the capacitor to provide DC stabilization.

Do I need a dual (split) power supply for an op-amp integrator?
Generally, yes. Because the output must ramp both positive and negative in response to AC inputs, a dual supply (e.g., ±12V or ±5V) is standard. If you must use a single supply, you have to bias the non-inverting input to a mid-rail virtual ground (e.g., 2.5V on a 5V system) and AC-couple your input signal, which severely limits your low-frequency integration capabilities.