integration time

Integration time is the window during which an instrument accumulates and averages a signal before outputting a reading. In practice, this means the sensor collects charge, current, or voltage data over a defined span, then sums or averages it. Longer integration windows reduce noise by spreading random fluctuations across more samples, leaving the true signal stronger relative to the jitter. A voltmeter set to integrate for 1 second will reject noise frequencies above roughly 1 Hz; one integrating for 10 milliseconds will pass through faster variations but leave more noise behind.

Different instruments use integration time differently. Lock-in amplifiers employ it as a filter time constant, typically ranging from microseconds to seconds depending on the modulation frequency and required sensitivity. Charge-coupled devices (CCDs) and complementary metal-oxide-semiconductor (CMOS) image sensors use it as the exposure time: photons strike the detector for exactly this long before the pixel values are read out. Digital multimeters often let you select 'fast' (around 100 ms) or 'slow' (1 second or more) modes, which are really integration time settings in disguise.

The penalty for long integration times is latency and reduced measurement rate. If your integration window is one second, you can only acquire one new data point per second. In real-time control loops or fast transient measurement, this becomes a serious limitation. Conversely, short integration times mean noisier readings and lower sensitivity. The choice always involves trade-off: grab data quickly with more noise, or wait longer for cleaner numbers.

Matched filtering and bandwidth

The name 'integration' comes from the mathematical act of summing or integrating the signal over time. Formally, longer integration acts as a low-pass filter whose cutoff frequency is inversely proportional to the integration window. Integrate for T seconds, and you attenuate frequencies above roughly 1/T hertz. This is why mains hum (50 or 60 Hz) vanishes with a 1-second integration on a bench instrument: the averaging window swallows the oscillation whole.

In sensor design, integration time also determines dynamic range behavior. A CCD pixel that integrates too long blooms: accumulated charge overflows into adjacent pixels. Too short, and you collect too little light in dim conditions, wasting sensor area for noise. Proper integration time selection is one of the first steps in designing a measurement system for a specific signal bandwidth and noise floor, because all downstream choices about gain, filtering, and sampling rate build on it.