error
To show or contain an error or fault.
error: when a circuit or system behaves outside spec
In electrical engineering, to error means a circuit, component, or system deviates from its intended performance. This is not the same as a simple fault or failure. An error is a measurable departure from the correct value, state, or function, often small enough that the system continues to operate but produces unreliable results. A digital circuit may error when its output voltage drifts outside tolerance. An analog sensor may error when temperature causes its output to shift 0.5% from nominal. These are working systems that are giving wrong answers.
Errors accumulate in different ways depending on circuit type. In precision instrumentation, errors compound: a 1% error in a voltage divider stage may cascade through gain stages, producing 10% error at the output. In digital systems, a single bit error in a communications line can corrupt an entire data packet. Timing errors in synchronous circuits cause setup or hold violations that make flip-flops metastable. Thermal errors shift component values; a resistor with a temperature coefficient of 100 ppm per degree Celsius will error significantly over a 50-degree range.
Real designs build in error margins. A logic family rated for 0 to 5 volts typically defines a high input as anything above 2.0 volts, and a low input as anything below 0.8 volts. This hysteresis band allows the circuit to error by up to 2 volts on a high signal and still be recognized correctly. Without such margins, noise and component tolerances would make the system unreliable. Power supplies are specified to error no more than plus-or-minus 5% from their nominal output under load changes and line voltage variations.
Errors also occur in measurement and control loops. A feedback system may error if the sensor has lag, hysteresis, or nonlinearity. A thermostat reading 2 degrees high will error in its control response, overheating the space. A strain gauge with nonlinear output will error when used beyond its linear range. Compensation networks and calibration procedures are designed specifically to reduce these errors.
The electrical trades distinguish error from failure: a failed component produces no output or catastrophic output; an erroring component produces plausible but incorrect output. This distinction matters in quality control and diagnostics. A shorted transformer fails immediately. A transformer with turns-ratio error of 2% may run for years delivering slightly wrong voltage, causing subtle problems downstream that are hard to trace. This is why manufacturing tolerances, aging factors, and operating condition ranges are specified for every component in critical circuits.