Static linearity metrics: “Linearity error is the maximum deviation in step size from the ideal step size.” Evaluate this definition.

Introduction / Context:
Two commonly reported static accuracy metrics for data converters are Differential Nonlinearity (DNL) and Integral Nonlinearity (INL). Although both concern linearity, they quantify different aspects. Confusing them leads to misinterpretation of datasheets and incorrect performance predictions.

Given Data / Assumptions:

• DNL and INL are defined relative to an ideal DAC transfer curve.
• “Step size” refers to the output change when incrementing the code by one LSB.
• Ideal DAC step size is constant over the entire transfer characteristic.

Concept / Approach:
DNL is the maximum deviation of any actual code-to-code step size from one ideal LSB. In words: “maximum deviation in step size from ideal step size” fits DNL. By contrast, INL (often called linearity error) is the maximum deviation of the DAC’s actual transfer function from a straight reference line (endpoint or best-fit line). INL reflects accumulated error across codes, not individual step distortions.

Step-by-Step Solution:

Verification / Alternative check:
Consult converter literature: plots of DNL show step-by-step differences; INL plots show cumulative deviation versus an ideal straight line. A device may have low DNL yet show noticeable INL, and vice versa.

Why Other Options Are Wrong:

Common Pitfalls:
Assuming monotonicity from good INL alone; monotonicity relates more directly to DNL staying above −1 LSB.

Final Answer:
Incorrect (that describes DNL, not linearity error)