Calculus I · Unit 2B · lesson
Manufacturing Tolerances and Error Propagation
Learning objectives
Use differentials and relative error to estimate manufacturing uncertainty.
How Measurement Error Moves Through a Formula
Before the formulas
In Manufacturing Tolerances and Error Propagation, the derivative converts a formula into a decision-relevant local statement. The useful question is often not merely "what is the rate?" but "how much does a small change matter here, in these units, under these assumptions?"
Build the model in stages and retain intermediate quantities. This makes unit checks possible and reveals which parameter drives the result. If measured data are involved, report uncertainty and avoid claiming precision the inputs do not support.
Read this graph as text
Input tolerance becomes output uncertainty through the derivative. A small measurement interval around x maps to an output interval around f(x) . The local slope estimates how much the output uncertainty is amplified or reduced. Near x=2 for f(x)=x 2 , the slope is 4 . An input uncertainty of about 0.1 therefore produces an output uncertainty of roughly 0.4 . The derivative is acting as a local amplification factor.
Every relationship in input tolerance becomes output uncertainty through the derivative is identified with written labels plus distinct solid, dashed, dotted, double, marker, or pattern cues; color is never the only carrier of meaning.
Why it matters: This visual connects differentials to manufacturing, measurement, and numerical sensitivity. It should make the phrase "error propagation" geometric rather than merely symbolic.
A small measurement interval around x maps to an output interval around f(x). The local slope estimates how much the output uncertainty is amplified or reduced.
A tolerance is an interval, and the derivative predicts how that interval propagates
Manufactured dimensions are never exact. If an input dimension may vary by , the derivative estimates the corresponding output variation . Large sensitivity means a small machining error can create a large performance error.
The estimate is local and assumes the tolerance is small enough for linearization. Safety-critical work may require rigorous bounds rather than a first-order approximation.
A cylindrical component has fixed height mm and measured radius
Its volume is
With height fixed,
Therefore
The relative error is approximately
So the volume uncertainty is about .
Conditioning versus measurement quality
A precise instrument can still feed an unstable formula, and a stable formula cannot recover information absent from a poor measurement. Numerical analysis separates the conditioning of the mathematical problem from the error introduced by the algorithm and data.
After the explanation
Use the section idea
Treat each derivative model as a conditional claim whose variables, units, assumptions, calibration range, and limitations remain visible.
A useful model connects a measurable input to a measurable output, while its derivative describes local sensitivity inside a stated domain.
Define the relationship and objective, differentiate, evaluate candidates or rates, then test sign, scale, units, and assumption sensitivity.
Extending a fitted model outside its data range or presenting medication, stopping-distance, or business outputs without the assumptions that shape them.
What observation would falsify the model, and how would the conclusion change if its strongest assumption failed?
Source & rights
Original instruction with traceable references.
BetterGrades-original composition declared by source handoff; owner provenance review required before public release
Reference textbooks remain rights-separated and are not published as application assets. Any direct adaptation requires separate identification and attribution.