JC Applied Technology cheat sheet

Design process order: define the problem, research constraints, generate alternatives, select, prototype, test, iterate; testing feeds back into earlier stages rather than ending the cycle.

Constraint types: functional (what it must do), physical (size, mass, material), economic (budget, unit cost), legal (codes, safety standards), and ethical/environmental limits.

Trade-off analysis: when criteria conflict, weight each criterion, score alternatives against the weights, and document the rejected options and the reason for rejection.

Prototype fidelity rule: use low-fidelity models (sketch, cardboard, breadboard) to test uncertain assumptions cheaply; raise fidelity only after the assumption survives testing.

Failure analysis tools: root-cause (5 Whys, fishbone) for events already observed; FMEA ranks risk as severity x occurrence x detection to prioritize fixes before failure.

Iteration convergence test: a design is ready to release when successive test cycles produce diminishing changes in measured performance against the stated criteria, not when time runs out.

Documentation standard: record date, version, test conditions, raw measurements, and the decision made, so results are reproducible by someone who was not present.

Human factors: usability testing with representative users reveals interface errors that expert review misses; sample size can be small if tasks and failure modes are observed directly.

Measurement uncertainty: report a value as measured value plus or minus uncertainty; the uncertainty of a sum or difference is the root-sum-square of component uncertainties for independent sources.

Resolution versus accuracy: resolution is the smallest display increment; accuracy is closeness to the true value; a high-resolution instrument can still be inaccurate.

Calibration: compare an instrument against a traceable reference and record the correction; calibration does not repair drift, it documents and compensates for it.

Signal sampling: the Nyquist criterion requires sampling above twice the highest frequency of interest; undersampling produces aliasing that cannot be removed by filtering afterward.

Sensor selection: match range, sensitivity, response time, and output type (analog voltage, current loop, digital bus) to the data-acquisition input; a 4-20 mA loop is chosen for noise immunity over long cable runs.

Control loop basics: proportional term responds to present error, integral removes steady-state offset, derivative damps overshoot; too much integral gain causes oscillation and windup.

Power budget: total load current equals the sum of device currents; battery runtime is capacity in amp-hours divided by average current, derated for temperature and conversion losses.

Data integrity: averaging N independent samples reduces random noise by the square root of N; it does not reduce systematic bias.

Material property selection: stiffness (modulus) governs deflection under load, strength governs failure, toughness governs energy absorbed before fracture, and hardness governs wear resistance.

Stress equals force divided by cross-sectional area; strain equals change in length divided by original length; the slope of the linear region of a stress-strain curve is the elastic modulus.

Safety factor equals the material's failure strength divided by the maximum expected stress; it accounts for load uncertainty, material variability, and consequences of failure.

Joining decision rule: fasteners allow disassembly and accommodate dissimilar materials; adhesives spread load over area but are sensitive to surface preparation and temperature; welding suits same-metal structural joints.

Corrosion control: choose a more noble material pairing, apply barrier coatings, or design drainage so moisture cannot pool; galvanic corrosion accelerates where dissimilar metals meet in an electrolyte.

Manufacturing process fit: injection molding for high-volume polymer parts with complex geometry, machining for low-volume precision, casting for complex metal shapes, additive manufacturing for one-offs and internal channels.

Tolerance stack-up: worst-case method sums all tolerances, statistical (root-sum-square) method assumes independent variation and yields a tighter but riskier assembly tolerance.

Sustainability metric: embodied energy and recyclability are evaluated per functional unit over the full life cycle, not per kilogram of material, to compare designs fairly.