22-Mec-B5 Product Design and Development · December 2014

Question 4 of 7: The product designer and the process engineer compared

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Notes on this paper

National Exams, December 2014 — 07-Mec-B5 Product Design and Development. Three hours. Open book; no calculator permitted. Question 1 must be completed and is worth 40 marks; four of the six remaining questions are chosen, each worth 15 marks, for 100 marks in total. Only the first five questions as they appear in the answer book are marked, and the paper states that most answers are expected in essay form or as tables, figures and charts, with clarity and organisation carrying weight.

The paper prints 40 + 6 × 15 = 130 marks and a candidate attempts 40 + 4 × 15 = 100 of them. All seven questions are answered below, because this set is a study resource rather than an examination script. The published marking scheme on the last source page splits Question 1 as 6 / 9 / 9 / 6 / 4 / 6 and each 15-mark question into its own parts, and the answers below are proportioned to that split. The arithmetic is kept deliberately light — no calculator is allowed.

Reference texts for this subject

K. T. Ulrich and S. D. Eppinger, Product Design and Development — the framework text for this exam code: mission statement, customer needs, target specifications, concept generation, concept selection by Pugh screening and weighted scoring, design for manufacturing, product development economics.
G. E. Dieter and L. C. Schmidt, Engineering Design — problem definition, decision matrices, design for the environment, life-cycle costing, and the intellectual property chapter behind Question 3.
G. Pahl and W. Beitz, Engineering Design: A Systematic Approach — the systematic conceptual / embodiment / detail sequence and function structures.
G. Boothroyd, P. Dewhurst and W. Knight, Product Design for Manufacture and Assembly — part-count reduction, handling and insertion time, the design-for-assembly index invoked in Question 2.
M. F. Ashby, Materials Selection in Mechanical Design — translate, screen, rank by material index, seek supporting information; the property charts used in Question 7.
S. Kalpakjian and S. R. Schmid, Manufacturing Engineering and Technology — process selection, tooling cost and economic batch size, process capability.
I. Gibson, D. Rosen and B. Stucker, Additive Manufacturing Technologies — process families, design for additive manufacturing and qualification, for Question 2.
Canadian context: the Patent Act, Industrial Design Act, Trademarks Act and Copyright Act administered by the Canadian Intellectual Property Office (Question 3); the Canada Consumer Product Safety Act, the Motor Vehicle Safety Act with CMVSS 213, the Medical Devices Regulations under the Food and Drugs Act, and CSA Group standards accredited through the Standards Council of Canada (Question 6); ISO 14040/14044 for the life-cycle reasoning in Question 1.

Question 4: The product designer and the process engineer compared (15 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part A — Two different thought processes (6 marks)

The essential difference is what each engineer is allowed to treat as unknown. The industrial design engineer starts with a need that is not yet a specification and must discover what the product should be; the manufacturing process engineer starts with a part drawing that already exists and must discover how to make it repeatably. One is searching for the right problem, the other is searching for a capable solution to a problem already stated.

The industrial design engineer diverges before converging. The work runs from an ill-defined mission statement through customer needs, target specifications, concept generation, concept selection, prototyping and refinement — and deliberately widens the option space before narrowing it, because the cost of missing a better concept is far higher than the cost of generating several. Requirements are negotiable and are traded against each other; a specification that cannot be met is renegotiated with the customer. The dominant risk is building the wrong thing well, so the work is front-loaded with user contact, and iteration is by physical or virtual prototype. The mindset is comparative and qualitative early, quantitative late.

The manufacturing process engineer converges from the start. The output is fixed by the drawing and its tolerances, so the search is over process parameters, equipment and sequence. The work runs from process selection through capability studies, designed experiments, pilot runs, control-plan development and ramp to rate. Requirements are not negotiable in the same way — a tolerance is a contract with the designer — so the variable is the process window. The dominant risk is variation: a process that makes a good first part and an unpredictable thousandth part has failed. The mindset is statistical from the beginning, because the object of study is a distribution rather than an artefact.

What they share. Both work by structured iteration against explicit criteria, both must respect the same constraints of cost, safety, regulation and schedule, and both fail in the same characteristic way — by optimising a sub-system while ignoring an interface. That shared failure mode is exactly why the two roles must overlap in time rather than run in series: design for manufacture works only if the process engineer is in the room while the geometry is still soft, and concurrent engineering is the name for refusing to let the drawing be thrown over the wall.

Part B — How each represents critical data (6 marks)

The industrial design engineer represents data in forms that keep the link between what the customer wants and what the product does visible:

A customer-need hierarchy with importance ratings, gathered from interviews and observation.
A house of quality relating each need to the engineering metrics that serve it, with the correlation roof showing which metrics fight each other.
A target specification sheet: metric, unit, marginal value, ideal value, verification method — the form used in Question 1 part D.
Concept sketches and function structures, then Pugh and weighted decision matrices for selection.
3D CAD with model-based definition, tolerance stack-up analyses, finite-element and computational fluid dynamics results shown as fields over the geometry.
Prototype test reports against the specification sheet, and a requirements traceability matrix that shows every requirement is verified by something.

The manufacturing process engineer represents data in forms whose subject is variation over time:

A process flow diagram and a value-stream map with cycle times, inventory and takt.
A process failure modes and effects analysis with severity, occurrence and detection ratings, feeding a control plan that names the characteristic, the specification, the gauge, the sample size and the reaction rule.
Designed experiments presented as main-effect and interaction plots or response surfaces, which is how a process window is located rather than guessed.
Statistical process control charts, typically mean and range charts, plus histograms against specification limits.
Capability indices and gauge repeatability and reproducibility studies — the latter first, because a capability number computed with an incapable gauge is meaningless.
First-pass yield, scrap and rework Pareto charts and overall equipment effectiveness trends.

The contrast is that the designer's artefacts are mostly matrices linking requirements to solutions, while the process engineer's are mostly time series and distributions. One is asking “is this the right thing”, the other “is this thing still the same as it was yesterday”.

Part C — How each assesses success (3 marks)

The industrial design engineer measures success by whether the product meets its target specification at the ideal rather than the marginal values, by customer acceptance and market share, by warranty return and field failure rates, by the development cost and time actually spent against plan, and by the margin the product earns. The honest long-run measure is the warranty and returns data, because it is the only one that reports on needs the team failed to discover.

The manufacturing process engineer measures success by capability and stability, by first-pass yield and scrap, by cycle time against takt, by cost per part, by overall equipment effectiveness, and by safety performance. Capability is the headline number, and it has to be read correctly.

Given. A characteristic specified as 25.00 ± 0.05 mm, on which a capability study returns a mean of 25.012 mm and a standard deviation of 0.0104 mm.

Find. The potential and actual capability indices, and what they tell the process engineer to do next.

Potential capability compares the spread with the tolerance band. $C_{p}=\dfrac{\text{USL}-\text{LSL}}{6\sigma}=\dfrac{25.05-24.95}{6(0.0104)}=\dfrac{0.100}{0.0624}=1.603$.
Actual capability also penalises being off centre. $C_{pu}=\dfrac{\text{USL}-\mu}{3\sigma}=\dfrac{0.038}{0.0312}=1.218$ and $C_{pl}=\dfrac{\mu-\text{LSL}}{3\sigma}=\dfrac{0.062}{0.0312}=1.987$, so $$\boxed{C_{pk}=\min(C_{pu},C_{pl})=1.218}$$
Interpret. The process is inherently capable — $C_{p}=1.603$ clears the usual 1.33 requirement comfortably — but it is running 12 µm above nominal, and that offset alone drags $C_{pk}$ down to 1.218. The action is to recentre, not to reduce variation: shifting the mean to 25.000 would raise $C_{pk}$ to the full 1.603 with no change to the equipment. A designer reading only the second number would wrongly conclude the tolerance had to be opened.

Question 4 — results.
Quantity	Result
Potential capability	C_p = 1.603
Upper and lower capability	C_pu = 1.218, C_pl = 1.987
Actual capability	C_pk = 1.218
Mean offset from nominal	+0.012 mm
Recommended action	Recentre the process; C_pk then rises to 1.603

Topic. The complementary mindsets of product design and process development, the artefacts each uses to represent critical data, and the measures by which each judges success.

Key relations. $C_{p}=(\text{USL}-\text{LSL})/6\sigma$; $C_{pk}=\min\!\left[(\text{USL}-\mu)/3\sigma,\ (\mu-\text{LSL})/3\sigma\right]$; a common acceptance rule is $C_{pk}\ge 1.33$; $C_{p}=C_{pk}$ only when the process is centred.

Why this works. Each role's tools follow from the uncertainty it owns. The designer owns uncertainty about what is wanted, so the tools are matrices that keep needs, metrics and concepts explicitly connected. The process engineer owns uncertainty about repeatability, so the tools are statistical and describe a distribution over time. The pair of capability indices is the cleanest illustration: one measures the spread the process could achieve, the other measures what it is achieving where it actually sits, and the gap between them is a management instruction.

Common pitfall. Running the two roles in series, so that the process engineer first sees the geometry after the drawing is released and can only report what is unmakeable. The technical pitfall is quoting a capability index without checking the measurement system first, since gauge variation inflates the apparent process variation and makes a capable process look incapable.

Source. Ulrich and Eppinger, Product Design and Development, on the development process, concept selection and design for manufacturing; Dieter and Schmidt, Engineering Design, on problem definition and concurrent engineering; Kalpakjian and Schmid, Manufacturing Engineering and Technology, on statistical process control and process capability.

← Question 3 Next: Question 5 →