22-Mec-B5 Product Design and Development · May 2018

Question 7 of 7: Phases of a New Product Development Process

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

Notes on this paper

Paper format. National Exams, May 2018, 16-Mec-B5 Product Design and Development — THREE (3) hours, OPEN BOOK, one approved Casio or Sharp calculator permitted. Question 1 is compulsory and carries 40 marks; four of the remaining six questions are chosen, each worth 15 marks, for 100 marks in total. Only the first five questions appearing in the answer book are marked. Most answers are expected in essay form or as tables, figures and charts, and clarity and organisation carry marks in their own right.

Scope of this solution. All seven questions are answered in full, not the five a candidate would attempt, so that the paper works as a study resource. Where the examiner offers a choice of product, one is selected and carried consistently through every part, which is exactly what the question's own guidance note asks for. Numeric illustrations are engineering estimates built from stated, ordinary data; every one of them.

Reference texts for 22-Mec-B5.

Ulrich, K. T. and Eppinger, S. D., Product Design and Development, 6th ed., McGraw-Hill — the framework text for this exam code (development process, concept selection, product architecture, DFM, economics).
Dieter, G. E. and Schmidt, L. C., Engineering Design, 5th ed., McGraw-Hill — problem definition, requirement cascades, material and process selection, design for X.
Pahl, G., Beitz, W., Feldhusen, J. and Grote, K.-H., Engineering Design: A Systematic Approach, 3rd ed., Springer — systematic embodiment design and the function structure.
Boothroyd, G., Dewhurst, P. and Knight, W., Product Design for Manufacture and Assembly, 3rd ed., CRC Press — the DFA index and process-based cost estimating.
Ashby, M. F., Materials Selection in Mechanical Design, 5th ed., Butterworth-Heinemann — material indices, screening and ranking, process-cost models.
Kalpakjian, S. and Schmid, S. R., Manufacturing Engineering and Technology, 7th ed., Pearson — process capability and manufacturing constraints on design.
Cooper, R. G., Winning at New Products, 5th ed., Basic Books — stage-gate governance and gate criteria.
Vaver, D., Intellectual Property Law: Copyright, Patents, Trade-marks, 2nd ed., Irwin Law, together with the Canadian Intellectual Property Office (CIPO) guides — the Canadian statutory frame for Question 4.

Question 7: Phases of a New Product Development Process (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 — Five phases of the development process (5 marks)

Phase 1, planning. The phase before the project is a project. The organisation assesses technology and market opportunity, decides which markets and which segment to serve, sets the target price point and volume, and produces a mission statement: the product description, the business goals, the primary market and the assumptions and constraints the team may not change. The output is not a design; it is permission to spend, plus the constraints under which the spending happens.

Phase 2, concept development. Customer needs are identified and translated into target specifications with measurable values. The function is decomposed, concepts are generated for each sub-function and combined, and the resulting concept set is screened against a datum and then scored against weighted criteria. Concepts are tested with users, an economic analysis is built, and one concept is selected. This is where the greatest leverage in the entire programme sits: the concept commits the majority of the eventual cost while almost none of it has yet been spent.

Phase 3, system-level design. The product architecture is defined — the arrangement of functional elements into physical chunks and, crucially, the interfaces between them. The product is decomposed into subsystems with allocated budgets for mass, cost, power and performance; make-or-buy decisions are taken; the assembly scheme and the final assembly flow are laid out; and the geometric layout is fixed. Interfaces frozen here are the most expensive things in the programme to change later, because a change to an interface is a change to two subsystems and every schedule that depends on them.

Phase 4, detail design. The complete specification of every part: geometry, material, tolerance, finish, and the identification of purchased parts. The process plan is established and tooling designed for every fabricated part. The output is the control documentation for the product — the released model set, the drawings, the bill of materials, the process plans. This is the longest phase in headcount-hours and the one most often under-resourced, because it looks routine.

Phase 5, testing, refinement and production ramp-up. Pre-production units are built, first from prototype tooling and then from production tooling, and are used for design verification and then production validation: performance, durability, reliability, regulatory certification and, finally, capability demonstration on production processes. The workforce is trained, the line is proved out, and output climbs to rate. Early production is often supplied to selected customers before general launch precisely so that problems surface where they can still be contained.

Part B — Establishing success within each phase before moving on (5 marks)

The mechanism is a gate: a decision point with pre-agreed, evidence-based exit criteria and a decision-maker who may say no. The criteria are set when the phase begins, not when it ends, and the deliverable list is what makes the gate reviewable rather than a status meeting.

Phase	Exit criteria (evidence, not opinion)
Planning	Mission statement approved; market and volume assumptions documented with their sources; business case shows an acceptable return; resources and a chief engineer committed by name.
Concept development	Customer needs documented and ranked; target specifications with measurable values and verification methods; a scored concept-selection matrix with weights agreed in advance and a sensitivity check; concept validated with users; economic model updated.
System-level design	Architecture and all interfaces frozen and under change control; subsystem budgets allocated with a named owner and a contingency reserve; make-or-buy decided and key suppliers selected; assembly flow defined; long-lead tooling identified.
Detail design	Complete design release with every drawing and model at production revision; DFM and DFA review closed with actions dispositioned; tolerance stacks complete; design FMEA closed; tooling kick-off approved.
Testing and ramp-up	Design verification and production validation complete against the requirement matrix; reliability demonstrated at the required confidence; regulatory certification granted; process capability $C_{pk}\ge 1.33$ on all critical characteristics; run-at-rate achieved at target first-pass yield.

The gate is also the point at which the business case is re-tested, because the information available has changed since the last gate. The standard instrument is the expected commercial value, which prices both the technical and the commercial risk.

Given. At the system-level design gate: commercial present value if successful $PV$ = CAD 38.0 M; probability of commercial success $P_{cs}=0.75$; remaining commercialisation cost $C$ = CAD 9.0 M; probability of technical success $P_{ts}=0.85$; remaining development cost $D$ = CAD 3.2 M.

Find. The expected commercial value at the gate and the productivity index used to rank it against competing projects.

Evaluate the gate arithmetic. Working outward from commercial success through technical success, $$ECV=\bigl[(PV\times P_{cs})-C\bigr]P_{ts}-D=\bigl[(38.0\times 0.75)-9.0\bigr]0.85-3.2=\boxed{\text{CAD }13.375\ \text{M}}$$ and the productivity index, which ranks projects by value returned per unit of the constrained resource, is $$PI=\frac{ECV}{D}=\frac{13.375}{3.2}=\boxed{4.18}$$
Read the ramp-up criterion in the same units. A capability of $C_{pk}=1.33$ places the nearer specification limit at four standard deviations from the mean, so the expected defect rate is $$2\bigl[1-\Phi(4)\bigr]=2(3.167\times 10^{-5})=\boxed{63\ \text{ppm}}$$ which is what the gate criterion actually promises the customer, and is the honest way to state it to a plant that is being asked to hold it.

Two governance points make the difference between a real gate and a ceremonial one. The gate must be able to kill or hold the project, and killing must be an acceptable outcome for the people who present at it, or the criteria will be quietly relaxed to fit whatever was achieved. And a criterion that cannot be evidenced — "the design is mature" — is not a criterion; each one must name the artefact that proves it.

Part C — Benefits and challenges of compressing the phases (5 marks)

Benefits. Earlier launch earns revenue earlier and for longer, which in a market with a fixed window is worth more than an equivalent cost saving; it reduces the risk that the market or the technology moves under the programme; it captures first-mover position and the pricing that goes with it; and it lowers the carrying cost of a project team, since a shorter programme is a cheaper one at the same monthly burn. There is a real quality benefit too, often overlooked: overlapping phases forces manufacturing, service and suppliers into the design early, which is the same concurrency that produces better designs rather than merely faster ones.

Challenges. Overlap means starting a phase on information that is not yet stable, so the downstream work may have to be redone when the upstream decision changes. Tooling committed before validation is capital at risk. Verification is the phase that gets squeezed, because it is last and its cost is visible, and squeezing it moves failures from the test laboratory into the field where they cost an order of magnitude more. Suppliers are asked for quotes on unstable data and price the uncertainty back in. And the team runs at sustained overload, which raises error rates precisely when the error consequences are largest.

The right way to answer "how much overlap" is to model it, because the saving is linear in the overlap and the expected rework is not.

Given. Phase durations of 5, 7, 9, 6 and 4 months, so 31 months strictly sequential, and 26 months of downstream work available to overlap. Each phase is started a fraction $f$ of the downstream duration early. The probability that an upstream change forces rework grows with the overlap as $p(f)=1.4f$, and rework consumes a fraction $r = 0.8$ of the overlapped work.

Find. The overlap fraction that maximises the net saving, the resulting programme duration, and the overlap beyond which compression costs more than it saves.

Write the net saving as a function of the overlap. The gross saving is $S(f)=26f$ months. The expected rework is the probability of a change times the fraction reworked times the overlapped work, so $$R(f)=p(f)\,r\,(26f)=1.4f\times 0.8\times 26f=29.12f^{2}$$ and the net saving is $N(f)=26f-29.12f^{2}$.
Maximise, and locate the point where compression stops paying. Setting $dN/df = 26-58.24f = 0$ gives $$f^{*}=\frac{26}{2(29.12)}=\boxed{0.446}$$ with a net saving of $N(f^{*})=11.607-5.803=5.80$ months, so the programme runs $31-5.80=\boxed{25.2\ \text{months}}$ rather than 31. The raw compressed schedule would have been 19.4 months; the difference is the expected rework, and a plan that quotes 19.4 months is quoting a schedule with the risk removed from it. Net saving returns to zero at $$f_{\text{be}}=\frac{26}{29.12}=0.893$$ so beyond about 89 per cent overlap the expected rework consumes the entire benefit.

The sequential plan against the optimum overlap. The red block is the expected rework the overlap buys; a programme plan that omits it is not a compressed plan, it is an optimistic one.

Three practical conclusions follow from the shape of that curve, and they are what the question is really asking for. First, the optimum overlap is substantial but well short of total, so the choice is not between sequential and concurrent but about how much of each interface to overlap. Second, the parameters that set the optimum are $p$ and $r$ — how likely an upstream change is and how much of the downstream work it destroys — so the highest-value action is not to overlap harder but to reduce those parameters: freeze interfaces early, release information in controlled increments with a stated maturity level, and design the downstream work so that a change invalidates only part of it. Third, compression should never be applied uniformly. Verification and validation are the phases whose compression converts directly into field failures, so the honest programme compresses the phases where rework is cheap and protects the ones where it is not.

Check: the overlap model assumes a single rework-probability coefficient applied uniformly across all four phase interfaces, and that rework is proportional to the overlapped duration. A real programme would calibrate $p$ and $r$ per interface from its own change history, and the coefficient of 1.4 used here should be read as an illustrative value in the range typical of programmes with unstable upstream requirements, not as a general constant.

Topic. The generic development process and its phases; stage-gate governance and gate economics; concurrent engineering and the economics of schedule compression.

Key relations. Expected commercial value, $ECV=[(PV\cdot P_{cs})-C]P_{ts}-D$; productivity index, $PI = ECV/D$; capability to defect rate, $\text{ppm}=2[1-\Phi(3C_{pk})]\times 10^{6}$; overlap model, $N(f)=Df-p_1 r D f^{2}$ with optimum $f^{*}=1/(2p_1 r)$.

Why this works. The expected commercial value multiplies through the risks in the order they are resolved — commercial payoff first, then the technical gate that must be passed to reach it — so it prices a project the way an option is priced rather than the way a certainty is. The productivity index then divides by the constrained resource, which is why it, rather than raw value, is the correct ranking criterion when development capacity is the bottleneck. The overlap optimum exists because the benefit is linear in the overlap while the expected rework is quadratic: the probability of a disruptive change and the amount of work exposed to it both grow with $f$, so their product grows as $f^{2}$ and eventually overtakes a linear saving. That quadratic form is the entire reason a finite optimum exists, and it is why intuition — which sees only the linear benefit — systematically over-compresses.

Common pitfall. Quoting the raw compressed duration as the plan. A schedule built from the gross saving alone assumes no upstream change will ever occur, which is exactly the assumption the overlap makes false; the resulting plan is missed, and it is missed in the verification phase at the end, where the schedule pressure is greatest and the consequences worst. The parallel pitfall in Part B is a gate that cannot say no, which converts governance into reporting and lets a weak concept consume the detail-design budget.

Source. Ulrich and Eppinger, Product Design and Development, 6th ed., Ch. 2 (development processes and organisations) and Ch. 16–18; Cooper, Winning at New Products, 5th ed., Ch. 4–6 (stage-gate and portfolio methods); Dieter and Schmidt, Engineering Design, 5th ed., Ch. 2 and Ch. 18.

← Question 6 Back to the paper →