Question 6 of 7: Communicating a design to the manufacturing team

Question 6: Communicating a design to the manufacturing team (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 — Four key pieces of information the designer must communicate

1. The product definition — toleranced geometry with a datum reference frame. This is the geometric truth of the part: every dimension, the tolerance on each, the datum features that establish the coordinate system from which those dimensions are measured, the surface finish, and the identification of which characteristics are critical to quality. A dimension without a tolerance is not information, because no process is exact; a tolerance without a datum reference frame is ambiguous, because the same feature measured from a different origin gives a different answer. Geometric dimensioning and tolerancing per ASME Y14.5 exists precisely to remove that ambiguity, and it also states function — a position tolerance at maximum material condition tells the factory that the requirement is assembly clearance and grants them the bonus tolerance that follows.

2. The bill of materials and the material and component specification. What every part is made of, to what grade and condition, with what finish and what restricted-substance requirements; which purchased components are approved, from which sources, and whether equivalents may be substituted and by whose authority. "Nylon" is not a specification; "PA6 with 30 per cent glass by mass, from the approved source list, natural colour, no regrind" is. Most substitution disputes are traceable to a bill of materials that stated a family instead of a grade.

3. The process and assembly definition. The assembly sequence and orientation, the joining methods and their parameters — torque values and sequences, weld schedules, adhesive cure profiles, press-fit forces — the fixtures and tooling required, handling constraints, and any process the design depends on for its properties. This is where design intent that is invisible in the geometry lives: a bolted joint is only as good as its preload, and a moulded part is only as strong as its cooling profile allowed it to be.

4. The acceptance criteria and the control plan. What is inspected or tested, on what sample and at what frequency, against what limits, using what gauge, and what happens when it fails. Attached to it are the first-article inspection requirement, the process-capability expectation on critical characteristics, and the engineering change route, so that both parties know how the definition is allowed to move. Without this the factory has been told what to build but not what "right" means, and it will define the term itself.

Two further items are frequently and wrongly left implicit: the volume and takt forecast, which determines whether the process route the designer assumed is even economic, and the packaging and logistics definition, which routinely turns out to constrain the geometry.

Part B — Challenges when the manufacturing team speaks a different language

The obvious challenge, technical vocabulary, is the least of it, because vocabulary can be looked up. The damaging problems are the ones that do not announce themselves.

Unstated defaults differ by engineering culture. Projection angle is the classic case: a drawing read in first-angle convention when it was drawn in third produces a mirror-image part that passes every dimensional check. Decimal commas against decimal points, thousands separators, date order, thread standards, "standard" fits, and default general tolerances all vary by region, and none of them is stated on the drawing because in each culture everybody knows.

False confidence in translated text. A work instruction translated once and then revised in the source language silently goes stale, and a translator without process knowledge will render "hand tight" or "snug" in a way that changes the preload. Terms with a precise engineering meaning and a loose everyday meaning — tolerance, stress, load, capability — are the ones that translate worst.

Loss of the informal channel, amplified by time zones. Co-located teams settle ambiguity in thirty seconds at a bench. Across a language barrier and a time-zone gap, the same ambiguity becomes a written round trip, and the round trip is the real cost.

Given. Design in Toronto (UTC−5), process engineering in Stuttgart (UTC+1) and assembly in Shenzhen (UTC+8), each working 08:00 to 17:00 local. Find. The synchronous communication available to the three sites.

1. Put every site on one clock. Converting each local window to UTC gives Toronto 13:00-22:00, Stuttgart 07:00-16:00 and Shenzhen 00:00-09:00.
2. Intersect the windows pairwise. Toronto and Stuttgart overlap 13:00-16:00, a window of 3.0 h; Stuttgart and Shenzhen overlap 07:00-09:00, a window of 2.0 h; Toronto and Shenzhen do not overlap at all. The three-way intersection is therefore $\boxed{0.0\ \text{hours}}$.
3. Test whether goodwill fixes it. Extending everyone to a 07:00-19:00 local day — a twelve-hour day, well beyond what is sustainable — raises the pairs to 6.0 h and 5.0 h but leaves the three-way intersection at $\boxed{0.0\ \text{hours}}$. The constraint is structural, not a matter of effort.
4. Count the channels that must carry the load instead. Three sites give $3(3-1)/2 = 3$ site-to-site links, but a 14-person cross-functional team has $n(n-1)/2 = 14 \times 13/2 = \boxed{91}$ possible person-to-person channels, almost none of which can ever be synchronous.
5. Price the consequence. Because the cost of a defect rises about tenfold per stage, an ambiguous surface-finish callout caught at first-article inspection costs of the order of CAD 3 800, the same ambiguity caught at production launch about CAD 38 000, and one that escapes to the field about CAD 380 000. Each day of round-trip delay pushes ambiguities one step further down that ladder, which is the real price of the missing window.

The conclusion is structural and it is worth stating explicitly, because it is what the question is testing: with a three-way overlap of zero, "we will hold a daily call" is not a plan. The team must be asynchronous by default, and any synchronous window has to be bought with somebody's evening — 12:00 UTC reaches all three, at 07:00 in Toronto, 13:00 in Stuttgart and 20:00 in Shenzhen — and the burden of that hour should rotate rather than always falling on the site with least authority.

Part C — Three tools to facilitate communication across a language barrier

Tool 1: model-based definition, so that the authority is geometry rather than text. The annotated 3D model carries the tolerances, datums and surface requirements attached to the features themselves and is released as the legal definition of the product, with drawings reduced to derived views. Geometry does not translate, so ambiguity of language is removed at the source rather than managed downstream, and the projection-angle failure mode disappears entirely because there is no projection. Pair it with a short controlled glossary — the two hundred terms that actually appear on this programme, agreed and versioned in both languages — so the residual text is at least consistent.

Tool 2: visual work instructions and physical error-proofing. Replace text-based assembly instructions with photographic or rendered sequences using standardized symbols (ISO 7000 for equipment symbols, ISO 7010 for safety), colour-coded parts, and torque values shown as numerals with units. Then take the further step of designing the ambiguity out physically: fixtures that only accept the correct part in the correct orientation, connectors that cannot mate the wrong way, asymmetric mounting patterns. Error-proofing is the most reliable form of cross-language communication there is, because it does not rely on the instruction being read at all.

Tool 3: a shared product lifecycle management system with a closed-loop confirmation ritual. A single controlled repository ensures both sites work from the same revision, with change orders routed and acknowledged rather than emailed. The closing element is what makes it work across a language barrier: require the manufacturing team to produce their own process sheet and a first-article inspection report against the definition, and review their document. A teach-back demonstrates comprehension; an acknowledgement demonstrates only receipt, and across a language and hierarchy boundary the difference between the two is where the defects live. Supplement with recorded video walk-throughs, which can be replayed and re-subtitled, and with augmented-reality remote assistance for the first build.