Quality Control Off-Site

Two terms that get muddled are worth separating first. Quality assurance (QA) is about the *system* that makes quality happen: the documented processes, the approved methods, the training, the checked jigs - everything set up so that good output is the normal result of following the process. Quality control (QC) is the *checking*: the inspections, measurements and tests that verify a given piece actually meets the requirement. QA is 'are we set up to do it right?'; QC is 'did this one come out right?' A good factory runs both - a sound process (QA) with verification built in (QC) - and leans on QA especially, because a process that reliably produces good work needs less frantic inspection to catch bad work.

The factory's signature advantage is inspection at the bench, at the right moment. Because the work flows through stations (lesson 6.1), quality can be checked *between* operations, while the relevant work is still exposed. First-fix services are inspected and tested while the wall is still open; a frame's squareness is gauged before it is sheathed and the geometry locked; a weld is checked before it is clad over. These are the quality gates of lesson 6.1, and the figure shows them sitting between stations. The payoff is captured by a simple, brutal rule that holds across all making: the cost of fixing a defect rises steeply the later it is found. Caught at the bench, a defect is a minute's rework; caught at the end of the line, it may mean disassembly; caught on site after craning, it is a major operation; caught in service, it can be a failure and a claim. Bench inspection attacks defects where they are cheapest to fix.

The other half of factory QC is repeatability of the check itself. On a site, an inspection depends on the inspector being there, at the right moment, seeing the right thing before it is covered - hit and miss by nature. In a factory, the check is a fixed step in the sequence, done the same way every time, often with a go/no-go gauge or a jig that makes pass/fail a matter of fit rather than opinion. Standardising the check is as important as standardising the part: a repeatable, well-placed, objective check catches what a one-off judgement on a chaotic site would miss. This is the quiet reason factory quality is more *consistent* - not that factory workers are better, but that the checking is systematic.

The factory's deepest quality advantage is one a site struggles to match at all: traceability - the ability to know, after the fact, exactly how a given unit was made. Because production is organised, documented and repetitive, a factory can record for each unit which batch of material went into it, which jig setting and machine produced it, which operator built it, which inspector signed each gate, the date and time, and the results of any tests. Modern factories attach this to each unit with a serial number, barcode or tag, building a birth certificate that travels with the module. Increasingly this links to the digital model (lesson 6.4 and Module 10's digital thread), so the as-built record is tied to the as-designed intent.

Why does this matter so much? First, when something does go wrong, you can find the extent. If a batch of sealant is later found defective, or a jig is discovered to have drifted out of tolerance, traceability lets you identify *exactly which units* are affected - rather than suspecting all of them, or worse, none. That turns a potential disaster into a bounded, fixable problem. Second, it enables accountability and learning: patterns in defects can be traced to their source (a particular shift, machine or material) and fixed at root, feeding the continuous improvement of lesson 6.2. Third, it supports approvals and warranties: a manufacturer can demonstrate to a building-control authority, a client or an insurer that each unit was made to the approved process and passed its checks - which is part of how off-site systems earn regulatory acceptance.

Alongside traceability sits testing. A factory can test in ways a site cannot easily arrange: sample units or assemblies can be subjected to controlled load, air-tightness, water-penetration, acoustic or fire tests, often as part of getting a system approved, and routine checks (torque on bolts, moisture content of timber, dimensional checks against the jig) run continuously on the line. The key distinction, and one this course insists on, is between routine production checks (which verify that units match an already-approved, already-tested design) and the binding performance testing and certification of the system itself - the fire-resistance rating, the structural capacity, the acoustic separation. The latter is engineering and approvals work belonging to qualified specialists and the manufacturer's tested system, not something a production-line check establishes. Factory testing confirms conformance to a proven design; it does not, on its own, prove the design.

Now the honest counterweight, and the single most important idea in this lesson. Everything that makes factory quality strong - repetition, jigs, a fixed process, identical units - has a shadow. The same repetition that stamps a proven detail perfectly onto every unit will stamp a hidden flaw just as perfectly onto every unit. On a chaotic site, errors tend to be random and local: one corner out of square, one duct clash, each different, each caught (or not) on its own. In a factory, an error built into the *setup* is systematic - it repeats identically across the whole production run until someone notices. A jig set two millimetres out makes every frame two millimetres out. A wrong batch of fixings goes into hundreds of modules. A detail that is subtly flawed in the design is manufactured flawlessly, hundreds of times, before anyone questions it. The systematic-error figure shows this double edge: the good detail replicated on the left, the single flaw replicated on the right.

This changes where quality effort must go. On a site you inspect *output*, sampling finished work. In a factory you must above all guard the setup and the first article: check the jig is true before the run, verify the first unit off a new setup exhaustively (first-article inspection), confirm the material batch is right, and re-check whenever anything changes. A defect in unit one that is caught is a non-event; the same defect uncaught is now in every unit. This is why serious factory QC obsesses over calibration, setup verification and change control - the leverage points where a single error becomes a thousand. It is also an argument for the traceability above: when a systematic error is found late, the record is what lets you contain it.

There is a sobering real-world lesson here that off-site must take seriously: because modular systems industrialise the detail, a flawed detail industrialises too. A connection, a fire-stopping detail or a waterproofing junction that is wrong in the prototype is wrong in every unit - and may be discovered only after many buildings are complete. This is precisely why the binding details - structural connections, fire-stopping and compartmentation, waterproofing - must be got right by qualified engineers and proven in the manufacturer's tested system *before* the run, not discovered on the line. The factory does not forgive a bad detail; it multiplies it. Respecting that is the difference between off-site's quality promise and its quality risk.

The final limit is one of scope. The factory controls only what it makes. It can deliver a near-perfect module, inspected and traced - and that module still has to be craned onto a foundation it did not build, joined to other modules at connections made on site, sealed at weather joints formed in the open, and tied into drainage, services and the public realm by site trades. All of that is interface work, and it happens back in exactly the uncontrolled, weather-exposed, hard-to-inspect conditions the factory was meant to escape. Off-site does not abolish site quality problems; it concentrates them at the interfaces - and those interfaces are now more critical, because a handful of connections hold together work that was otherwise made to a high standard.

This creates a characteristic off-site risk: a project can have superb modules and still fail at the joints. The weather-tightness of a building made of perfect modules depends entirely on the seals between them; the structural integrity depends on the connections; the fire and acoustic separation depends on the junctions being closed correctly on site. A defect at an interface can undo the quality of everything the factory did - and these interfaces are often hidden quickly as assembly proceeds, recreating the site's concealment problem at the very points that matter most. Good off-site practice therefore treats interface inspection as a first-class activity: defined hold points on site where a connection or seal is checked before it is covered, clear responsibility for who inspects the joint between the factory's scope and the site's, and details designed (Module 5) to be inspectable and tolerant.

So the honest summary of off-site quality is two-sided. On one side, the factory genuinely tends to beat the site: accessible bench inspection at the right moment, repeatable objective checks, traceability and testing, and a process that makes good output normal. On the other, repetition can replicate a systematic error across every unit, routine checks confirm conformance but do not prove the design, and the factory controls nothing at the site interfaces where failures concentrate. The design response runs right through this course: get the binding details right with the specialists and the manufacturer's tested system before the run; guard the setup and the first article; and design and inspect the interfaces as carefully as the modules. Quality off-site is not automatic - it is a different, and in many ways better, game with its own distinctive ways to lose.