Designing around a tool that does not exist yet

Committing to geometry before the means of making it was available.

Designing around a tool that does not exist yet means committing a drawing to a feature, a channel width, a bend radius, a bore diameter, before the machine or cutter capable of producing it has been ordered, tested, or in some cases even built, a bet that the tool will arrive and perform as specified by the time the drawing needs it.

A door ordered before the doorway is framed

Specialised tooling, whether a custom cutter, a purpose-built fixture or a machine capable of a feature nothing in the shop could produce, often takes longer to arrive and prove out than the rest of a project's design work. That leaves two paths. The project can wait for the tool to exist and be verified before finalising the geometry, losing that time from the schedule, or it can commit the drawing to the tool's expected capability while the tool is being built and risk finding out at the end that the two do not quite agree. The schedule here could not absorb the wait, so the second path was taken and the risk was lived with.

Ordering a custom door to a specified size before its doorway has been framed is the same kind of bet. The door is built to a fixed size in one workshop while the doorway is framed to the same size somewhere else, and neither can be held up against the other and corrected until delivery day. It usually works, because both trades are working from one drawing and know how much tolerance the job can absorb, but it depends on both sides trusting that drawing completely. The manifold's channel geometry, drawn around a cutter that had been ordered but not delivered, relied on the drawing as the only agreement between a committed design and a tool still being built to match it.

Finding the few features that carried the risk

Betting on an unproven tool sounds reckless stated plainly, but the risk was narrower than that. The manufacturer had built comparable tooling before, and the specification sat well inside their demonstrated experience, closer to a variation on a known capability than a first attempt. The real uncertainty sat in a few dimensions pushed toward the edge of what they had previously achieved. Managing the risk meant identifying exactly which features those were and building tolerance and inspection margin around them, while the rest of the tool was treated as the known quantity it was.

That concentration of attention paid off. On the features that carried real risk, the tool arrived close enough to its specification that the manifold needed only a small adjustment instead of a redesign, and a team spreading its limited review time evenly across every dimension would have had little left for the ones that mattered.

A fallback sat in reserve throughout: a less ambitious version of the same manifold that could be made on equipment already proven in the shop, with a larger footprint and a less elegant channel layout. A bet worth making still needs a way to recover if it fails, and with a working alternative ready, the schedule was never hostage to the new tool in the way the drawing alone suggested.

The habit carried into later projects. Features that needed an unproven tool's new capability were separated from the much larger number that did not, and only that smaller group was exposed to the risk. A drawing built this way fails gracefully if the tool underperforms, since only those few features need reworking, while a drawing that spreads the dependency across every dimension turns a small shortfall into a wholesale redesign.

My key error with this

A vendor told me they could produce very long channels at a scale far smaller than anything I had designed for, and I did not really believe them, so I spent an unnecessary amount of effort designing around the capability instead of to it, hedging the layout so that the design would still function if the channels came back shorter or less consistent than promised. That hedging cost real space and real complexity in a design where both were scarce. When the parts eventually arrived they met the specification they had been quoted against, and my caution turned out to have been the expensive part of the project. The deeper mistake underneath it was assuming that my intuition transferred, since I had spent years developing a feel for how fluid behaves in channels of a size I could see, and at that much smaller scale the behaviour I was quietly bracing for was not the behaviour that governs. What replaced the belief is that intuition is calibrated to the scale you built it at, and that a supplier working daily at a scale I had never worked at is a better source of truth about that scale than my instincts are.