A single order is a self-contained event. A long project is a series of orders made against a design that is still settling, and the commercial behaviour that suits one is wrong for the other. Buyers who plan a build across several stages need a relationship with a fabricator that can absorb change without losing the thread of the design.
The difficulty is not manufacturing. A workshop can make a part twice. The difficulty is memory: which version was agreed, which change was approved, and whether the second part is meant to match the first or to represent the newer idea.
A repeat order is easy when four things exist: a drawing, a revision, a tool and a record. With all four, ordering again is mostly a matter of confirming that nothing has changed and issuing the instruction.
Losing any one of them shifts the work back into engineering. Without a drawing, somebody re-measures. Without a revision, nobody knows which version was made. Without a tool, the fixture has to be rebuilt. Without a record, the material and finish may drift. The buyer's own filing is as important as the workshop's, because a repeat request usually starts from the buyer's side of the table.
Parts made from a documented drawing are far easier to repeat than parts made to a sample, and that is worth remembering at the start of a build. A part arising from a measured sample is repeated through the drawing the engineer produced, so that drawing becomes the real specification.
Not every second order should repeat the first. As a build develops, a bracket moves, a joint is replaced with a better type, or a clearance turns out to be tighter than expected. Ordering the same part again simply because it was ordered before is a way of repeating a problem.
The useful question is whether the reason for the change is understood. If the original part worked and the new requirement is different, a new version is right. If the original part did not work, the change should be described as a correction rather than as a variation, and the drawing should be updated so the next order does not go back to the faulty version.
This distinction matters most on welded assemblies where one component affects another. A change to a midpipe, for instance, can alter the alignment of everything behind it, so a change agreed in isolation can have consequences further down the system.
On a multi-stage project, the buyer should keep a simple version log. One line per part, giving the date of the order, the revision supplied and the reason for any change, is enough to prevent most confusion.
The log makes a specific question answerable: which version is currently on the vehicle. That question comes up whenever a new stage is planned, and it is surprisingly difficult to answer from memory once several parts have been swapped in and out.
Version control also protects against accidental regression. Where a change was made to solve a problem, ordering from an older drawing would bring the problem back. A maintained log catches that before the part is made, which is the only point at which it is easy to catch.
A change should be issued in writing, against a named drawing and a named revision. Verbal instructions given to whoever answers the phone will reach the bench in some form, and that form may not be the one intended.
A useful change request contains three things: what is changing, why it is changing, and whether previous parts are affected. The last point is the one most often omitted and the one a fabricator needs most. If a change affects earlier parts, the shop may need to advise on compatibility rather than simply make the new version.
It also helps to say whether the change is final or exploratory. A shop that knows a decision is still open will treat the tooling differently from one that believes the design is now frozen, and that judgement affects how much is committed to a titanium fabricated component or any other made-to-order item.
Tooling does not stay new. Fixtures wear, stops shift, templates lose an edge. Over a long project the effect is gradual, and gradual changes are the ones nobody notices until a new part will not fit alongside an old one.
There is a practical response. When a repeat order is placed after a significant interval, it is reasonable to ask whether the tooling is still in the condition it was in at the start, and whether the new parts should be checked against an original sample rather than only against the drawing. Buying one part to confirm a design is still valid is a normal request, not a criticism.
Matching old and new parts matters most where they have to join. An assembly that has to bolt between two existing components inherits the tolerance of both, and stainless and titanium work differently in this respect, which is why a stainless system already on the car is worth measuring before a new section is ordered to meet it.
There is a version of this problem that has nothing to do with tooling. Materials, finishes and finishes of parts change over the years as suppliers change or ranges are revised, so an identical drawing can still produce a part that does not look identical to its predecessor.
Acknowledging that early avoids disappointment. If the second part must be visually matched to the first, say so and supply a reference. If it only has to fit and function, say that instead, and the shop will not spend effort on a cosmetic match nobody needs.
The same logic applies across the whole project, whether the components are exhaust, intake or cooling related. A custom intercooler fitted in stage one may be joined by new piping in stage three, and the joining is where old and new have to agree.
Repeat orders and changes both benefit from the same discipline. Written references, a stated intention and an honest statement of what is still uncertain will keep a long project coherent.
A long project is easier for everyone when the paperwork accumulates instead of being reconstructed. Tigindustry has worked from a consistent sequence since 2009, where a drawing with dimensions, material and special requirements leads to tooling, samples and measurement before mass production is confirmed. Our titanium exhaust systems show what a settled design looks like in production form, the fabricated parts range shows the assemblies that typically repeat, and the air intake work shows how a project can grow through stages. Change is normal in custom work. Losing the record of it is what makes it painful.
