Casting is not always the cheaper route, and it is not always the better one. It becomes economic at a particular combination of complexity, volume and required consistency, and below that combination it is an expensive way to make a shape that could have been fabricated. Knowing where that point sits is as much a part of component design as knowing the wall thickness.
The pattern is the reason. A pattern is a fixed cost that has to be amortised across the production run. At low volume, that fixed cost dominates and the per part price is unattractive. At high volume, it fades and the low marginal cost of a cast part takes over. The crossover is not a single number, because it depends on the complexity of the shape and on the cost of the alternative.
What makes a casting uneconomic is therefore usually one of a small number of conditions, and each of them can be identified before the tooling is committed.
A simple shape, made from plate or tube, is usually cheaper to fabricate than to cast. The fabrication uses standard material, standard cutting and standard welding, and it needs no pattern at all. The casting route would require a pattern, a proof, a first article and a qualification cycle to produce something that a fabricator can build from a cutting list.
Complexity is where casting earns its place. Internal passages, integrated bosses, varying section thickness, compound curvature and features that would be difficult or impossible to weld are the things a pattern can produce at almost no marginal cost once it exists. A casting that reproduces a simple box has bought nothing.
The judgement is not about whether the part could be cast. Almost anything can be cast. It is about whether the shape contains enough complexity to repay the tooling, and whether that complexity would be expensive to achieve another way.
Below a certain volume, the pattern cost cannot be recovered through the per part price. This is the most common reason a casting programme fails to make sense, and it is the one most often ignored because the per part price quoted for a high volume looks attractive in isolation.
The comparison has to be made across the actual quantity required, including the first article, the qualification castings and any scrap allowance. A programme that needs fifty parts and buys a pattern for them is paying for tooling that will never be amortised, and the effective unit cost is far above the quoted figure.
It is worth stating the obvious: the correct comparison includes everything. Pattern cost, proof cost, first article inspection, any pattern maintenance, and the cost of the time spent managing the programme. A fabrication quote that looks higher per unit may still be cheaper overall at low volume.
A third condition is a specification that the casting process cannot meet economically. Very thin walls, very tight tolerances on as cast features, or acceptance criteria that require sound metal everywhere will push the process into a regime where yield falls and cost rises faster than volume can compensate.
This is not a failure of casting. It is a mismatch between the requirement and the route. Sometimes the answer is to change the requirement, for example by allowing machining on a feature that was specified as cast. Sometimes the answer is to change the route. Either way, the mismatch is worth identifying early, because it will otherwise be discovered through scrap.
A drawing that demands machined tolerances on cast surfaces is the clearest example. The process can sometimes achieve them and will charge for the attempt, and the charge will be reflected in the price of every part, including the ones that pass.
The last of these deserves emphasis. A casting programme assumes a stable design. A component that is expected to change every few months is a poor candidate for a pattern, because every change is a tooling change and the amortisation never completes.
An honest comparison puts the casting and the fabrication side by side across the same quantity, with the same inspection and the same acceptance criteria. That is harder than it sounds, because the two routes have different cost structures and the comparison is easy to skew by comparing a mature casting price against a prototype fabrication price, or the reverse.
The fabricated alternatives are well established, and the range in the aluminium fuel tank range shows how much can be achieved without tooling when the shape allows it. That is the benchmark a casting has to beat, and it is a real benchmark rather than a theoretical one.
When a casting stops paying, it usually stops for a reason that was visible at the start: the shape was simple, the volume was low, the requirements exceeded the process, or the design was not stable. None of those are surprising in hindsight. The value of the analysis is in seeing them in advance, before the pattern is cut and the money is committed.
The right conclusion is not that casting is expensive. It is that casting is a tooling based route, and tooling based routes reward volume, complexity and stability. Where those three are present, casting is very hard to beat. Where they are absent, it is very hard to justify.
A design that changes every few months never finishes amortising its tooling. That is not a reason to avoid casting, but it is a reason to wait until the geometry settles. TIG aluminium welding absorbs change far more cheaply, which is why prototypes are usually welded.
Every pattern change is a tooling cost, a proof and a delay, and a programme that needs several of them has quietly become a fabrication programme with extra steps. aluminium water tank range and custom parts work is often chosen for that reason alone.
The comparison only means something when both routes are priced at the same quantity, with the same inspection and the same acceptance. TIG stainless steel welding and titanium fabricated components quotes should be read the same way as a casting quote, item by item.
