Carrying a subassembly across instead of designing a new one

Choosing an existing module over a better clean-sheet design, and why.

Carrying an existing subassembly across into a new product instead of designing a fresh one is usually the right call because a subassembly is a shape plus every hour of testing, tooling and supplier qualification that already went into proving it works, and a clean-sheet replacement has to earn back all of that invested value before it is actually better and not merely different.

A tow bar that is already fitted

Buying a second-hand car with a tow bar already fitted by a previous owner, even though it sits in a slightly awkward position for what will actually be towed, illustrates the trade plainly. A tow bar that already exists, already works and is already rated for a real load saves far more time and money than a perfectly positioned new one would be worth, because fitting a new one means sourcing a part, booking a fitter and paying for labour, all in exchange for an improvement that is mostly cosmetic. The awkwardly placed but working tow bar wins almost every time, and the same logic applies to an engineering subassembly that is not perfectly optimised for its new role but is already proven, already tooled, and already sitting on a shelf.

What sits behind a proven module

A subassembly (a wiring loom, a pedal box, a cooling module) represents a much larger investment than its own bill of materials suggests. Behind the parts sits every failure that was found and fixed during its development, every supplier relationship that already delivers it reliably, and every piece of tooling that already exists to make it repeatably. A theoretically better clean-sheet replacement starts from zero on all three fronts, however much better its geometry looks on a drawing.

The real choice is therefore between a known, proven quantity and an unproven one that promises to be superior once it too has been through the same development the old one already received. That promise is sometimes worth the wait and the risk. Often the theoretical improvement turns out to be smaller than the very real, very predictable cost of re-earning everything the old subassembly had already earned.

Measuring the gap that a redesign would close

This changes how a carry-over decision gets evaluated. A fresh design almost always looks better on a drawing, where none of its real-world problems have shown up yet, so asking whether it would be better in isolation settles nothing. The useful question is whether the specific, measurable gap between the old subassembly's performance and what the new product needs is large enough to justify paying, in time, money and risk, for a redesign to close it. Where that gap is small, or where the old subassembly's shortfall touches nothing the new product has to do, carrying it across is usually the correct engineering decision, even though it will always look like the less ambitious one on paper.

The same reasoning can quietly become an excuse to avoid a redesign that is overdue. A subassembly carried forward across several successive products without being re-evaluated can become a permanent, unquestioned fixture long after the gap between what it offers and what the product needs has grown large enough to justify replacing it, simply because nobody ran the comparison again with fresh numbers. The habit worth building is a periodic re-check of the same trade-off, since the value of an inherited subassembly erodes as that gap widens, and a comparison that favoured carry-over five years ago may well favour a redesign today. The tow bar on the second-hand car is a good buy until the owner starts towing a caravan it was never rated for, at which point the old one has to be replaced whatever it saved at the start.

My key error with this

I had assumed that carrying an existing subassembly across would constrain the hull design, since committing to a module's mounting points and envelope before the hull shape is settled feels like designing the interesting part around the boring one. It worked in exactly the opposite direction. Because the standard components and their interfaces stayed fixed, I did not have to redraw a single interface while I explored the shape around them, and the CAD work I saved, easily more than ten hours across the exercise, turned directly into five hull variations that I could put through quick simulation and then actually build. Had I redesigned the mounting arrangement for each variant, I would have produced perhaps two and understood neither of them particularly well. What replaced the belief is that a fixed interface is not a restriction on exploration, it is what makes exploration cheap enough to be worth doing, because the variable you are studying is only isolated when everything around it is held still.