Test the Risky Part Before You Pay for the Whole Thing

The instinct with a new invention is to have it built properly: a full design, a polished prototype, something you could put in front of anyone. I understand the pull. I also think it's the most expensive way to find out whether an idea works.

A proof of concept should test the one thing you're unsure of, and nothing else. A rough version of that one thing costs a fraction of a finished prototype, and it tells you whether the finished one is worth building.

A robot in a grain bin

Well before we took on the Uxbridge shop, an inventor came to us with a good idea: a robot that travels inside a grain bin, carrying sensors through the grain. They had a large quote to build the whole robot.

That money would probably have paid for a working product. But nobody yet knew whether the core idea worked. Can wheels move through grain at all? Which wheel design?

Our advice was to leave the robot alone for now. Strap a motor to a piece of plywood. 3D print one of the wheel designs. Run it with an Arduino, a small, cheap controller board. See if it moves.

What the plywood buys you

The rig is worth building because of the decision it lets you make. If the wheels spin in place, the inventor has kept most of that large quote in the bank and learned what the hard problem really is; if they move, the inventor walks into the next conversation, with an investor or a builder, holding a result instead of a hope.

Either way, the money spent was small and the question was answered.

Rough, but honest

Scrappy is a stage. The standard doesn't change. The test rig can be plywood and tape, but the test still has to be measured honestly, and when the idea moves on to real parts, those parts are made to the drawing.

There is a place for full-scale development: a large team, testing to the last detail, a polished version at the end. Some firms do it well. Once the core idea is proven, it can be a sensible next step. Before then, it's a lot of money spent on parts of the design that may still change.

Where this step sits in the whole path, from writing the idea down to protecting it, is set out in my guide to the order of operations for a new idea.

The same idea for molded products

It works for plastic parts too. Before paying for a production mold, you can prove the part molds at all.

A single-cavity aluminum mold with hand-loaded inserts in place of side actions can do that at a fraction of the cost of a production tool. A hand-loaded insert is a loose piece of metal that forms an undercut or a hole. An operator places it in the mold before each shot and takes it out of the finished part, so the mold needs no moving slides. It's slow per part, which doesn't matter when the goal is to learn. If the part needs changing, you find out on a mold built to be changed. More on choosing the metal is in aluminum or steel mold, choosing for your stage.

When a project is too new for anyone to quote, the same approach shapes the quote itself. That's its own post: how to get a quote on something nobody can quote.

Write down the one thing you're least sure will work. That's your first prototype.

Related services: Product realization and prototyping Shop capabilities