I make real things on a 3D printer, then show you how.

Most of what I print starts with a small annoyance of my own. An external drive trailing its cable beside my MacBook. MagSafe docks that are all made to stay on a desk, when I change where I work all the time: home, the office, outdoors. A 300 W charger too big and heavy for the desk, with nowhere to put it underneath, and nothing on Shopee that fits both the charger and my desk leg. So I printed one myself.

Sometimes it’s someone else’s problem. Uyên Bee, a business that raises and supplies stingless bees, wanted a proper door for their hives in place of the bottle or bit of pipe stuck on the hive mouth — in colours that suit the box, with their brand on it. A furniture business needed a few sample floor tiles to judge a design before paying hundreds of millions of đồng for a production mould.

And sometimes it’s just curiosity. I don’t run, but when I saw F1 circuit frames on Pinterest I tried making one from a run on Strava. That frame is now part of the maps you can order here, alongside 3D relief maps of a province, a street or a journey. Or the time I tried the Snapmaker U1, four print heads on one machine, to see what multi-colour printing really costs: the same four-colour dragon is about 16 hours and 165 g of waste on a single head, by the slicer’s estimate, and under 3 hours on the U1.

It goes the same way every time: sketch the idea, model it, print a prototype, use it, fix it. Often I model straight on the iPad, in Shapr3D — the hive door took one morning, away from my computer. When I’m making an accessory for something that already exists, I model that thing first — the wooden hive box, here — so the sizes and positions are easy to judge.

The first print is rarely right. The MagSafe dock took four prototypes for three problems. The first, after a cassette tape, folded flat — and was far bigger than the phone. The second was cut down. The third solved the cable and the fold. Only the fourth was tidy, good-looking and close to the idea I started with.

Working isn’t enough. For me a good thing also has to feel good in the hand and look good. The drive case first opened and closed with magnets; it worked, but the motion had no satisfaction to it, so I changed it to a slide — and gave it Snorlax’s face, because I’m fond of Pokémon. And delivered isn’t always finished: once the hive doors were in use, two things showed up — the door is a little long, and the cap has nowhere to live — so I’m still working on it.

By trade I’m a production optimisation engineer, and the habits of that work come with me to the bench: measure, test, change, test again — and design for the process, not only for the part. I print on Bambu Lab and Snapmaker machines, and on a 3D printer those habits come down to a few rules.

Measure, don’t guess. Between measuring and a finished print there’s error everywhere: the measuring tool, the printer, the shrinkage of the plastic. So for the drive case I printed a small test piece for the width and the curve of the drive first, and only then put the numbers into the design. A tape measure for the desk leg; for the back of the MacBook, a magnet-detecting card, to find where Apple put its magnets.

Test small before printing big. Anything that has to fit — a pin in a map, the route printed as a separate piece, the figures that stand at the start and finish of a run — gets a small clearance coupon first, and the gap is chosen by how it actually fits, not by a number on paper.

No one should have to babysit a print. I don’t sink magnets in mid-print: it risks the layers splitting, it means stopping the printer to do it by hand, and I don’t like having to watch a print to step in. What can be printed separately is printed separately. The floor tile had pins on both faces, so 95 % of it hung in the air; I took the pins off one face, printed them on their own, and left a locating hole for each, so after printing they simply push in — no support, less plastic, less time.

When a project’s done I film it and write it up in the journal: what I needed, the options, why I chose one, what didn’t work and why. Every project here comes with its video. I hope that if you’re stuck on the same problem, it gives you one more way through.

The Flow comes from two meanings of the word. The first is those hours of complete focus when I’m making something: there’s only the model I’m drawing and the print that’s running, and by the time I look up, hours have gone by without my noticing.

The second is flow as in a stream: smooth, supple, things moving naturally from an idea to an object you can hold in your hand. That’s how I’d like every project here to go.