I'm building a pen plotter from scratch (badly, in public, on purpose)

Right. Here's the thing I've decided to do. I'm building a pen plotter from scratch.

No no, not buy one. Build one. From a bare microcontroller and a bag of components and a truly heroic amount of not knowing what I'm doing.

Now, before anyone helpfully comments this: yes, I am aware you can buy one. An AxiDraw could be on my desk by Thursday, drawing gorgeous things by Friday, and I'd have learned precisely nothing except how to open a box, which is a skill I already possess quite well thank you very much.

I don't primarily want to create my plotter drawings from something pre-built. I want to know how a line in a file becomes a line on paper: the maths, the motors, the bit in the middle where a coordinate turns into electricity. And I can't buy that. I'll have to go the long way round like a wally...

So. Long way round it is... Here's my goal:

What I want at the end of this

Most plotters look like they were designed by someone who'd been formally notified that joy is not permitted in the workshop. Beige. Grey. Boxy.

I want the exact opposite. Mechanism as the whole point. Cogs and carriages on show, in colours like something that's escaped from a Bauhaus disco. Unapologetic. If it's living in my house it can bloody well earn its place as art itself.

And then, what it draws:

Generative cartography. Coastlines, constellations, contours, terrain. Maps that probably aren't trying to help you get anywhere functionally, but celebrate the art.

Geometric pattern. A compass and straightedge system that's been running maths for a thousand years and still isn't exhausted, and if that doesn't make you feel something then, well quite frankly, I'm not sure we can be friends.

Both of them are structure and beauty the same thing, rather than opposite things. Art and engineering. Both of them are pen work. A plotter is the correct machine for the job and I've decided it's my job to make one.

Flashback to 1993

I need to tell you about a very small computer.

When I was a kid I had a VTech PreComputer 1000. Sold as an educational toy; I think it had a great typing course, and trivia games, that sort of vibe. And I'm fairly confident nobody in my house hold had the faintest bloody idea that hidden inside it was a BASIC interpreter, let alone what BASIC was.

Specs (since you asked): It was a Z80 clone. Slightly under 2KB of memory. A small single line, twenty character LCD. And, crucially, no way whatsoever to save anything. If you wanted to keep a program you'd written, you wrote it out. On paper. WITH A PEN. Like some sort of bloody Victorian, by candlelight maybe.

I typed PRINT "HELLO" into it and the machine said HELLO back to me and I can still tell you exactly how that felt. The curiosity was overwhelming; even though I had no one to explain to me what this all really meant; the bigger picture. What did this mean???

I can also tell you what my brother said. Sixteen, entirely composed of football at that point, looked at a machine displaying the word HELLO on a twenty character screen, and asked what was the point of that.

Reader, at seven years old, I did not have a comeback.

I've had thirty one years to think of one, so here it is: the point, is that I made a machine do a thing. That's it. That's the whole point. And in 2026, in a world flooding with AI - it is still the point. I'm about to spend six months and an unreasonable amount of my own grown up money proving it.

Fast forward to 2000: School kept all the good bits in separate classrooms

STEM wasn't a word yet, or at least it hadn't reached my bit of the world. I went to a girls' school in Norfolk where electronics was so far off the menu, it wasn't even something you were told you couldn't do. Nobody sat me down and said "electronics isn't for you, Lisa." They didn't need to. It simply wasn't in the building. The subjects that were in the building stayed rigidly, almost religiously, in their own lanes.

Physics: equations on a whiteboard, delivered with roughly the joy of a parking ticket. Electromagnetic fields with iron fillings, yes. Bright orange ammeters, yes. Soldering, resistors, and controllers, no.

Maths: also equations but two floors up... absolutely no acknowledgement that these might be the same equations that Physics two floors down were vibing on.

Art: entire other end of the building, possibly a different postcode. Amazing, but traditional.

DT: a workshop, a bandsaw, and a brief strongly implying I make a small pine box. That the teacher then promptly tore apart in front of the class to prove a point.

Let's not even discuss IT... I'm speaking as the only generation who's first lesson was their teacher showing how to navigate to the Blue Peter website for the first time.

These subjects that are, if you squint even slightly, the SAME SUBJECT viewed from four angles. And in those eleven years not one lesson let them touched each other. That was the ninties for you.

Which is a proper shame, because all the good stuff lives in the overlap. A geometric pattern is maths you can hang on a wall. A stepper motor is physics doing something worth building. A plotter is all four at once and it is not remotely a coincidence that it's the thing I want to build. I've quilted with trig, laser cut key rings, 3D modelled and printed pikachus.

So we're here now - I'm aware I spent a very long time quietly believing the hardware end of this belonged to other people, and that belief got installed early, entirely by absence. No forty, with my own bank balance - nobody can stop me. Who makes the rules?

Turns out the thing I was interested in at nine was correct - and curiousity is a thing to persue, I recommend it to all of you.

Let's get back to the brief
I've written enough briefs for other people, received my fair few as a UX designer too. Feels only fair to write myself a proper one, so that in four months when I'm sat on the floor surrounded by stepper motors, neon cogs and belts and swearing, there's something to point at. Documentation is part of the reason I set up lab.lilypixel.

What it has to do:
Take an SVG, draw it accurately, and be able to draw it again tomorrow and get the same thing. A3 to start with. Pen up, pen down, return to origin and still know where origin is. That last one is the whole game, honestly. Anything can flail a pen about. Knowing precisely where you left it is the hard part.

What it has to look like:
Exposed. Fluoro. Laser cut acrylic in colours that would get you thrown out of a proper engineering firm. BECAUSE I CAN. Every moving part visible, because the moving parts are the reason I'm doing this and hiding them in a box would be boring.

The rules:
No kits. I can buy raw components, motors, belts, rails, bearings, the stuff you'd have to be an actual factory to make.
I can design and 3D print or laser cut them myself too.
I cannot buy anything that already knows how to be a plotter.
No pre-built motion system, no off the shelf controller board doing the clever bit for me.
I write the firmware, in MicroPython, on the Pico, even when that's demonstrably the slower and stupider option, until I know what the next step up is for myself. Especially then.

Definition of done:
A coastline, a constellation, a Strava map, a mathematical representation in art form. Plotted, framed, on my wall, and I can explain every single stage between the file and the ink without hand waving ambiguously over any of it. If there's a bit I can't explain, it isn't finished. The art continues after that.

What's out of scope:
I don't expect it to be good as an AxiDraw.
It will not be fast.
It is not a product, there'll be no shop (maybe for the art... one day), and if you want one you'll have to build your own, which is rather the point. But I'm here to chat.

So...That's the brief! Oh god... Time to go and learn literally everything.

Next post: the four things that stopped a single LED blinking for an entire week, and the plan for what comes after.