Automatic Litter Box 3.0 – Starting the Design Over

🐈 Automatic Litter Box 3.0 – Starting the Design Over, Keeping Every Wire

I have decided to start the whole thing over.

Version 3.0.

Before anyone panics, I am not throwing the project in a drawer and walking away. The electronics are staying. The motor is staying. The code is staying. All of that already works and I have spent far too many evenings on it to redo any of it.

What I am redoing is the physical design.

The box, the frame, the mounts, the housing around the mechanism. All of that is being drawn from scratch.

And the reason is one small technique I only half understood a few months ago.

Click to enlarge

🔥 The Thing That Quietly Changed Everything

When I built the inner box back in August, I joined the large printed sections together with a soldering iron.

You melt PLA into the seam, let it cool, and the two parts become one solid piece.

At the time I treated it as a repair trick. A way to get around the fact that my parts kept coming out bigger than the printer could handle.

The build plate on my Bambu Lab P1S is 256 mm.

For a long time that number was a hard wall. If a part did not fit on the plate, then the part was simply wrong and I had to break it up, add brackets, add screws and find a way to bolt it all back together.

Welding removed that wall.

Not reduced it. Removed it.

The build plate is no longer the limit. The design is. I can draw a part at whatever size actually makes sense for the machine, and then slice it into printable chunks afterwards and fuse them back into one piece.

That sounds like a tiny shift. In practice it changes almost every decision I make in the model.

🧱 Why Not Just Patch Version 2

The honest answer is that version 2 was built around the old rule.

Every part was shaped to fit the plate. That meant parts were split in awkward places, held together with brackets, and the brackets were held together with screws.

And screws were exactly where things went wrong last time.

If you read the earlier post, the rotating mechanism kept jamming at the same spot because a few screw heads were sticking into the gear path. That whole category of problem comes from building a big thing out of lots of little screwed-together things.

Once I stopped needing screws to hold the large parts together, a huge number of those joints just disappeared.

Patching version 2 would have meant carrying all of those old compromises forward.

It was easier to start clean.

📦 The New Inner Box

The colours in these renders are not decoration.

Each colour is a separate printed piece.

Once they come off the printer they get welded together into one rigid part, and the seams get cleaned up afterwards. So the back wall you see here is really five pieces pretending to be one.

Click to enlarge

The mounting arms on the ends used to be separate printed brackets that I bolted on. Now they are part of the same welded piece. The motor mounts sit where they are supposed to sit instead of where a bracket happened to allow.

Click to enlarge

It is a much heavier and more stubborn way to build a part.

But it is also far more rigid, and it stops fighting me every time I want to move a mounting point two millimetres to the left.

⚙ The Whole Shell Together

The render at the top of this post is roughly where the full assembly stands.

The gear drive, the NEMA 23 and the TB6600 that drives it all bolt straight into the new frame. None of that hardware changed. It was already sized and tested, so the new box is drawn around it rather than the other way round.

The flat grid panels in front are the sieve surfaces I keep going on about. The grid is the whole point: clean litter should fall straight through the openings while the clumped waste rides over the top and stays behind. Designing those as welded panels means I can make them as wide as the drum instead of as wide as the plate.

Whether the hole size is actually right is a completely separate argument that real cat litter is going to win or lose for me later.

🔌 The Part I Am Deliberately Not Touching

This is the important one, and it is the reason starting over does not scare me.

Every electronic part of this project stays exactly as it is:

• Arduino Uno R4 WiFi as the controller
• NEMA 23 stepper and TB6600 driver for the rotation
• KY-023 joystick and the 1.3" OLED for local control
• Record and replay running against Firestore, so the saved motor sequence lives in the cloud and survives a restart

All of that is already programmed and working.

I can drop the new box onto the exact same electronics, press replay, and it should run the sequence I recorded months ago.

That is the whole idea behind keeping the hardware. The movement is a solved problem. I do not want to re-solve it just because I changed the plastic around it.

🧪 Where The Project Actually Stands

• Motor, driver and controller (kept from 2.0)
• Local joystick and OLED control (kept)
• Record and replay from Firestore (kept)
• PLA welding worked out as a real building method
• New frame and inner box designed around welded segments
• Print the new box in segments and weld it
• Mount the existing electronics into the new frame
• Sieve panels tested with real litter
• Full cleaning cycle in the new design

So on paper version 3.0 looks like a step backwards. In reality almost everything on that list that matters is already ticked. I am just giving all of it a better body to live in.

🔜 Next Step: Print It In Pieces

The next milestone is not another render.

It is the first big welded part coming off the printer in segments and going back together as one piece.

I am expecting the first few welds to look rough. Some of the seams on the old inner box were genuinely ugly before I cleaned them.

That is fine.

If the part comes out rigid and the old electronics still replay their sequence inside it, then version 3.0 has already done the one thing version 2.0 never could.

My P1S is going to be busy again.