Skip to content

Machine Week

This weeks task was to Design a machine that includes mechanism + actuation + automation + application Build the mechanical parts and operate it manually Actuate and automate your machine Document the group project while also individually Document your individual contribution

My Contribution

In my part of machine week, I: - Designed the shaft - Designed the fully 3d printed bearings - Designed the feed mechanism - Redesigned the stepper holder with the right dimensions - Redesigned the PCB, first one used the wrong symbol for the a4988 - Coded the machine

A link to the machine week page, which includes all of the files and documentation, can be found here

Reflection

Machine week was a disaster. We went into it expecting it to take longer than we expected, maybe like a week we hoped, but it ended up taking 4. Virtually every problem we could of run into happened. We had drivers break, the first board we milled wasn't designed right, some 3d prints weren't designed right, 1/3 of the pcb simply stopped working, loosing one of our steppers, the wire we had was way too thick for a project that was supposed to take a week.

3D design

the 3d design portion was really fun. I applied out of the box thinking to all 3 designs. For the shaft, I designed it to be connected in segments, which when glued together with epoxy proved really strong. For the fully 3d printed bearings, I started by looking online, but all of them used ball bearings that failed to print every time on my machine. I redesigned those bearings to be a 2 part one with a bit more friction, but that worked flawlessly and had the exact perfect dimensions to go in the holder. For the feed mechanism, I started by trying to recreate the example project, but after it failed multiple times due to the filament not being strong enough, I based my design off nerf flywheels. This needed electrical tape to get the grip right, as I could never get it to be the perfect grip with just the 3d print alone.

PCB Design

Im proud of this design. As my first double sided board intended for milling, It used very thick traces and never had any trace failures. It did however have a failure when somebody pulled a driver out really fast, which broke a connection that we thought was secured under hot glue. Besides for that, the driver worked flawlessly and didn't contribute to any noise within the motors

Code

Im also proud of the code, as it was my first time making code for a automated machine, but I've been interested in making commercial machines 10x cheaper, and using an esp instead of a raspberry pi is a good start on cost cutting.

Problems

The main issues right now are - It's not powerful enough to bend thick wires - It requires 2 cables, microcontroller power and motor power - Its not uniform, just the assembly itself is different colored 3d prints combined on wood with really visible epoxy stains

Improvements

Im not going to sugercoat it, our machine kind of sucked. It had a novel way to control it, but physically it could barely bend the wire. The biggest improvement is to add in proper gearing. right now, the feeder struggles if the wire isn't exactly the right grip, the z axis stepper that stopped working had to be run really slowly so that it could work, and the bender portion required borderline unsafe overcurrenting to get it to bend the wire we used. If we designed this for better gearing, then none of these would be issues. We also didn't even need the extra speed, so It's just something we should of done but didn't. Another place where we could improved, epically compared to the example we based ours on, is the size. If we designed with integration in mind, the board could be hidden, and the machine could be way smaller.