Rafael Rebolleda — Fab Academy 2020 Documentation

Project Development

These are the initial high-level thoughts about the project development. I foresee these will change rapidly in time, as it often happens as we learn and, well, things happen :)

Spiral 1: Ideation

  1. Pitch the projects
  2. Learn from other projects in Fab Academy and elsewhere
  3. Look for potential challenges
  4. Check the availability of initial components

Potential Challenges

The main challenge seems to be to pass information and power through the spinning axis, both in between circles and down to the main board after the main motor.

We've got a couple of ideas of how to go about it, based on good old LP-records.

We've also given some thought on how to manufacture some of these pieces with a combinations of 3D printing, small ball bearings and a few hacks:

It also seems these devices are also available (d'oh), so it might be worth exploring the differences between the approaches.

In relation to the assignments

In general terms, these are the main parts of the project machine and how they relate to the relevant assignments. Of course other topics will impact the project too :)

Computer Aided Design

Computer Controlled Cutting

3D printing

Electronics Design

Computer Controlled Machining

Embedded Programming

input Devices

Output Devices

Mechanical & Machine Design

Spiral 2: Proof of concept

  1. Design and build a FabKit
  2. Connect the LEDs and the motor
  3. Design a simple program to test control of all components

The cornerstone of the project is the FabKit that will control all input and output, so this is where I'll put my initial efforts.

Week Assignments used for the project

Spiral 3: Design, build and test basic components

  1. Design and build the box
  2. Design and build the main circle
  3. Design and build gears
  4. Test the motor speed and torque

Week Assignments used for the project

The box

This was a fairly easy process, which can be seen in more detail in the W6: Computer Controlled Machining assignment.

The main circle

A fairly straightforward process, I tried a couple different designs:

It's super thin and light. However, I may do another one with a 15 mm stock, so drilling holes is less risky.

We also installed a bearing in the base of the box as part of the mechanical design.

On issue though is that the depth of the box is too shallow for a mobile camera to see the whole ring. It's just too close.

Therefore, I need to reconsider perspective and point of shooting. Perhaps even the shape itself... maybe a circle is not the best option. In fact, I ended giving it some thought and came up with better shaped for the final outcome:

The gears

I want to multiply the motor's rpm, so that means the motor needs a big gear and the axis that's connected to the circle will have a small one.

I used geargenerator.com to create the gears and download them as DXF for further procesing before going to laser cut.

Gears are laser-cut in two 3mm plexiglas, which are then bound together with chloroform:

The motor

The motor I was using so far, the classic 8BYJ-48 turns out to have faily slow rpms and not much torque, at least for my purposes:

After asking and looking around for a bit, I settled on the a continuous servo motor. It doesn't have speed control, but it has much more torque, the latter being more relevant for me than the former. Initial tests were promising:

It seems the pieces were coming together :)

Spiral 4: First attempt at putting everything together

  1. Communicate with the board over one spinning axis
  2. Add some simple interface elements to control the machine
  3. Consider adding a second, free spinning circle inside
  4. Design the interface

Week Assignments used for the project

This turned out to be a spiral full of failures and hard learnings. With time running out, some decisions will need to be made.

The LP tracks

With the motor and the lights testes out, I though it'd be a piece of cake to set up the "LP" tracks to communicate over the axis. Sure enough, building the board was easy enough.

Setting them up in the machine was also easy. Here it is in action with the new shape and hooked to the servo through the gears:

I though setting the tracks to communicate with the board would be an easy part in the whole project, but it ended up being pretty complicated, to the point of resulting unusable.

I did a hand operated test. You can see it sort of looks OK:

In practice, however, connection is very fiddly. It works most of the time, but it fails enough times to not be usable.

I design and implemented different approaches, but all failed differently:

I also tried different approaches on the LP itself, with cables both on top and directly on the track. Most worked most of the time, but none worked reliably.

All in all, this was a big set back, and with time running out, we decided to resort to the slip ring I had as backup.

The Plan B

This will require a different, wider shape, with different seating, rebuidling the gear socket and 3D printing the pass-through transmission, and probably many more things.

Here's an idea of how it should work:

This is quite a set back in the planning, which means I'll probably have to drop the inner ring. I'm hoping to still get good shots out of it :)

This is a new piece modeled in CAD:

And here's the final outcome:

And here it is in action:

We're back on track!

In the meantime, this is how I'm thinking about the interface:

I also started painting the interior elements in black:

Spiral 5

  1. Communicate with the board over one spinning axis
  2. Add some simple interface elements to control the machine
  3. Put everything together
  4. Create and experiment with different light and movement programs
  5. Add information and decoration stickers

Week Assignments used for the project

Finishing the ring

I cut the strip of neopixels in three different segments and linked them together. I had to do it twice because I mess up with the IN/OUT order for the strips, which is very relevant! Used black tape and black cable all over.

Building the User Interface

I run some tests of individual componentes

Following up on the design shown above, I cut a simple piece of black plexiglass to mount the buttons, pots and switches and connected all common grounds. as well as adding pull_up resistors:

Made some quick and dirty connections to test the UI

In the Input and Output assignments I go in more detail about how it works, but these are a couple of examples:

Adding a little decoration

I thoought I'd add some colour by means of the FAB logo, so I cut it in vinyl:

I also added four rails to prevent the top case from moving:

Integrating the electronics

As seen above, I also made a plexiglass case for the Hello Fabkit board so I could install it underneath the machine:

These are the I/O connections:

The Fab Light Paint Machine

A few sample videos:

A few sample stills:

Project Timeline

I thought I'd share some thoughts in retrospective, since it was such a weird time to go through Fab Academy. Below is a timeline of how the project unfolded:

Obviously, I started very late. Because of the pandemic, we were confined for a great deal of the duration of the Fab Academy, so that meant no access to the lab. Later on we had a bit of mobility, but the university was closed anyway, so it took a long time to get a window of (limited) work.

When that window opened, I talked with my instructor about different options to move forward with the project and all assignments, and we came up with a new final project that covered as many assignments as possible, including the machine, and seemd feasable with limited access to the lab (so I could do some work at home) and also with somewhat easy to buy components and materials.

I had planned (or hoped, really) to finish around a week earlier, in time for a final project presentation still in July, but a number of project setbacks and personal circumstances (time off to look after the kids and my elders) made that impossible.

As seen in the timeline, electronics design and production proved to be the most challenging aspect by far. It is the topic I've poured on more study and work of them all.

Other things required quite some work and trying out different designs, like the gears or the fixture to the frame, but weren't as initially defying as electronics design and production, for me anyway.

Other than that, I had quite a bit of a set back with the initial design to communicate the board with the frame (as described above in Spiral 4), but came up with a workaround which also led to unexpected yet interesting pieces of work. It felt frustrating at the time, and it felt like it took way more time than it actually did "physically".

That also has to do with the fact that I worked on the project many nights and weekends that I don't think I would have normally done had it been a "regular" year, so when seeing the timeline it needs to be considered that many of the days depicted there were very long in order to meet the deadlines.

Other things ended up being easier than I expected, like input / output or embedded programming.

All in all, I'd say the "worst" effect of the pandemic in my personal circumstances and project has been the limitation to try different things just to experiment, and also expand the options to try to solve setbacks. There were quite a few occassions where I just had to say "I don't have the time for more tries or approaches, I'll have to simplify and go this way". On the one hand it's sad, but on the other I think that's how "real life" projects work many times, so learning to make compromises to move forward is also a great learning.

In any case, working hard also means great joy and pride in the experience :)

Video Summary