Week 18 - Project Development

There are many tasks that needed to be done. We list the tasks ordered by spiral indicating what has been completed. We then list the tasks per spiral, and finally we discuss the planning.

Spirals

The full idea is to have a track with a lift mechanism actuated by driving a car onto the track. The car will be picked up by a lift mechanism, after which it will move down by gravity, making a u-turn, jumping while a camera takes a picture of the jump. It will come to a stop at a rotary platform that can hold about four cars that each can be launched onto the track again.

Below a subdivision of the spirals:

• Track and lift mechanism
• U-turn
• Make the car jump
• Create a rotary platform to store / launch cars
• Add a camera for a picture of a car jumping

Tasks

Below a list of the tasks for the first spiral:

• Investigate wood and the laser cut settings.
• Investigate flexures and the laser cut settings.
• Investigate parameters for press-fit.
• Investigate belt system and tensioning belt.
• Model parts for the lift mechanism.
• Investigate radius curve in the track.
• Model pillars to create an elevated track.
• Laser cut the parts for the track.
• Design the electronics for the sensor.
• Produce the electronics for the first sensor.
• Experiment with / develop the code for the sensor.
• Produce the electronics for the second sensor.
• Design the electronics for the lift mechanism.
• Produce the electronics for the lift mechanism.
• Develop the code for to test the lift mechanism.
• Integrate the car sensors with the lift mechanism.
• Test the lift mechanism actuation.
• Fine-tune the lift mechanism actuation.

A list for the second spiral, the u-turn:

• Investigate u-turn flexures.
• Model radius/width u-turn flexures.
• Model the rail of the u-turn flexures.
• Model the connectors between pillars and rail.

A list for the third spiral, a jumping car:

• Investigate radius curve for jumping.
• Design the ramp.
• Design the platform to land.
• Experiment with the car landing.
• Integrate with the rest of the track.

A list for the fourth spiral, the rotary platform:

• Investigate the radius in relation to parking cars
• Investigate the launch mechanism of a car.
• Investigate the height of the platform in relation to relaunching the car.
• Model the launch platform.
• Design the electronics for the rotary platform.
• Produce the electronics.
• Laser cut the launch platform.
• Integrate with the rest of the track.

A list for the fifth spiral, the camera system:

• Measure the speed of the jumping car.
• Find a solution for the camera to move with this speed.
• Create a rail system based on the Beehive project.
• Make a choice for a stepper or a DC motor.
• Design the electronics for the camera system.
• Design pillars with lighting for the camera.
• Produce the electronics for the camera system.
• Laser cut the pillars for the camera system.

Planning

Since I switched projects and invested too much in other topics in the weekly assignments, I had only three weeks to develop all of the above. The tasks as listed above are more or less listed in a dependency order, so for example, before designing the track, I have to know to what extent the flexures are doable. Tasks that are independent is the car sensor work in relation to the track work.

What worked well

There are several things that worked well: The flexures and press fit parameterization worked really well. For the press fit I have to make a note: for track pieces coming from the same sheet of wood, the parameters and adjusting them for press fit in the sheet worked well. However, I made use of two sheets of plywood and they both required different clearances for the press fit. This resulted in less than ideal connections between track pieces from different sheets of plywood.

The belt system and tensioning the belt worked well, but I had to adjust the trajectory of the belt a bit to tip over the car at the top. Because of how I wrote the software, it was easy to experiment with the speeds of the stepper motor.

The scripted geometry of the pillars worked well: I can make pillars for any height. This includes the cross bars that provide stability.

The electronics for the lift mechanism and the car sensors worked well, both the design and the production. There is a small issue with the car sensors where one of the sensors gives readings that are not so good as the other one, that behaves more as I expect. Luckily, it was possible to solve that in software. The software to detect cars worked well.

Remaining Questions

One of the remaining questions is whether I can improve the lift actuation mechanism. I may need to simply continue to fine-tune it or perhaps I need to change the design a bit. For example, I may need to improve the clip at the end of the belt that picks up the car.

Another question is why one of the sensors does not behave similarly to the other that behaves more or less how I expect. I could simply create a new sensor board to compare it with the existing two.

A very interesting question is whether I can find press fit parameters for different plywood sheets, even though the difference between them is small. (One was 3.50 mm, the other 3.60 mm.) I guess, this will simply mean laser cutting test pieces from one sheet, and then from the other and find a match that works for both sheets and works for "within" the sheet. I understand now that it would have been better if I had taken this approach, and estimating how much wood I needed.

I mainly learned that I should indeed have worked continuously on the final project. I think that this is what Fab Academy assumes, but I think it would perhaps be better to provide students a bit more guidance if the idea of the final project does not become more concrete after some weeks. I wouldn't advise anyone to try to finish the final project in the last three weeks because this will be very difficult when having setbacks, especially with those that I had.

I didn't like the approach we took in the machine week, just experimenting without having a 3D model, so I vowed to have a proper 3D model because this is the base on which you can start to iterate. I still stand by this, but I did not anticipate that CAD could be a barrier as well, as happened with me when it turned out that it the design was much more challenging than initially assumed. Because press fit designs rely heavily on very precise assemblies, there was not much I could do for a plan B, I simply needed to have the designs, and if the designs were right, laser cutting was in principle fast. In essence, I laser cut the whole track one day before the final presentation and I had quite some confidence in the outcome because I knew the design was more or less good.

Perhaps the most important lesson is that I underestimated the complexity of adding flexures to press fit and the fact that the belt system required many custom parts. In hindsight, I believe I based my experience on week 3 where I created the cardboard construction set. After I had the design, the laser cutting went very well and fast, but I never got to the stage that the race track I designed became a construction set because of all the custom parts around the lift mechanism.

Finally, I must conclude that project planning in the sense that Neil explained has never really worked for me. I had the same problem now that I had earlier when I was really determined to try supply-driven time management, which was in the molding and casting week. I remember making a schedule for which task I wanted to do when, but the first task was designing the face mask. I couldn't complete that in the time frame I planned for that, but everything was dependent on that. So, in short, I couldn't keep to the schedule and the same was true for the final project.

Since everything depended heavily on a design with an assembly and that component turned out to be more complex than anticipated becoming a bottleneck, there was not much I could do except for simply continuing the design.

I guess, I miss a discussion of the "critical path" and task dependencies in Neil's lectures about time management. Perhaps I simply missed it. It could also be that Neil or others naturally make sure that they start projects without too much heavy dependencies, but if this is the case, then that should be a more important lesson that how you manage your time. I think I will think more about this issue that I experience with time management in the future.

Fab Academy

• List the tasks completed.
• List the task that remain.
• Discuss what has worked.
• Discuss what hasn't worked.
• Discuss the remaining questions.
• Explain the planning.
• Reflect on the project.