In the Ni Lab, I was encouraged to test the waters in many different divisions of the operation. I learned about fabrication, using computational models such as DIC for shape sensing, and developing apparatus to test the soft robots themselves. I would like to share the process of development for a couple of sub projects I was working on.
Pneumatically Actuated Shape Sensor - Mould
PBAT (short for polybutylene adipate terephthalate) is a biodegradable random copolymer, specifically a copolyester of adipic acid, 1,4-butanediol and terephthalic acid. We were tasked to use this polymer to create a prototype that changes its shape when pumped with air.
Such properties are only plausible when there is a hollow design created. Moulds for hollow structures are extremely complicated to design and fabricate.
Usually, the pull-away test that tests the ability of a design to be mould pressed fails when it comes to the design of the mould itself. 3D printing turns into a precarious task with these designs.
Nevertheless, a system was developed to use a base 3D printed structure upon which void filling polymer components were added.
Isometric View
Objective
To develop a mould that could create pneumatically actuated shape sensors. Upon actuation, the sensors this mould would create would change its shape which would then be used for further analysis.
Process
We first evaluated the current method of creating this mould. The mould was laser engraved into acrylics using a CO2 laser. This method led to the development of several gaps in the sensor's final render.
Second, we developed the mould using CAD tools to make a completely independent mould fabrications system. This closed mould would assist in reducing the many holes that arose from the former prototyping process
Third, we checked the sanity of the final product by printing the mould several times to ensure if the intended result was met or not.
Design
For a mould, where we want air to fill must remain hollow. This means that when the liquid-state polymer is poured, we want the a structure to cordon off the flow into the areas where air is supposed to develop. This is difficult to develop since the middle piece would be left suspended to noting, creating an irregular design.
The solution to this is visible as the small tooth-pick like structure that is supporting the inner moulds that make the sensor hollow. A side view of the prototype is provided adjacent.
Linear Actuator
To test this sensor with pneumatics essentially means to fill the mould with small increments of air. Each unit volume that is added will change the shape of the soft robot prototype - measuring this is the most important part of the project we've been a part of.
To create a system that adds small increments of air, we wired the soft shape morphing prototype to a syringe that pumped in air. Moving the pump of the syringe to let air go in extremely controlled increments was the task at hand. We resorted to solving this with a linear actuator.
I wished to utilise a stepper motor - since their movements could be precisely controlled. Translating rotational motion to a linear motion took me back to middle school physics - the concept of an inclined plane wrapped around a thin rod.
I then drew a design that mounted a screw to the motor, add a sleeve that held the nut for the screw to move it inwards and outwards. This inward and outward motion pushed the arm in and out.
The CAD image above shows the motor mount, the hex bit that mounts the screw to the motor, and the pushing arm. The reason for this design the minimal dependency of non-essential 3D printed parts. Popular prototypes of the linear actuator do not have the moving components exposed in the open - this is to protect the design from various external interferences. However, in a controlled environment such as the lab, the bare-bones structures was extremely functional and possessed much utility in our requirements.
Push Arm
Comments