Overstimulated Tofu (Bio-Cube)

Overview

A responsive biomaterial cube that turns human proximity into a visible consequence. As viewers approach, the cube’s breathing mechanism shifts and deforms, revealing how observation and overconsumption can overwhelm a living system.

Keywords: Biocomposites, Fusion 360, Bambu Studio, Microcontroller, Actuators

Team Members: Bverly Yip with help from Yesuel Song, Matt Griffin, Alana and Audrey Oh (shop staff)

My role: Creating Biocomposite, 3D Modeling, Linear Actuator

Status: Completed, GitHub

Contents:

Mechanical Mechanism

Making the Bioplastic

3D Modeling & Printing

View Final Media & Interaction

Process

Mechanical Mechanism

Mechanical Mechanism: We were set on having the mechanism housed inside the box. We explored several internal movement systems, including offset motors, solenoids, gyroscopes, and vibrational motors. Ultimately, we chose to adapt an existing mechanism rather than build one from scratch, prioritizing reliable movement that could be fully enclosed and hidden.

Prototyping linear actuators

Approach 1: We 3D printed a linear actuator, but the gear mechanism kept getting stuck. Printing precise gear systems that rely on a fine space to work from existing files was difficult.

Approach 2: We pivoted fully to a scissor linear actuator. But during testing, friction caused inconsistent movement.

Making the Bioplastic

We want to make the cube out of an agar agar bio composite material. The tests and findings are as follows.

Approach 1 Agar Agar Glycerin Sheet:

Recipe:

8 grams of Agar
13.5 grams of Glycerin
400 ml of water

Observations & Objective:

We were able to achieve a full sheet with minimal strinkage. Goal now is to test how to make the sheet thicker and more rigid so it could hold a box mold instead of just decorating a structure.

After about a week of drying

Approach 2 Agar Agar, Glycerin, Psyllium, Calcium Carbonate Mold Box:

Recipe:

Water: 180 g
Agar: 5–6 g
Glycerine: 6–8 g
Calcium carbonate: 8–12 g
Psyllium 1 g (for crack resistance)

Casted into a small, cube mold

Observations & Objective: The solid mass (~11.5%) was too low. Air drying exposed the bottom, pulling on the walls and causing cracks. Future iterations need slower drying in the first 24 hours, then demolding and continued air drying. Stronger fillers may also help.

After about a week of drying

After over a week of drying

Approach 3 Agar Agar, Glycerin, Wood Flour Filler:

Recipe:

400ml water
16g agar
10g glycerin
40g of wood flour

Observations & Objective: Solid mass (~14%) was still too low. Target should be 25–30%. Low solids caused shriveling and warping as water evaporated. The texture also became too rigid for our needs.

Picture right after casting

Picture after the sheet has dried

Approach 4 Back to Agar Agar, Glycerin, but Cast to Cube Mold:

Recipe:

400 ml water
8g Agar
18g glycern
Gauze around the center of the cube

I put a bowl over the mold after casting to let it evaporate slowly for the first 24 hours

Observations & Objective: Rapid drying caused warping, but a partially wet cube worked well because the walls remained flexible. Drying can be controlled by storing in a sealed bag in the fridge.

24 hours after casting, initial demolding

After drying for a bit in the induction oven

Approach 5 Agar, Glycerin, MCC, Calcium Chloride: Last ditch effort to try fillers:

Recipe:

600 ml water
15g agar
20g glycerin
30g MCC
30g Calcium Chloride
Gauze around the center of the cube

Observations & Objective: This produced the best balance of flexibility and strength

Untitled spreadsheet

Casting to Silicone Molds. Our options were limited.

Our mechanism worked well in these molds, but

Problem: The molds didn't suit our vision because we wanted it to be a cube

Solution: scale down the mechanism (which we tried to avoid).

In the meantime, I cast a cube in our ideal size and slowed drying by storing it in the fridge for more control.

Note: Since I did notice that the previous test strips/molds were starting warp, I decided to leave our cube mold in the fridge to significantly slow down drying so we could have more control over it’s life span

3D Modeling & Printing

We switched to a smaller SG90 servo to fit the box’s tight spatial constraints and confirmed it had enough torque to drive the mechanism. However, balancing smooth scissor motion, wall force, and fit within the cube remained difficult. After feedback, we paired the compact mechanism with a larger servo for stability and torque, modifying the mounting and reducing mechanism height to make everything fit inside the box. We scaled down the mechanism in Bambu.

However, through playtesting, this mechanism proved unpredictable and difficult to scale within the box’s constraints. Rather than continuing to iterate on unstable tolerances, I learned Fusion 360 and switched to a rack-and-pinion mechanism.

Delicately carving out the hole for the proximity sensor

We battery-powered the circuit off the breadboard

First Iteration: We added a hole for the LiDAR sensor and programmed: Idle “breathing” motion, Light wall pulses, and Expansion when someone approaches, contraction when they leave

Second Iteration: I used an ultrasonic sensor. I programmed three behaviors: breathing when recognized, rapid expansion when approached, and slow continuous expansion when someone gets too close.