Sculpture 1 Process: Daphne
On this page you will see the process that brought my Daphne sculpture to life. It will run through the initial design process, prototyping journey and material research in a chronological order. This page includes all the fabrication files and code for this sculpture.
⍣RESEARCH AND IDEATION:⍣¶
Graph and more information can be found at National Environmental Education Foundation https://www.neefusa.org/story/climate-change/early-blooms-spring-how-climate-change-impacts-growing-seasons-and-you.
Early blooming cycles are a phenomena caused by climate change. As plants use a set of triggers to synchronize the timings of their growth and reproduction with their environmental conditions, the drier and warmer conditions brought about by climate change have resulted in progressively earlier blooming. This has wider consequences for entire ecosystems e.g impacting plant pollinators and animals who rely on timely production of nectar, fruit and seeds.
This video is a good summary of the wider context, as botanist and ecologist, Liana May explains how an understanding of phenology(the study of seasonal change in plants and animals) can be used to track the impacts of climate change.
I knew that I wanted to create a sculpture that would bloom progressively faster as the person approached it. I began sketching ideas and started to see Daphne on a 1 to 1 human scale with a large flower head that would meet the viewer eye to eye as it opened and closed.
I collated inspiration in a moodboard and looked at lots of timelapses of flowers to see what kind of motion I wanted.
I was quite inspired by the forms of orchids as they already had quite an otherworldly and anthropomorphic appearance and had many soft and intricate petals which make the opening and closing movement more dramatic.
I began sketching out my ideas and came up with this initials design. It is meant to be suspended at eye level and will have 4-6 moving parts around a central mechanism. I wanted it to have these long fabric arms, so that its silhouette was both reminiscent of a plant with leaves but also a figure.
Finally, I went looking for examples of mechanical flowers and started catagorising them into different types of mechanisms I would research and prototype. I had dipped my toe into this world during wearables week and had explored how I could use origami geometries to make some furling/ unfurling motions so knew someplaces to start looking in the search for relevant oribotics and linkage systems:
⏣PROTOTYPING MOTION⏣¶
From this research I selected 3 possible mechanisms to explore. I had never made anything kinetic before and my knowledge of mechanical engineering and physics is non-existent, so I decided my best way forward was to start making things with my hands and understanding through trial and error.
Gear and Rack¶
Diagram from Encyclopedia Britannica.
This mechanism is based on the movement of a gear across a rack.
However, in this case the rack moves back and forth rather than the gear. The gear is held in a fixed position above it. Depending on the distance the gear is able to move along the rack you can determine how much the gear turns. This allows the linear back and forth movement of the rack to be turned into a arc of various degrees.
To explore this mechanism I used this instructable: Solana the Sunflower. They had already made fabrication files for this mechanism. I used the .dwg files they include in their documentation to create a laser cut file and fabricate the pieces.
HERE IS IT IN ACTION:
ANALYSIS
-
I found this mechanism unreliable and reliant on a materials such as fishing line and elastic bands to create the motion which were not very resiliant and temperamental.
-
There was a lot of friction between the different parts which were designed to be fitted together very close. This created a lot of inertia in the movement.
-
The motion was hard to control and requiring a lot of force. It didn't give me hope that I could make it work with a motor if I had to keep changing the way I pulled the mechanism with my hands.
Multigear System¶
As I learnt in many of the assignment weeks, it was time to kill my darling. I moved on to exploring another mechanism I found on... you guessed it ✦ Instructables ✦.
I had been to see William Darrell's exhibition "On the Shores of Obliion" at Somer Gallery recently and was amazed by the graceful movement a multigear system could create.
I ended up finding this amazing project Kinetic Origami Interactive Art . I was drawn to this project as I had used a linear servo actuator in wearables week and had a good understanding of how this mechanism worked and could be adapted.
The mechanism is essentially a linear servo actuator that moves a plunger with gear teeth up and down. When this moves up and down it turns the gears fixed around it by hinges (much like the first mechanism) so that they move in a 90 degree arc.
I drew out my understanding and how I would adapt it to my project:
Then I got to work slicing the files with Prusa Silcer and 3D printing the pieces on the Creality Ender 3 (Ø1.75mm, Bowden) with 1.75 mm PLA Traffic White filament.
N.B I had some fit issues with the fabrication files provided and working with the tolerance of the printer. The ratios between the servo gear and pusher plunger were very off.
-
I scaled the plunger rack by 90% on the z and y axis but 80% on the X axis using my slicer software.
-
I similarly had to change the dimensions of the servo gear in rhino to fit the servo motor I was using (Tower Pro Micro Servo 9g SG90).
-
I also scaled this by 90% in Rhino, added a top covered the top of the gear and added a 2mm hole for the servo screw.
The .stl file for the adapted servo gear is downloadable on sketchfab:
Iteration 1:¶
I immediately felt like this was the right mechanism for Daphne. The movement was very smooth and predictable and could easily be modified by the length of the rack and the code and the model was simple enough that I could redesign parts and scale it to fit my design.
ANALYSIS
-
I would probably need to use a higher torque servo motor and therefore I would need to frankenstein into the file a different linear actuator e.g this one from thingiverse.
-
I really loved the 6 "arms" on the first model rather than 4, I think the blooming would be nicer and they would come together to make a more curved shape when the flower is closed.
-
The attachments on the gears were designed specifically for the 4 poles which give a very inorganic and boring shape. I would need to remodel the gears with attachments for a more complex petal design which wouldn't impediment the mechanism's movement.
Iteration 2: ¶
I began playing with the design in Rhino. I wanted to change the scale and make the mechanism so that it worked in the round. This meant adding at least another 2 gears into the system and ensuring they all still had contact with the central cog. I used the Array Polar function to make the mechanism work with 6 gears..
Here is the .stl file for the new model:
I printed this using the same filament and taking into consideration the same scaling issues from before.
Frankensteining in a linear actuator
Prior to printing I found a Linear Actuator that would fit the dimensions of the Hi Torque Servo motor I wanted to use: Tower Pro Digi Hi Torque Servo Motor MG 996R. I chose this motor due to its high torque, fixed rotation and metal gears.
The linear actuator I found was this one on Thingiverse I ensured that the rail was within the correct dimensions to fit my mechanism, then prepared and printed these files also.
I used Sency Super Glue and Book Screws to assemble the parts. This allowed for sturdy connections between them and smoother movement of the gears. I also glued some wire petals to each gear so that I could test range of motion of the mechanism:
HERE IS ITERATION 2 IN ACTION:
The final adjustments to the mechanism were as follow:
- The Gear attached to the Servo motor did not attach securly enough, perhaps from problems with the original design or the tolerance of the 3D printer. I remodelled the gear and added a hole for a small screw to be inserted and create a secure join between the gear and the motor.
- The petals needed to be attached securely in a way that didn't disrupt the movement of the gears. I measured how far the gears rotated against the central cog and worked out where it would be safe to attach wires. I then used this information to extend the gear and add holes for attaching the petals. I designed this in Rhino using the boolean and push/pull commands.
⑆ELECTRONIC INTERACTION⑆¶
Now that I had the mechanism up and running. I started to integrate the electronics and write the code.
Components
* Xiao ESP32C3 as I had used this microcontroller in other projects and found it very versitile. It is also important for my ultrasonic sensor that pins have Pulse width modulation, which many of the Xiao pins do.
* The HCSR04 Ultrasonic Sensor as this seemed to be the most well documented proximity sensor and had a 2cm - 4m range which would be plenty sufficient for my project. This module uses echo location to work out how far away an object is in its environment. The module has 4 pins: power and ground as well as Trigger (which triggers the transmitter to send out an ultrasonic pulse) and the Echo (which counts the time from when the pulse leaves to when it it received again).
* And of course the Tower Pro Digi Hi Torque Servo Motor MG 996R. Ensure you get the fixed version for this project.
* From the data sheets, I knew that the HCSR04 Ultrasonic Sensor had a working power of 5V and the servo needed between 4.8 V and 6V. Therefore, a 5V operation with a 2-3A power supply would be appropriate to power both the Sensor and the Servo.
I used these tutorials to get a basic understanding of how the ulrasonic sensor HCSR04 would work and how I could get it to interact with a servo.
I have linked Rachel De Barros's starter code here which I used to initiate and start getting readings from the sensor. I used this as a starter in Chat GPT to generate my final code:
Below is the schematic for the circuit and image of the breadboard for reference:
WIRING:
-
Connect the signal pin on your servo to A02 and the power and ground to the power rails.
-
Connect the power and ground on the sensor to the 5V and ground pin on the Xiao Board and the Trigger pin to A04 and Echo to A03
-
Connect a common ground between the Xiao Board and the power rail.
THE CODE:
This code will map the delay on the servo sweep from position 0 to 180 degrees to the distance the object is from the sensor in cm.
If the object is beyond 60cm, the servo will simply sweep back and forth very slowly.
- #include <ESP32Servo.h>
const int servoPin = 2;
const int trigPin = 4;
const int echoPin = 3;
float distanceCM = 0;
float lastValidDistance = 100;
Servo myServo;
int sweepPos = 0;
bool sweepingForward = true;
unsigned long lastSensorRead = 0;
const unsigned long sensorInterval = 50;
unsigned long lastSweepUpdate = 0;
unsigned long sweepDelay = 60;
void setup() {
ESP32PWM::allocateTimer(0);
ESP32PWM::allocateTimer(1);
ESP32PWM::allocateTimer(2);
ESP32PWM::allocateTimer(3);
myServo.setPeriodHertz(50);
pinMode(trigPin, OUTPUT);
pinMode(echoPin, INPUT);
Serial.begin(115200);
myServo.attach(servoPin);
myServo.write(0);
}
void loop() {
unsigned long now = millis();
// === Distance Reading ===
if (now - lastSensorRead >= sensorInterval) {
lastSensorRead = now;
// Trigger ultrasonic sensor
digitalWrite(trigPin, LOW);
delayMicroseconds(2);
digitalWrite(trigPin, HIGH);
delayMicroseconds(10);
digitalWrite(trigPin, LOW);
// Use pulseIn with timeout to avoid blocking too long
unsigned long duration = pulseIn(echoPin, HIGH, 25000); // 25ms timeout = ~4.3m max
float newDistance = (duration * 0.034) / 2;
// Valid reading
if (newDistance > 0 && newDistance <= 200) {
distanceCM = newDistance;
lastValidDistance = distanceCM;
Serial.print("Distance: ");
Serial.print(distanceCM);
Serial.println(" cm");
if (distanceCM <= 100) {
sweepDelay = map(distanceCM, 30, 100, 5, 50);
} else {
sweepDelay = 60;
}
} else {
// Ignore glitch reading
Serial.println("Invalid distance reading (0 or >200), ignoring.");
}
}
// === Non-blocking sweep motion ===
if (now - lastSweepUpdate >= sweepDelay) {
lastSweepUpdate = now;
myServo.write(sweepPos);
if (sweepingForward) {
sweepPos += 3;
if (sweepPos >= 180) {
sweepPos = 180;
sweepingForward = false;
}
} else {
sweepPos -= 3;
if (sweepPos <= 0) {
sweepPos = 0;
sweepingForward = true;
}
}
}
}
The videos below document the process of:
Getting the right sweep on the servo to open and close the petals.
Working out how to arrange the wire petals as to increase the drama of the movement without impedimenting the gears.
Working out the range for the mapping function that was appropriate.
☺︎MATERIAL UPGRADE AND BIO FABRICATION☺︎¶
Next it was time to work on the ✰aesthetic✰ elements of the piece.
Choosing a Material¶
First, I needed to find a material for the petals that was lightweight enough that it wouldn't put to much additional load on the motor but still have enough structure to it that it could hold a shape.
I did research into chinese millinery techniques and diy flower making tutorials on youtube such as these:
I decided that the best approach would be to create layers of heat formed organza with flower making wire (regular wire wrapped in paper) between the layers to hold the shape and attach them to the mechanism.
I did an initial test with some spare organza and hand cutting the petal shape. Unfortunately, the transparency of the material meant that you could easily see the wires and the overall look was not refined. You could also see the glue I used to adhere the 2 layers of fabric together around the wire.
I knew I would need a lot more layers to disguise the wires as much as possible and it would be better to sew the layers together for a cleaner finish.
I also knew that I wanted to colour the petals with botanical dyes so that the colours would change and translate throughout the life of the sculpture and also have a connection ot the natural world.
To use natural dyes I would need to use a natural fibre. I decided I would make some of the petals out of 100% silk organza.
I took a trip to the fabric store and looked at many options, but the silk organza had the nicest matte, semi-opaque quality to it but still had some structure. Unfortuantely, this fabric is incredibly expensive so I could only use it on the inner petals, the rest of the layers of petals would have to be a white synthetic organza which matched the colour and texture the closest. If there were not limitations to my budget, then the whole flower would be in the silk organza.
I ended up with 0.5m of silk organza and 2m of synthetic organza.
Lasercutting¶
To layer up the petals I would need them to be exactly the same shape. I would also need a large number of them to make the final petal more opaque. Therefore, Laser cutting the fabric was the obvious solution.
I sketched out some petal shapes whilst looking at various images of biotanics that inspired me and then used the SKETCH tool in Rhino to create the shapes I wanted.
You can ⍣ DOWNLOAD ⍣the .dxf file for lasercutting these petals below 1
LASER TIPS:
-
To test that I liked the shapes I first cut them in paper and was very happy with them.
-
I then proceeded to stretch the organza taunt over the laser cutting bed and securing it with paper tape. Its important that the fabric is flat and ironed before cutting so that there are as little variations in the pieces as possible.
-
I used a very low power and high speed to cut this fabric as its was basically paper and I didn't want brown singey marks on the edges!
Constructing the Outer Petals¶
With the pieces cut, I set about constructing each petal. I did some experimenting and worked out the best method of layering up each organza petal was as so:
PETAL CONSTRUCTION:
- Lay one organza petal flat.
- Cut three long pieces of Vlisofix about 0.5cm thick. Lay one in the center and then one diagonally to either side of this piece. Cut the ends at an angle so the two outer pieces meet the center piece in a V shape.
- Remove the bottom layer of protective sheet from the Vlisofix and layer a piece of baking paper over the top of everything.
- Use an iron to fix these pieces in this position.
- Peel the top protective layer from the Vlisofix. Place one piece of flower making wire on each piece of vlisofix.
- Align another organza petal over the top of all of this and secure together with a clip.
- Use the baking paper again and iron everything together.
- Your wire should be sandwiched between the two organza pieces and the 2 petals held together by the surrounding adhesive.
- Now layer up the remaining 4 organza petals, two on the top side of the piece you have made and two on the bottom.
- With a white or transparent thread, sew stitches neatly around the wire ensuring all 6 layers are held together tightly at the top, bottom and middle of each wire. I used some hair clips to hold all the layers in place whilst I sowed.
- Twist the remaining wire together to make one stem.
I was super happy with these fluffy and organic looking petals. I think the many layers and light material created a more dramatic movement and the volume was enough to hide the mechanisms.
Testing Movement¶
I attached everything together using the holes I added to the gears. I threaded the wire stems through one hole and then wrapped it round the gear before threading it through another so that it was secure.
I then tested if the motor could still move the mechanism with the additional weight and that the sensor interaction was still as intended:
All was well and I was happy with the clear movement of the motor, so I moved on to dyeing and constructing the inner petals, that I would attach to the main big gear at the center.
Natural Dyes¶
For this week I drew upon what we had learnt during Biochromes Week with Cecilia Raspanti. On my assignment page you can see a full break down of these processes and explanation of words I will use below:
With the petals I wanted subtle transitions of colour and a patchy, painted application so that they would look as organic as possible. I drew inspiration from these kinds of images, where there were heavy layers of petals and a gradient of colours in the center. This would work nicely with the blooming motion as the coloured center would be revealled as it opens.
Preparation of Fibres¶
Because I was only dying a very small amount of material for the center of the flower, I was quite imprecise with weighing the fibre. I knew I would only need a very small amount of eveything to scour, mordant and dye them.
It is not essential to Scour silk fibres and you do not want to damage them with heat, so I simply used a gentle, natural soap (ecover) to rinse the silk petals in luke warm water(it is importnat not to shock the fibres at any point in the process with extreme temperatures).
I also scoured some test pieces of cotton as well with one table spoon of sodium carbonate for 40 minutes. I did this as it is a cheaper material. It allowed me to get an idea of what the dye colours were before testing them with the silk organza.
Scouring the cotton fabric.
Mordanting¶
A mordant is a connector between the Textile and the colour, it fixes the colour to the fibres.
I used ALUMINIUM POTASSIUM SULPHATE(Aluin) to mordant both the cotton and silk fibres at roughly 15% of my weight of fibres which was again about 1 tablespoon for such a small amount of fabric added to a small pot of water.
I gently heated this for 45 minutes at a gentle heat as to not damage the silk fibres.
Dye Baths¶
I researched the colours I wanted by referencing the dye sample book that myself and Carolina Beirão made during Biochromes week.
I decided I wanted an olivey green, a range of yellows and a deep, cool purple.
For this reason I choose to make 4 dyebaths:
- Weld - cool yellow
- Chinese Rhubarb- warm yellow
- Campeche (logwood)- cool purple, can be mixed with weld to make a cool light green.
- Red Onion Skin- olive green.
I measured out my dye matter and put them in individual pots:
- Weld: 10g
- Chinese Rhubarb: 8g
- Campeche: 10g
- Red Onion Skins: 6 onions worth.
I put a small amount of water in each pan and gently heated for about 1 hours until I got the deep colours I wanted.
I dipped and painted my test fabric with the dyes and mixtures of them to test what colours I could make and which went nicely together. I tried different application techniques from folding, scrunching and sipping before washing off the dye in places to using different hard and soft brushes.
I eventually settled on a combination of a very quick dip in the weld and campeche, washing off the dye immediately and painting on a gradient of the red onion skin green and a mixture of weld and campeche at the bottom of the petal and down the paper stem.
I arranged the pieces intuitively, folding pieces together and sewing them together to make the bunch in the center.
I was really happy with how the colours turned out and the way they blended together.
Final motion test with all parts¶
✎FINAL TOUCHES✎¶
To finish off the sculpture I transfered all of the electronics onto protoboard and secured my power supply wiring. I glued the protoboard to the flat side of the mechanism, running the sensor down and under the flower head:
I also super glued on 2 arms to the flower head for attaching the long pieces of fabric that hang down in my design.
I lasercut these from 2mm white acrylic sheet and you can download the .dxf file for the arms here: 2.
I attached 2 lengths of 2m organza (a shimmery blue/ yellow colour from A.boeken) to each arm using thread and suspended the whole sculpture via fishingline.
Please see my Project Page for the final installation and interaction with the work!