hero-image
HOME
hero-image
project-highlight-image

3D Printed Beam Strain Rate Analysis

hero-image
Riley Ovalle

Project Timeline

Oct 2025 - Nov-2025

OVERVIEW

Designed and constructed a data acquisition system to measure the effectiveness of five different PLA infill patterns against strain and creep under a constant load. Key functionalities include the integration of a Wheatstone bridge strain gauge, an LM358B operational amplifier, and an Arduino R4 board for continuous analog data logging. The system utilized a custom-fabricated hanger assembly to apply a constant 500-gram load for 72 hours. The results identified that the Honeycomb pattern was the most resistant to deformation, exhibiting a maximum strain of approximately 0.007.

HighlightS

  • Designed and implemented a data acquisition system to measure strain and creep, integrating an Arduino R4, Wheatstone bridge strain gauge, and operational amplifier
  • Conducted a three-day, controlled experiment on five different PLA infill patterns, applying a constant 500-gram load and logging 4,320 analog data points per beam
  • Determined the optimal material performance, finding that the Honeycomb infill pattern exhibited the lowest maximum strain (approximately 0.007), providing the greatest resistance to deformation

SKILLS

SolidWorksArduino IDECircuit Design 3D Printing Data Acquisition Data Analysis Error Analysis

SUPPORTING MATERIALS

Additional Details

Problem Statement

The conventional manufacturing processes using assembly fixtures  lack the precision required for manufacturing miniaturized ultrasound devices at high-volume production. This is required to fit inside a small catheter. Achieving the necessary +/- 15 micron accuracy for critical assembly steps, including UV adhesive dispensing, part alignment, and clamping, often relies on manual or semi-automated methods. These methods are prone to human error, time-consuming, and inconsistent, leading to increased scrap rates, higher production costs, and limitations in scaling manufacturing output. Therefore, there is a critical need for an automated solution that can consistently deliver ultra-high precision across multiple assembly stages to enable reliable and efficient manufacturing of these advanced miniaturized components.

Bill of Materials (BOM)

The following table lists the components used in the prototype, including part numbers, quantities, materials, estimated costs, and potential suppliers.

Item

Component

Part Number

Qty

Material

Cost ($)

Supplier

Notes

1

Heat Sink

637-20ABPE

1

Aluminum

25.00

McMaster-Carr

100x100x50 mm, extruded aluminum

2

Cooling Fan (12V, 40 mm)

AFB0412SHB

1

Plastic/Metal

10.00

DigiKey

Low-noise, 35 dB max, 12V DC

3

Device Housing

Custom (3D-printed)

1

PLA

15.00

University 3D Print Lab

FDM-printed, 200x150x100 mm

4

Thermal Insulation Foam

851-074

0.5 m²

Polyurethane Foam

8.00

Amazon

Cut to fit optics compartment

5

Temperature Sensor

DS18B20

1

N/A

12.00

Adafruit

±0.5°C accuracy, digital output

6

Arduino Uno

A000066

1

N/A

25.00

Arduino Store

Runs Python PID via serial interface

7

Fan Muffler

Custom (3D-printed)

1

PLA

5.00

University 3D Print Lab

Reduces fan noise

8

Fasteners (Screws, M3)

91292A112

10

Stainless Steel

3.00

McMaster-Carr

M3x10 mm, for securing components

9

Thermal Paste

AS5-3.5G

1

Silicone-based

5.00

Amazon

Improves heat transfer to sink

Motor Review

Basic PID Controller Script

PYTHON
import time class PIDController: """A PID controller for precise control in robotic systems. Attributes: kp (float): Proportional gain. ki (float): Integral gain. kd (float): Derivative gain. setpoint (float): Desired target value. prev_error (float): Previous error for derivative calculation. integral (float): Accumulated integral term. dt (float): Time step in seconds. """ def __init__(self, kp: float, ki: float, kd: float, setpoint: float = 0.0): """Initialize PID controller with gains and setpoint. Args: kp: Proportional gain for error response. ki: Integral gain for accumulated error. kd: Derivative gain for error rate of change. setpoint: Desired target value (default: 0.0). """ self.kp = kp self.ki = ki self.kd = kd self.setpoint = setpoint self.prev_error = 0.0 self.integral = 0.0 self.dt = 0.01 def compute(self, current_value: float) -> float: """Compute PID output based on current system value. Args: current_value: Current measured value of the system. Returns: float: Control signal to adjust the system. """ # Calculate error error = self.setpoint - current_value # Proportional term p_term = self.kp * error # Integral term self.integral += error * self.dt i_term = self.ki * self.integral # Derivative term derivative = (error - self.prev_error) / self.dt d_term = self.kd * derivative # Calculate total output output = p_term + i_term + d_term # Update previous error self.prev_error = error return output if __name__ == "__main__": # Initialize PID controller pid = PIDController(kp=1.0, ki=0.1, kd=0.05, setpoint=10.0) # Simulate mode (e.g., motor position) current_value = 0.0 for _ in range(100): control_signal = pid.compute(current_value) # Simulate system response: position updates based on control signal current_value += control_signal * 0.1 print(f"Current Value: {current_value:.2f}, " f"Control Signal: {control_signal:.2f}") time.sleep(pid.dt)


lowinertia
Portfolio Builder for Engineers
Created by Aram Lee
© 2025 Low Inertia. All rights reserved.