A Comparative Analysis of Suspension Geometries on Handling Dynamics
Advisor: Noble
Internship Location: Vanguard Auto Repair
Onsite Mentor: Josiah Liu, Owner, Vanguard Auto Repair
My interest in mechanical engineering stems from a fascination with cars, so I had to do something with cars. My experience in physics and robotics provides the foundation needed to design custom components and interpret the resulting force data.
The project involves building a modular 1:10 scale RC chassis equipped with interchangeable suspension arms. I will use a 3D printer to fabricate custom mounts for the different geometries. Testing will occur on a controlled ""test track"" where I can measure lap times and use an onboard sensor to track acceleration.
The final product will be a Functional Research Prototype and a Digital Dataset. The data will include 3D CAD files of my designs, high-speed video analysis of the suspension travel, and comparative graphs showing how each geometry handled specific terrain impacts.
Project Blog
Week 1: Project Introduction
Suspension geometry is one of the most critical variables in vehicle dynamics, yet its effects are difficult to isolate in full-scale testing due to the interference of driver input, weight distribution, and safety constraints. This project addresses that problem by applying the engineering principle of Similitude: using a precisely scaled model to investigate mechanical behavior that would otherwise be impractical or dangerous to study directly. I am examining how two fundamentally different suspension architectures, the Double Wishbone and the Trailing Arm, influence lateral stability and traction across two distinct surface conditions: asphalt and gravel.
My motivation for this project comes from a long-standing interest in mechanical engineering and automotive design. Having studied physics and worked in robotics, I recognized that a 1:10 scale RC chassis offers a legitimate and controllable platform for real engineering analysis. The central question this project seeks to answer is: How do Double Wishbone and Trailing Arm suspension geometries comparatively affect the lateral stability and tire-contact consistency of a 1:10 scale vehicle on standardized gravel and asphalt terrains?
I hypothesize that the Double Wishbone system will demonstrate superior performance on asphalt due to its more precise camber control during cornering, resulting in greater tire-contact patch consistency and lower lateral acceleration variance. However, I predict the Trailing Arm system will prove more competitive or even superior on gravel, where its longitudinal compliance and simpler geometry better absorb irregular surface impacts. Ultimately, I expect the data to show that no single suspension design is universally optimal, and that terrain type is a decisive factor in determining mechanical efficiency.
Week 2: Methodology
I have chosen a controlled experimental method, in which a single independent variable (suspension geometry) is systematically altered while all other variables are held constant. This method was selected because it allows for direct causal comparison between the two suspension systems, minimizing the influence of confounding factors such as driver error, vehicle weight variation, or inconsistent terrain. The use of a 1:10 scale RC platform further strengthens experimental control, as it enables repeatable high-speed trials in a contained environment that would be unsafe and cost-prohibitive at full scale.
Steps
CAD Design
Before any physical fabrication begins, I will design both the Double Wishbone and Trailing Arm suspension systems in Onshape. This step is essential because it allows precise geometric relationships, such as camber angle, suspension travel range, and mounting point alignment, to be verified digitally before committing to a physical build. Designing both systems on the same base also ensures that each shares identical mounting interfaces on the chassis, which is critical for isolating the suspension geometry as the sole variable.
Fabrication and Assembly
Once designs are finalized, components will be printed using high-strength PLA+ filament, with SLA resin or nylon as contingency materials if durability proves insufficient under load. The chassis will be assembled with a telemetry-capable Electronic Speed Controller (ESC), enabling consistent, measurable power output across all trials.
Bench Testing and Calibration
Prior to field testing, each suspension configuration will undergo a standardized Drop Test, in which the chassis is released from a fixed height onto a flat surface. The rate of oscillation decay will be measured using the onboard MPU6050 accelerometer to calculate the Damping Ratio of each setup. This calibration step ensures that both configurations operate within the same damping range (ideally the industry-standard "critically damped" zone) before any terrain trials begin. Skipping this step risks comparing a well-tuned suspension against a poorly tuned one, which would compromise the validity of subsequent field data.
Controlled Field Trials
Each suspension type will complete 10 timed trials on asphalt and 10 on gravel, for a total of 40 runs. The track layout will remain identical across all trials, and vehicle speed will be regulated through consistent ESC throttle settings. The MPU6050 accelerometer will log lateral G-force data continuously throughout each run. Ten trials per surface per configuration was chosen to produce a statistically meaningful sample size, allowing for the identification of consistent patterns while accounting for natural run-to-run variation.
Data Analysis and Comparison
Following field testing, accelerometer data will be compiled and analyzed to compare lateral acceleration variance, oscillation frequency, and overall stability metrics between the two systems across both terrains. High-speed video footage will supplement the sensor data by providing visual confirmation of suspension travel and tire-contact behavior during cornering and surface transitions. Results will be presented as comparative graphs, allowing direct visual identification of performance differences between geometries and across terrain types.
Limitations
Material Constraints of 3D-Printed Components
PLA+ and even SLA resin are fundamentally different materials from the aluminum, steel, or carbon fiber used in real suspension systems. Under repeated high-impact loading, printed parts may flex, fatigue, or fracture in ways that genuine suspension components would not. This means that some of the performance data may reflect material behavior as much as geometric design. While switching to stronger filament types mitigates this risk, it cannot eliminate it, and the findings should be interpreted with this constraint in mind.
Designer Inexperience and Geometric Accuracy
As a student rather than a professional mechanical engineer, my CAD designs and fabricated components may not perfectly replicate the true geometric properties of Double Wishbone and Trailing Arm systems as they are implemented in real automotive engineering. Subtle inaccuracies in joint placement, camber angles, or suspension travel range could mean that any observed performance differences partially reflect the quality of my designs rather than the inherent characteristics of the suspension geometries themselves. This limits the degree to which conclusions can be generalized beyond this specific prototype, and future iterations of this project would benefit from design review by an experienced engineer.
Week 3: Getting started in CAD
I have begun modeling the chassis in Onshape, laying out the foundational geometry that will serve as the base for both the Double Wishbone and Trailing Arm suspension configurations. I chose Onshape over Fusion for this project because Onshape better supports large assemblies with many moving parts. Getting the chassis right from the start means that both suspension systems will share identical mounting points and structural dimensions, ensuring that any performance differences observed during testing can be attributed to the suspension geometry itself and not inconsistencies in the build.
Every joint placement and mounting bracket must be carefully thought through before anything is sent to the printer, because errors at this stage are much easier to fix on a screen than on a physical part. As the chassis model develops, the next steps will involve designing the interchangeable suspension arms and running preliminary simulations to check for clearance and range of motion issues. Unfortunately, my image isn't uploading at the moment. Pictures and more updates to come as the design progresses.
Week 4: "Big things coming for my build!" he said
Progress on the project has been steady this week across multiple fronts. On the design side, I have been applying the scientific method to iteratively improve the Double Wishbone suspension assembly: printing a version, identifying what needs to change, refining the CAD model, and printing again. A few iterations have already come off the resin printer, and each one has brought the geometry closer to where it needs to be in terms of fitment, joint clearance, and structural integrity. The resin printer produces parts with high levels of detail and rigidity, enabling meaningful comparisons between versions. However, due to the print volume limitation, larger pieces such as the underbody will have to be printed on a filament printer or CNC machined from aluminum.
On the electronics side, I successfully wired the motor and servo to the RC receiver and tested them to make sure they work. I also made my first visit to a local hobby store, where I had a really helpful conversation with the person at the counter about parts selection and what components would best suit the scale and speed requirements of the project. That kind of hands-on expertise is hard to find online (looking through forums is hard work), and it gave me a lot to think about going forward. I am planning to return to the store again this week to follow up and source some additional components. They also have free pizza.
Photos from this stage of the build are attached below.
Week 5: Trailing Arm
With the Double Wishbone assembly progressing well, I've shifted my attention to designing the second suspension configuration, the Trailing Arm system. This phase has involved a lot of careful thought about how the geometry of a trailing arm differs fundamentally from the wishbone setup, and how to translate those differences accurately into a printable CAD model. Where the Double Wishbone relies on two lateral control arms to manage wheel motion, the Trailing Arm mounts behind the wheel hub and controls movement longitudinally, which changes how forces are distributed through the chassis during impacts and cornering. Getting those relationships right in the model is critical if the final comparison between the two systems is going to mean anything scientifically.
One of the more significant design decisions made this week was the choice to split the base plate into two halves since the Double Wishbone is mounted to a compact bulkhead while the Trailing Arm hinges way further from the wheels. This is a meaningful step forward for the integrity of the experiment, as a successful modular chassis isolates suspension geometry as the sole variable and reduces the chance that structural differences between two separate builds could influence the results. An All Wheel Drive drivetrain seems plausible as of now, but if I can't find driveshafts that are interchangeable for the two systems, I will have to resort to Front Wheel Drive. There is still refinement to be done before this assembly is ready for the printer, but the foundation is in place.
Next week, I am planning a visit to my mechanic shop to get some real-world perspective on how these systems are designed, built, and serviced at full scale, which should inform some of the finer design details still to be worked out.
Week 6
A few important decisions and milestones came together this week even as some physical progress has been temporarily paused waiting on a resin order to arrive before printing can continue. On the design side, a key decision was made to confirm a Front-Wheel Drive powertrain configuration for the chassis. By routing drive power exclusively to the front wheels, the rear suspension becomes a purely passive system, which significantly reduces the number of variables at play during testing and allows any differences in handling and stability to be attributed more cleanly to suspension geometry alone.
The mechanic shop visit also took place this week, and it proved to be a valuable experience. Getting a firsthand look at full-scale suspension components and talking through how these systems behave under real load conditions added useful context to what has so far been largely a CAD and research-driven process. Seeing the physical scale and complexity of both Double Wishbone and trailing arm setups reinforced several design decisions already made and raised a few new considerations to think through. In the meantime, wheels and differential components are also on the way, so once the resin arrives and those parts come in, the build should be picking up pace considerably. More updates to follow as everything comes together.
Week 7: Lost the Plot
I got the wheels and differential components from the hobby store this week, which was an exciting milestone until they didn't quite match the current designs. The physical parts introduced some dimensional and fitment discrepancies since the online shop didn't include spec sheets, meaning I had to redesign several components in CAD to accommodate the actual hardware. It is a frustrating but familiar part of the engineering process, and the ability to adapt the design around physical constraints is just as important a skill as getting the original design right in the first place. The revised components are ready to print, and the changes made should actually result in a more accurate and better-fitted assembly overall.
However, printing has hit a temporary roadblock. The resin printer has run into some technical issues that have put fabrication on hold for the time being. To keep the project moving forward without losing momentum, I will be borrowing a printer from my robotics team while the situation gets sorted out. It is a minor setback in the grand scheme of things, and the redesigned components will be ready to go as soon as printer access is confirmed. The goal is to get parts printing again early next week and get the build back on track. I'm also ordering the sensors soon.
So fried.
Week 8: Suffering From Success
The car is moving under its own power, and the first real driving tests this week went better than expected. The chassis rolls smoothly, the steering is responsive, and overall the vehicle handles well across both surface types. However, this early success has introduced an unexpected concern: the 28-turn (number of turns refers to how many times the copper wire is wrapped around each pole of the motor's armature or stator) brushed motor is generating significantly more power and speed than anticipated, even with the current high gear ratio. While a fast and capable vehicle is satisfying to see, if the car is moving fast enough to effectively skip over surface irregularities rather than responding to them, the suspension may not be engaging meaningfully enough to produce measurable differences between the two geometry configurations. In other words, the vehicle may be performing too well for the bumps and terrain variations to show up clearly in the data.
Options on the table include adjusting the gearing to reduce top speed, running trials at controlled throttle levels, or revisiting the terrain setup to introduce more pronounced surface irregularities that the suspension cannot simply power through. I'm most likely not going to have time to change the design again. Regardless of how it gets resolved, data collection is on the horizon and will begin soon.
Week 9: Data Collection Chore
Data collection is officially underway, and for the most part things are going well. The chassis is performing consistently across trials and the onboard accelerometer is logging readable, usable data, an encouraging sign after all the design and fabrication work that led to this point. As predicted, however, the motor power is proving to be a complicating factor. Even at reduced output, the vehicle moves fast enough that surface irregularities don't engage the suspension as dramatically as would be ideal, which may compress the performance differences between the two suspension configurations in the final dataset. It is a limitation worth acknowledging honestly, and one that was anticipated to some degree based on observations from last week.
On the throttle consistency issue, a practical compromise has been reached. The ESC controller has a built-in setting that caps the motor at 50% throttle, which is the lowest limit the system allows. While this reduces speed meaningfully compared to full power, it still results in a fairly fast vehicle. Ideally, the drivetrain would be re-geared to bring speeds down further and give the suspension more opportunity to respond to terrain inputs, but with the project presentation approaching, there simply isn't enough time to make that change before trials need to be completed. The 50% throttle cap will be applied consistently across all remaining trials, which at the very least ensures that speed is held constant as a controlled variable even if it isn't at the most experimentally ideal level. Data collection will continue through the week.
Week 10: The End is Near
With physical testing complete, the focus this week shifted entirely to processing and visualizing the raw sensor logs. The Jupyter Notebook script reads in the collected data and produces two key plots for each suspension configuration and terrain type: shock travel and body roll, both graphed against time ticks on the x-axis. Seeing the suspension behavior rendered visually for the first time was a satisfying moment as the curves tell a story that the raw numbers alone don't. The plots make it clear how each geometry responds to surface inputs, how quickly oscillations decay, and where the two systems diverge in behavior. As a mechanical engineering student, getting clean, readable graphs out of raw sensor data required some wrangling, but Jupyter Notebook made the iteration process pretty straightforward.
As for the results themselves, they came out largely as expected. The trends align with the hypothesis, and the data is consistent enough across trials to support meaningful conclusions. Presentation prep is nearly finished, with the graphs now integrated and the narrative tightened up. It has been a long process from the first CAD sketch to a complete dataset with Python-generated visualizations, and it is rewarding to see it all come together in the final stretch. More to come after the presentation.
Post 11: Reflection
Looking back on this project as a whole, the thing that stands out most is how much was accomplished by keeping the scope focused. The modular chassis design was one of the best early decisions I made, as building a single vehicle with interchangeable suspension arms rather than two separate builds kept the experiment clean and the variable isolation intact. The Python analysis pipeline also worked out well, as having a repeatable script that could parse raw sensor logs and generate consistent graphs made the transition from data to presentation smooth and professional. And on a broader level, the use of Similitude as the theoretical foundation gave the project academic legitimacy that a more casual build-and-test approach would have lacked. It grounded every design decision in real engineering reasoning.
That said, there are things that would be done differently given another pass at this. The motor and gearing selection deserved more research upfront: the speed issue was manageable, but it introduced a limitation that affected the quality of the data and required a workaround rather than a solution. For anyone continuing this work, the first recommendation would be to plan the drivetrain around the terrain, not the other way around. The second would be to seek a professional or experienced engineer to review the CAD geometry before printing. Student-designed suspension components are a reasonable starting point, but independent validation would meaningfully strengthen the findings. The significance of this research lies in demonstrating that Similitude-based scale testing is not just an abstract engineering concept; it is something a motivated student can actually execute with accessible tools and produce real, interpretable data from. That has value for anyone in robotics, automotive engineering, or physics looking for a rigorous project framework.
As for where this work could go next: re-gearing the drivetrain is the most immediate extension, as it would allow terrain inputs to more fully engage the suspension and produce cleaner performance differences in the data. Beyond that, this same platform could be used to test additional suspension geometries, different spring and damper configurations, or even active suspension control systems. The chassis and the data pipeline are both reusable. More broadly, the Similitude methodology applied here could be extended to other disciplines such as structural impact testing, aerodynamic studies, or even robotics locomotion research. On a personal note, getting to spend a trimester doing a structured, rigorous project on something I am genuinely passionate about has been one of the more rewarding academic experiences I have had. It is one thing to study engineering and another to actually do it.