Horizontal Suspension System

Abstract

Vehicle camber is the angle of the tire with respect to the axis of the car body, and is highly influential on vehicle control. A modified design of a double wishbone suspension system is presented. The design provides vehicles with a dynamically adjustable camber angle that can induce negative camber without an external force acting on the wheel. This design feature is made possible by an adjustable spread of the upper arm, controlled by a motor. Ultimaker S3 and Ultimaker S5 printers were used for fused filament fabrication (FFF) printing, and Formlabs Form 2, Form 3+, and Form 3b+ were used for stereolithography (SLA) printing. PLA was selected to be used in the FFF printers, and clear resin and durable resin was selected to be used in the SLA printers. The AM process speeds up the manufacturing process during prototyping, allowing for quick corrections.

Introduction

1.1 Background

Suspension systems are primarily used in automobiles to absorb bumps in the road and provide the occupants of the vehicle with a safe and comfortable ride. There are a range of suspension systems used in modern vehicles, each with their own advantages and disadvantages. 

Double wishbone suspension, shown in figure 1, was commercialized by Citroën in the 1930’s (1). The main components of the suspension system are the upper and lower control arms, and the shock absorber. The shock absorber is fixed to the body of the vehicle, and the control arms are attached using pin joints. When a force is applied to the wheel, for example when cornering or passing over a bump, the spring is compressed and the control arms tilt. The shorter upper arm induces negative camber when the arms are tilted.

Figure 1: Double wishbone compressible suspension system (2)

Vehicle camber is the angle of the tire with respect to the axis of the car body. Positive and negative camber angles are shown in figure 2. Camber is highly influential on vehicle control. Both positive and negative camber angles have their own advantages. Positive camber improves the steering axis inclination and provides more stability when traveling in a straight line. Negative camber increases the contact area between the wheels and the road when cornering which increases traction. Most personal vehicles have positive camber that is set during wheel alignment because the steering has less resistance. Off-road and racing vehicles often use negative camber because of the increased stability.

Figure 2: Camber angles (3)

1.2 Design

The design is based on a double wishbone suspension system for the ability to have adjustable camber. The proposed design provides vehicles with a dynamically adjustable camber angle that can induce negative camber without an external force acting on the wheel. This design feature is made possible by the adjustable spread of the upper control arm. The spread of the control arm can be controlled remotely by a motor. The adjustable camber would allow for adjustments to the camber angle, based on the drivers preferences, without a trip to the mechanic. The design is shown in figure 3.

Figure 3: Suspension system design

1.3 Prototype Production

Additive manufacturing was chosen to produce this prototype because it is a low cost method to produce parts in a short amount of time. These features are desirable for prototyping because the design requires a large number of working parts that may require re-printing and re-design. The specific processes and materials available for each part can be selected based on their mechanical properties for optimal functionality. Additive manufacturing is also a more sustainable option compared to subtractive manufacturing because of the reduction in waste materials.

Methods and Materials

2.1 Materials

Ultimaker S3 and Ultimaker S5 printers were used for fused filament fabrication (FFF) printing, and Formlabs Form 2, Form 3+, and Form 3b+ were used for stereolithography (SLA) printing. The maximum printing dimensions of the FFF printers and the SLA printers are 330x240x300mm and 145x145x185mm respectively. The materials available for use in the FFF printers are polylactic acid (PLA), PA66/PA6 (nylon), and polyethylene terephthalate glycol (PETG). The material available for use in the SLA printers is clear resin containing urethane dimethacrylate (55-75%), methacrylate monomer (15-25%), and photoinitiator (<0.9%).

2.2 Method and Material Selection

Materials for the prototype were selected based on their material strength, thermal expansion, and printed accuracy. The final materials for each part were selected based on their mechanical properties and print performance. The mechanical properties of the available materials are shown in table 1.

Table 1: Comparison of material properties

Material Young’s Modulus (MPa) MAX. Yield Stress (MPa) Flexural Modulus (MPa) Flexural Strength (MPa) Printer
PLA 2346.5 49.5 3150 103 FFF
Nylon 580 27.8 463.5 24 FFF
PETG 1939 46.2 1882 78.9 FFF
Clear Resin (post-cured) 2800 N/A 65 2200 SLA

PLA was selected to be used in the FFF printers, and clear resin was selected to be used in the SLA printers based on the mechanical properties comparison. PLA was selected for its high Young’s modulus and maximum yield stress. FFF printed PLA was selected as the process and material suitable for bulk structures because of the ability for the material to resist deformations when under loading. SLA printed clear resin was selected for functional parts for its transparent properties that allow for viewing of the functionality.

The prototype was scaled by printing the largest part near the maximum size of the FFF print bed, in the range of 300x200x300mm. The size of all other parts were determined based on this parameter. The large prints allow for higher resolution of complex structures.

Results

3.1 Surface Finish

For the FFF processes, the surface finish was rough as anticipated. For the resin printed parts using the SLA process, the surface finish met expectations. The majority of the surfaces were smooth and operated with low friction. However, excess material was present in some spots. This excess material was particularly problematic in moving parts such as the ball-joint structure and the threaded parts. In the case where the surface finish was not acceptable for the required tolerances, post processing was considered. The post processing techniques include sanding, and polishing. The materials machining properties must be considered prior to executing these procedures to ensure that the part does not fail while being processed. Clear resin was found to respond well to post processing whereas PLA required more delicate post processing due to its thermoplastic properties. Once sanded down, the ball and socket joint operated smoothly. Post processing of the threaded parts did not respond to post processing as well.

Support structures were found to influence the surface finish and consequent joinery of the parts. Support structures are used in both FFF and SLA printers. The removal of these structures and the post print surface processing required to remove the support residue is both a time consuming process and can compromise the surface properties of the material. Minimisation of supporting structures is desirable in all cases. Through test printing, it was shown that the orientation of parts in the printers has a significant effect on the support structure requirements. The effect of orientation was more influential in FFF printers. As shown in figure 4, there may be two different orientations of the same part where one requires no supporting structures, and one requires significant supporting structures. Finding the self supporting orientation reduces print time and material requirements as well as increases the surface finish.

Figure 4: Effect of ring orientation on support structure requirements (shown in green)

3.2 Dimensional Tolerances

The initial design was produced with small dimensional tolerances between the connecting parts. Upon printing the parts it was shown that the thermal expansion of the materials should be considered when designing the dimensional tolerances. It was found that, where possible, connecting parts should be printed using the same material and additional space should be included for connections.

The initial design involved printing a screw with a triangle-shaped thread. The thread would screw into the connectors threaded insert. Upon printing, it was found that the poor dimensional accuracy of the part resulted in the screw jamming inside the connector during the rotation. Inspection of the part showed the following problems.

  • The triangular thread’s outer edge was not sharp enough. This lowered the grip between the inner and outer thread which caused the slip.
  • The thread pitch is too small because the printer’s resolution was not high enough to print the thread without blurring the design.
  • The support material residue that remained on the screw requires that the design is updated to increase the space between the inner and outer threading. However, the triangular threading does not allow large tolerances without losing its ability to grip the internal threading.

To mitigate these problems, the screw was redesigned with square threading. The initial and final screw threading are shown in figure 5a and 5b respectively.

Figure 5. a): Screw with triangular threading and b): Screw with square threading

3.3 Mechanical Properties

The PLA parts printed using FFF were able to sustain the wait of the main structure and the normal torsion forces during operation as expected. The clear resin parts printed using SLA had transparent properties and had the structural stability that was expected. Overall, the materials selected fulfilled their functional requirements as anticipated and no material failures or large deformations have been observed.

3.4 Degrees of Freedom

Upon assembly of the first prototype, it was found that the connector to the upper control arm, as shown in figure 6, was redundant as it increased the degree of freedom of the system without having a critical need. To maintain the degree of freedom, it was identified that the part must either be removed or constrained.

Figure 6: Redundant control arm connector labeled in assembled structure

The initial design included a guide to hold the screw connected, restricting its movement along Ux and limiting its degrees of freedom. The screw connector guide is shown within the assembled part in figure 7. Once assembled, and the functionality of the screw was tested, it was found that the screw connector also needs a rotational degree of freedom along Ux. The updated design is shown in figure 8.

Figure 7: Initial design with screw connector guide

Figure 8: Newly designed screw holder

Discussion

4.1 Thermal Expansion

The heat that the materials are exposed to during their respective additive manufacturing processes causes thermal expansion. Upon cooling of the part back to room temperature, shrinkage occurs. Shrinkage of parts was identified for both FFF printed PLA and SLA printed clear resin. Shrinkage was identified as the main cause for dimensional deviation from the digital model. Quantifying the shrinkage of each part and predicting the printed dimension of the digital parts is a difficult task as shrinkage involves several factors such as the printing temperature, the material used, the part shape and the printer used. Through trial and error, empirical correlations for the alterations required for the printed parts to fit like the digital parts was determined. The following tables summarizes our findings.

Diameter >=8mm & <10mm >=10mm & <15mm >=15mm
Pin -0.4mm -0.5mm -0.6mm
Hole +0.4mm +0.5mm +0.6mm

Table 2: Pin joint clearance fit

Diameter <15mm >=15mm & <20mm
Ball -0.3mm +0.4mm
Socket +0.3mm -0.4mm

Table 3: Ball joint clearance fit

Following these empirical findings, the prototype was assembled as anticipated.

4.3 Design Alterations

The following structural changes were made that made the system function properly.

  • The relative position of the connector to one of the upper control arms was fixed which constrained the connector. Essentially, the connector and the upper arm became a single unit, getting rid of additional degrees of freedom.
  • The screw guide was completely removed and replaced with a newly designed structure that allows for fluid rotation of the screw.
  • The new holder structure allows the screw connector to only have linear and rotary degrees of freedom along Ux.

During the assembly process and functionality trials, it was observed that some moving parts didn’t displace as smoothly as anticipated even with the addition of a large tolerance. This was attributed to the dimensional inaccuracy and unavoidable limitations of additive manufacturing processes. Neither FFF or SLA processes were able to produce fine details accurately. For example, in some sections, sharp edges and corners were required. However, the printed parts resulted in a fileted corner rather than a sharp one. These limitations resulted in the geometries shown in figure 9 below. The filet introduces dimensional inaccuracy and does not allow for a smooth rotation between the internal and external threading. As the screw rotates inside the threaded hole, the tolerance is not enough for the thread due to the interference of the filet.

Figure 9. a) The digital design in Solidworks and b) What the printed part looks like

4.4 Residual Material

For the SLA process, support material remains attached to the surfaces. Although most of the material can be easily removed, some of the material remains on the surface in small lumps. The lumps inside the threaded hole are particularly difficult to remove. During the rotation of the screw inside the threaded hole, interference occurs between the internal lumps and the screw which inhibits smooth movement. These problems increase the friction between the screw and the hole to the point where both parts are vulnerable to operational damage. Different designs of the thread shape, pitch and depth were tested but no solution was found.

Changing the material used for the SLA process from clear resin to durable resin appeared to mitigate the issues that we were having with the SLA printed parts. Durable resin has a larger maximum elongation  and a smaller Young’s modulus when compared to clear resin, shown in table 4. These properties allow for the printed parts to better handle small deformations while still being resistant to large displacements and high external forces. Durable resin also has a lower coefficient of friction. Unlike clear resin, durable resin does not wear as much when the parts rub together. Durable resin’s flexibility helps the structure handle deformations when inaccurate dimensions or interferences occur, avoiding deformation and failure.

Material Young’s Modulus (MPa) Ultimate Tensile Strength (MPa) Elongation at Failure (%)
Clear Resin 2800 65 6.2
Durable Resin 1000 28 55

Table 4: Property comparison of clear and durable resin

Conclusions

5.0 Conclusion

The prototype is functioning properly as a real suspension and can adjust its camber angle as proposed. Thus it is proved that this structure could be one of the potential ways of adjusting the camber angle dynamically.

Also, we find that there are many technical problems that need to be solved so that it can really be equipped in a vehicle. How to reduce the overall size would be one of the problems that needs to be solved. Currently the upper suspension arm is still very big. However, there is not so much space for the suspension under the hood of a car. 

How to hold the spinning rod is also a problem for the structure.

During manufacturing and assembling, we also find some advantages and disadvantages of the AM process. The AM process speeds up the manufacturing process during our prototyping. It is especially important for such a whole new design, since we have to change the design very often to see which design or what parameter does well. 

While we enjoy the convenience and speed of the AM process, there are still many disadvantages that need to be solved. During our prototyping, we also try to solve those disadvantages and try to make the AM process can be used in a larger range..  

5.1 Dimension accuracy and assemble

The first thing is about dimension accuracy. It is largely limited by its own principle and the shrinkage of materials during cooling down. No matter if it is the FFF process or the SLA process, which is used in our prototype, they have some accuracy limitations.

 For the FFF process, it has nozzle size (we use a 0.4mm diameter nozzle ). So, every detail which is smaller than 0.4mm can not be printed. For the SLA process, the resin is cured by UV laser beam. Although the spot of the beam is very small when it is projected to the resin, it will cure not only the resin exactly at this point but also the resin around the spot. It means that the laser beam will cure an area while we just want it to cure a tiny point. 

From above, we can easily imagine that those processes will have their natural disadvantages in accuracy. Also, the shrinkage contributes to other parts of inaccuracy. But this time, we can do some adjustments to mitigate the inaccuracy.

By experiment, we figure out a clearance fit table during design. For example the pin joint, if we make the hole a little bigger and keep the pin as the same size, the hole will shrink to just the right size to fit the pin after printing (only in idealized condition, in fact there will be some surface defects which means we should also design the pin a little bit small.) By doing that, we can make sure different parts can fit each other well in assembly. 

In fact, there are more things to consider if we want the AM process to go wider. Different AM processes will have different finishing quality and dimensional accuracy. We can not use the same standard to make parts by different AM processes. For example in our case the pin is printed by the FFF process while the hole is printed by the SLA process. How to make those parts come from different processes fit well. Even, how to fit the parts from the AM process to the parts that come from traditional manufacturing. We also think that creating a cross-process standard is also very important for using AM in a larger field.

5.2 Surface finishing and post-process

Sometimes the surface finishing of the 3D printing process is not as good as we want, for example, the FFF process or the SLS process. Post-processing will be crucial at this time. Some material is not good at post-processing or it needs some special post-processing procedure. Based on this, we also need to consider that during our design and manufacturing.

5.3 Design for manufacturing (DOM)

It is similar to the AM process to the traditional manufacturing process that we should consider some limitations of the manufacturing itself during our design. For example, we divide a whole part into many smaller parts. Such that we can choose different materials for different segments. It is not only because the whole part is too big to be printed by a 3D printer, but also, by doing this, we can optimize the efficiency of 3D printing. 

Also, some 3D printing processes like FFF are not good at printing overhanging structures. Thus, we will decrease the place of overhang of a part during design after we decide we will use the FFF process for this part. 

References

  1. 1. Double Wishbone: Derivation and History. Groupsevenpeugeot.blogspot.com. (2022). Retrieved 5 May 2022, from https://groupsevenpeugeot.blogspot.com/2012/10/dou ble-wishbone-derivation-and-history.html.
  2. Gysen, B. (2022). Generalized harmonic modeling technique for 2D electromagnetic problems: applied to the design of a direct-drive active suspension system. Eindhoven University of Technology research portal. Retrieved 5 May 2022, from https://research.tue.nl/en/publications/generalizedharmonic-modeling-technique-for-2delectromagnetic-pr.
  3. Learn Camber, Caster, and Toe | Suspension Alignments. Comeanddriveit.com. (2022). Retrieved 5 May 2022, from https://www.comeanddriveit.com/suspension/cambercaster-toe.