IS-431: Base Mechanical Design

Mechanical Design

V-Model Traceability: This page documents the physical realisation of the base, addressing RM-3 (1-person portability) through transport hardware design and load calculations, and supporting RM-1 (Structural stability) through frame material and fabrication choices.

Frame Structure & Fabrication

The base frame is constructed from 70 × 50 mm Rectangular Hollow Section (RHS) mild steel tubing, joined by MIG welding. RHS was selected over solid bar, angle iron, or channel for several mechanical and practical reasons:

High second moment of area per unit weight: Hollow sections resist bending and torsion more efficiently than equivalent-weight solid sections, which is crucial for resisting overturning moments from punches.
Flat external faces: Enable straightforward mounting of the rigid top plate, brackets, and lower mechanisms without requiring extensive machining.
Weldability: Mild steel RHS is straightforward to MIG-weld, providing strong, reliable joints for the trapezoidal geometry.
Availability: SHS/RHS is readily available in standard sizes from local steel suppliers.

Wall Thickness Rationale

The RHS members are specified with a wall thickness of 3 mm. This sizing provides an optimal balance between total mass and structural integrity:

Weld penetration: Ensures adequate fusion depth at butt and fillet welds, avoiding the risk of burn-through or cold welds common on thinner-walled sections.
Mounting loads: Safely supports bolted connections for the flat mounted plate and mechanisms above without localized buckling or tearing at the bolt holes.
Impact resistance: The tubular sections must resist transmitted punching loads without permanent deformation at connection points.

Portability & Deployment

BoxBunny must be moved between storage areas, fabrication workshops, and public demonstration venues. The transport strategy was driven by requirement RM-3 (Portability): the base must allow transport by 1 person between locations without compromising the planted operational stance. The transport solution must not require disassembly of the robot and must function effectively on typical gym and corridor surfaces (e.g., rubber mats, concrete).

Planted vs Transport Stance

A key design principle is that transport features must not compromise the planted operational stance.

Planted: During training, the base rests flat on the floor with its full trapezoidal footprint in contact with the surface. Anti-slip pads were pasted at the bottom during assembly to maximise friction and stability.
Passive Wheels: The transport wheels are positioned slightly elevated relative to the base feet. They do not contribute to the support polygon or introduce rolling compliance while flat.
No Locks Required: This dual-mode behaviour eliminates the need for brake locks, deployment levers, or other complex mechanisms that could fail under punching impacts.

Transport Hardware

The base features two rear-mounted wheels located at the widest rear edge of the trapezoidal footprint, with parts sourced from MISUMI:

Component	Part Number	Material & Dimensions	Function
Transport wheels	RMNA100	Rubber, Ø100 mm	Smooth rolling on gym and corridor surfaces when tilted backward.
Wheel mounting pins	KCLSBF12-62	Stainless Steel (SUS304 eq.), Ø12 mm × 62 mm	Secure axle retention; standard clevis-pin format for easy replacement.

Transport Load Capacity Check

To ensure the transport hardware does not fail during tip-and-roll handling, a worst-case load check was performed. The maximum structural stress occurs when the robot is tilted to its balance point, placing the entire system weight onto the two rear wheels and their welded brackets.

1. Wheel Load & Factor of Safety

Assuming a fully assembled robot mass of 105 kg, the total weight (W) is:

              W = 105 kg × 9.81 m/s² ≈ 1030 N
            

When balanced entirely on the rear wheels, each wheel supports half the weight:

              Fwheel = 1030 N / 2 = 515 N
            

The MISUMI RMNA100 rubber wheels have an allowable static load of 1200 N per wheel.

              Factor of Safety (FoS) = 1200 N / 515 N ≈ 2.33
            

2. Clevis Pin Shear Stress

The 12 mm solid stainless steel pin is mounted in double shear. The effective shear area (A_shear) is twice the cross-sectional area:

              Ashear = 2 × (π × 6²) ≈ 226 mm²
            

The shear stress (τ) on the pin is:

              τ = 515 N / 226 mm² ≈ 2.28 MPa
            

With SUS304 stainless steel having a shear yield strength of approximately 150 MPa, the pin is massively over-specified and immune to shear failure under these loads.

This calculation confirms that both the 100 mm wheels and the welded steel mounting hardware are structurally sound for tilting and rolling the fully assembled robot without risk of yielding or wheel collapse.

Design Iterations

Phase 1: Early Robotics Fair implementation

The first implemented base configuration was used during the NUS CDE Robotics Fair in first half of semester. At this stage, the primary objective was safe and rapid deployment using materials that were immediately available. The robot was mounted on a 1220 × 580 mm wooden board, which served as the initial support platform for the early exhibition-stage build.

Interact with the 3D model to explore its structure. Toggle annotations to view key components.

Hide Annotations

_{Interactive 3D Model: Early Robotics Fair Wooden Board Base}

Although this wooden board was not intended to represent the final engineered base, it provided the first opportunity to assess the relationship between footprint, body placement, stability, and user working space. It allowed the robot to be assembled and demonstrated safely in public and gave the team an early reference point for space claim, support behaviour, and practical handling.

However, this arrangement had several limitations. The footprint was availability-driven rather than optimised, the load path into the floor was not intentionally structured, and the board occupied too much uniform plan area near the user. It also did not integrate naturally with wheel brackets, welded hardware, or a central mounting datum. This made it acceptable for early deployment, but not appropriate as a final engineered base.

Phase 2: Stability-driven rethinking of the footprint

Once the base was treated as an engineered subsystem rather than a temporary support, the key question became how the robot could remain stable without making the front working area too bulky for the user. This led to the shift toward a trapezoidal footprint. Instead of enlarging the base equally in all directions, the design narrowed the front and widened the rear. This was the most suitable mechanical compromise because the front remained compact where the user interacts most closely, while the rear gained support width where stability leverage is most useful.

Phase 3: Transition from flat platform to framed steel base

After the footprint logic was clarified, the next decision was how to realise it physically. A flat plate or board could provide area, but not a very clear structural load path. The project therefore moved toward a welded steel frame using rectangular hollow sections. This brought several mechanical advantages: better bending stiffness for the material used, clearer load transfer into the floor, easier bracket and subsystem integration, and a lower profile than a thick monolithic support. The chosen structural members were 70×50 mm rectangular steel tubes with 3 mm wall thickness.

The section size provides enough depth to act as a true structural frame member while still remaining close to the floor. The 3 mm wall thickness is a balanced prototype-scale choice because it is thick enough to be practical and tolerant in welding, more robust near local bracket and mounting regions than lighter-gauge tube, and still avoids unnecessary mass and fabrication burden. This is a sensible middle-ground choice for a welded prototype base.

Phase 4: Mounted plate as a structural datum

As the lower-mechanism stack matured, it became clear that the base did not only need to support weight; it also needed to define a geometrically reliable interface for the mechanisms above. This made the mounted plate critical. The mounted plate therefore had to be both sufficiently rigid not to flex significantly under subsystem loads and sufficiently flat not to induce misalignment into the rotation subsystem. To achieve this, the plate incorporates precise mounting holes for the rotation system to be directly secured. Additionally, the plate is screwed onto the steel tubes rather than welded; this modular attachment method allows for the easy integration of additional electronic component mounting plates. At the current stage, the mounted plate is confirmed to be flat and rigid, which is an important success in the mechanical-design process.

Phase 5: Integration of transport intent

The next iteration step was to incorporate the transport concept directly into the base rather than treating movement as an external logistics problem. This led to rear interfaces intended for wheel hardware and tip-and-roll handling. This decision is mechanically sensible because the wider rear already provides the best location for wheel integration, the rear is naturally farther from the user footwork zone, and moving the robot by tipping from the rear is compatible with portable gym-equipment handling logic.

Phase 6: Durability and corrosion thinking

Once the welded steel base became the final direction, corrosion had to be treated as part of the base design. Early observations of rust on steel members showed that both interior and exterior surfaces needed realistic protection. A key insight was that interior and exterior corrosion conditions are fundamentally different: exterior surfaces are accessible and paintable, whereas interior hollow sections are inaccessible and more prone to trapped moisture. This led to a dual-path strategy, with interior sections relying on forgiving, penetrating protective coatings and exterior surfaces relying on mechanical cleaning, anticorrosion primer, and paint. This is good mechanical practice because it recognises the real maintenance limits of hollow welded steel structures. Consequently, professional priming and powder coating were explicitly budgeted and integrated directly into the fabrication process.

Final Design State

The final base subsystem can therefore be summarised as follows:

it began as a 1220 mm × 580 mm wooden-board prototype base for early fair deployment,
it evolved into a purpose-built welded trapezoidal steel frame after the need for better footwork clearance, clearer load paths, and more intentional subsystem integration became clear,
the final frame uses 70 × 50 × 3 mm rectangular steel tubes,
the base keeps a narrower front and wider rear for user clearance and stability,
it incorporates a flat, rigid mounted plate as the geometric interface for the lower-mechanism stack,
it includes rear transport intent.

Hardware Component Specifications

Component	Specification	Function
Frame Tubing	70 × 50 × 3 mm RHS Mild Steel	Provides the main structural load path and resists overturning moments from punches.
Mounted Plate	Flat Rigid Steel Plate	Acts as a geometric datum and solid mounting interface for the lower-mechanism stack.
Transport Wheels	RMNA100 (Rubber, Ø100 mm)	Enables smooth rolling on gym and corridor surfaces when the base is tilted backward.
Wheel Mounting Pins	KCLSBF12-62 (SUS304, Ø12 mm)	Ensures secure axle retention; standard clevis-pin format allows for easy replacement.

Understanding this structural foundation is critical because it dictates how forces flow into the ground. When a punch lands on the elevated upper body, the resulting overturning moments and radial forces must be transmitted down through the entire mechanism stack and ultimately into this welded base geometry. The next section, Load & Stability Analysis, details how these structural loads and moments were quantified to ensure the robot resists forward tipping and remains completely stable under worst-case boxing conditions.

Prev: Design Ideation

Next: Load & Stability Analysis