Flat Belt Tire Tester · Senior Design

The customer's previous testing method placed a complete vehicle on a dynamometer roller. This made it difficult to isolate tire behavior from suspension and vehicle-system effects, and the curved roller did not reproduce a flat road contact patch.

The goal was to design a safe and cost-conscious machine that could independently test a GEM e2 tire at representative speed and load. The machine needed to support steady-state rotation and longitudinal slip testing while remaining within the available Mott Lab space and the $1,500 mechanical budget.

The mechanical team evaluated three primary approaches:

The flat-belt design received the highest weighted Pugh matrix score (3.5, versus 3.21 for the SAE dynamometer and 1.21 for the resistance roller) because it best balanced customer requirements, durability, usability, and cost.

A second weighted Pugh matrix then compared flat-belt layouts. A six-small-drum treadmill design with a tensioning drum scored highest at 3.13 and was selected, ahead of a two-small-drum system at 2.86 and a two-large-drum system at 2.64, the last of which most resembled commercial machines but would have blown the belt and drum budget.

The preliminary design used six rollers and a vertical gantry carriage. Three center rollers supported the belt beneath the tire, and a linear-bearing carriage moved the wheel vertically.

The final design changed both systems:

• Replaced the vertical gantry with a lower-cost pivot arm, then developed a parallel 4-bar after the pivot arm changed the wheel's camber during loading
• Replaced the three center rollers with a continuous support deck to reduce cost and sourced components
• Added a fully specified motor, step-up chain drivetrain, crowned drive roller, and tensioning system
• Changed the packaging from approximately 44" x 50" x 26.5" to approximately 50" x 53" x 48"
• Retained the treadmill-sourced belt and flat tire contact patch

• Selected a 5 HP, 1,800 RPM, 3-phase NAE PE184T-5-4-ODP motor
• Used the motor's 1.25 service factor for 6.25 HP of available service-factor capacity
• Calculated 4.27 HP at the typical 360 lbf tire load
• Calculated 6.11 HP at the 550 lbf upper design load
• Used a 45T motor sprocket and 32T roller sprocket for a 1.41:1 step-up
• Calculated a 2,531 RPM drive-roller speed and 26.4 mph belt speed
• Used the variable-frequency drive to limit the maximum speed to 25 mph
• Added an 18T idler sprocket on a slotted mount for chain tensioning
• Sized the #35 chain against a 194 lbf peak tension versus its 480 lbf working load, a factor of safety of 2.5

• Used an 18" long 6061-T6 aluminum tube with a 3.5" OD and 0.1875" wall
• Added a 0.030" diametric crown to center the belt
• Added an 85A polyurethane slip-on cover for belt traction
• Used bolted aluminum end caps to retain sealed R12-2RS ball bearings
• Used a 3/4" solid 1018 steel dead axle
• Chose a dead axle for cleaner shaft loading, simpler frame mounting, and easier belt replacement
• Calculated a roller bending factor of safety of 37.6 with 0.0017" mid-span deflection
• Calculated a dead-axle factor of safety of 20.1 with 0.0074" deflection
• Sized the R12-2RS bearings for a dynamic-load factor of safety of 5.5 and an L10 life of about 815 hours, roughly eight years of intermittent lab use

• Supported the belt contact patch with a 1/2" UHMW-PE wear sheet
• Fully supported the UHMW with six butted 2" x 2" x 1/4" steel box tubes
• Created a continuous 12" steel-backed contact area beneath the tire
• Removed the preliminary center rollers and intermediate steel plate
• Kept the UHMW under compression instead of allowing unsupported bending spans
• Routed the belt around one crowned drive roller and one idler return roller
• Tensioned the belt with steel cable and turnbuckles pulling against eyebolts
• Anchored the supporting welded frame to the floor

• Used a parallel 4-bar to keep the tire parallel to the belt during loading
• Extended the lower link into a 38" first-class lever
• Positioned the pivot at the center for a 1:1 force ratio
• Used a bottom-mounted 900 lbf linear actuator with a 6" stroke, running at about 61% of its rated force at the 550 lbf design load
• Produced approximately 3" of vertical wheel travel
• Used four 1" steel pivot shafts, each on two flanged self-lubricating 841-bronze sleeve bushings (eight total)
• Selected bronze bushings for slow oscillating motion, shock tolerance, and lower cost than ball bearings
• Installed a 1,100 lbf S-beam load cell inline with the actuator
• Used the 1:1 lever ratio so the load-cell reading directly represents wheel load
• Mounted the tire to a GEM e2 rear brake hub for future longitudinal-slip testing

The four-bar replaced an earlier single-lever concept for a structural reason, not just to hold constant camber. A single 2" x 3" x 1/4" lever carried the vertical load at a factor of safety of 3.6, but under combined braking that margin dropped toward the 1.5 minimum, because braking traction at the contact patch loaded the lever in weak-axis bending. The parallelogram geometry instead converts that traction largely into axial load along the arms, which the 2" x 2" x 1/4" tubes carry at a factor of safety above 100. Choosing the lever-based loading over the original linear gantry also cut the loading system's cost by roughly 90 percent, while the 1:1 ratio kept the load-cell reading directly usable with no calibration factor.

Before any FEA, I sized and checked every load-bearing subsystem with closed-form hand calculations against the project's 1.5 minimum dynamic factor of safety (2.0 target). The lowest analytical margins were the motor's continuous-torque check and the worst-case pivot-bushing bearing pressure, both wear or rotating items rather than primary structure.

A mechanical team member then performed Siemens NX FEA using the CAD I produced, to verify these predictions on the as-modeled geometry. The studies evaluated the belt frame, tire arm, and arm frame at the 550 lbf operating load and a 2,000 lbf overload condition.

All three assemblies remained below the modeled steel yield strength. The tire arm was the most highly stressed assembly, while the belt frame had the lowest stress and displacement.

I created the project BOM and purchase-order list to compare the final design against the $1,500 mechanical budget.

The purchase list totaled $1,490.37 and included:

• Bearings, fasteners, bushings, chain, sprockets, and tensioning hardware
• Custom pivot, roller, and actuator components
• 6061-T6 aluminum stock for the drive roller and end caps
• 1018 steel plate and shaft stock
• GEM rear brake hub
• Polyurethane roller cover
• Turnbuckles and eyebolts

The spend was concentrated at a few vendors: roughly $818 at McMaster-Carr for fasteners, bushings, turnbuckles, the 45T sprocket, and the roller cover; $240 for 1/4" and 3/8" 1018 steel plate from a scrap yard, cut by the shop machinist; $128 at Fabworks for laser-cut end-cap spacers, the motor plate, and pivot brackets; $115 at Online Metals for the drive-roller tube and end-cap plate; $85 for a used GEM rear brake hub from eBay; and the balance in sprockets, chain, and bearings from REV Robotics, West Coast Products, and Amazon.

To stay in budget, the project reused equipment already on hand: a treadmill belt (35% glass-filled polyester) and return roller sourced from Facebook Marketplace, the GEM tire and a Lebow 3136 500 kg load cell from Kettering University, structural tubing from campus stock, and the 5 HP motor and 900 lbf linear actuator purchased by the controls team.

• Designed physical guarding around rotating components
• Included provisions for emergency-stop integration
• Planned floor anchoring for the welded frame
• Provided mounting and packaging space for the motor, sensors, load cell, and controls hardware
• Designed around remote or separated operation during testing
• Coordinated belt speed, motor, actuator, load-cell, and wheel-speed requirements with the controls team
• Planned control-system failure modes to prevent overheating or unsafe continued operation

Physical validation is scheduled after manufacturing and controls integration.

The validation plan includes:

• Run the unloaded belt at 5, 10, 15, 20, and 25 mph
• Check belt tracking, tension, alignment, and bearing temperatures
• Apply static loads from 0 to 550 lbf in 100 lbf steps
• Test 350 and 550 lbf at both 10 and 25 mph
• Hold each combined load and speed condition for 10 minutes
• Test braking at full load and speed
• Validate the emergency stop under operating conditions
• Verify tire removal and replacement
• Measure speed and load accuracy
• Evaluate test-to-test repeatability