Original 36 × 8 ft frame

Iteration 01 · November 13, 2026. Preserved analysis and findings.

How much does the 36 × 8 ft dance deck bend when people gather on the party side? Explore the modeled steel frame under static gravity loads.

Roof load explorer

Engineering findings ↓

Peak member stress: 281 MPa

Peak deck drop: 3.71 in

Peak post compression: 4,810 lbf

Peak mount uplift: 387 lbf

Crowd footprint at peak stress: 12 × 4 ft, starting 14 ft from the front deck edge.

4,688 lb frame and plywood + 4,500 lb people = 9,188 lb

Stress governs at Deck edge girder. Dots mark the modeled posts.

Each readout is its maximum across the sampled placements.

Peak uplift at the rear passenger sill seat.

81% of assumed first yield. Below assumed yield in this static model. Dancing capacity and chassis attachments still require verification.

Can Earl carry 12, 15 or 20 dancers?

Twenty people with ¼-inch walls stay below assumed first yield in the gravity cases examined. The driver-side girder reaches 281 MPa against an assumed 345 MPa yield stress. Peak static deck movement reaches 3.71 in. Dancing on the parked or moving bus has no established occupancy rating.

Each row uses 225 lb per person and keeps dead load in place. The rear portal carries load in this comparison. Stress and drop are independently maximized over the sampled party-side crowd positions.

Maximum sampled stress (MPa) and deck drop (inches)
Dancers	⅛ in wall	3/16 in wall	¼ in wall
12 · 2,700 lb	321 MPa / 4.34 in	242 MPa / 3.26 in	204 MPa / 2.74 in
15 · 3,375 lb	378 MPa / 5.18 in	283 MPa / 3.86 in	237 MPa / 3.22 in
20 · 4,500 lb	454 MPa / 6.02 in	337 MPa / 4.46 in	281 MPa / 3.71 in

Why the crowd’s position matters

With ⅛-inch walls, 15 people in the centered 12 × 4 ft patch produce 331 MPa. Moving that same footprint along the deck raises the sampled peak to 354 MPa. Allowing a shorter 9 × 4 ft group raises it to 378 MPa. Twelve people in a 7.2 × 4 ft group reach 321 MPa, about 93% of assumed yield under gravity alone.

The workshop’s 20-person, ¼-inch case produces 126 MPa and 1.67 in of drop with people spread evenly. The centered party-side patch produces 264 MPa and 3.71 in. In the placement sweep, the highest stress occurs with the group starting 14 ft from the front edge; the largest drop occurs with it starting 11 ft from the front. The peak post force and hold-down demand are 4,810 lbf and 387 lbf across this sweep.

What thicker walls buy

The modeled frame and plywood weigh 2,778 lb at ⅛ in, 3,757 lb at 3/16 in and 4,688 lb at ¼ in. Increasing all frame walls to ¼ in adds about 1,909 lb to the elevated assembly. For 20 people, 3/16 in reaches 337 MPa, about 98% of assumed yield. The ¼-inch trial reaches 81%. These percentages compare unfactored elastic normal stress to assumed material yield. A design-strength and serviceability review still has to account for load combinations, dynamic response, connections and the vehicle.

Rear portal, posts and chassis loads

The source geometry has four interior posts and a front portal. The workshop adds a mirrored rear portal with an unverified chassis load path. Each portal transfers load from its two outboard posts through a sill to two assumed inboard rail seats. Each fixed foot can resist uplift in the model. Select “Absent” to see how losing the rear support changes the results.

At 20 people with ⅛-inch walls and the rear 12 × 4 ft crowd patch, omitting the rear portal increases deck drop from 1.35 to 6.37 in and peak stress from 219 to 479 MPa. For ¼-inch walls and 20 people, the no-rear-portal sweep reaches 294 MPa and 4.02 in. A bearing-only or slipping mount would redistribute force when a seat lifts. The actual rails, plates, welds, bolts and engine clearances require measurement and design checks.

The four-post design’s 100 × 60 mm edge girders are only 60 mm deep and span 21.09 ft between interior bearers. Girder bending controls the stress comparison. In the separate archived single-post geometry, the uniform 20-person case with ⅛-inch walls reached about 929 MPa in a girder. Its lower deck and different bracing make it a distinct earlier design.

Geometry, loads and solver checks

Source: Earl workshop geometry at revision 031363d, paired with Blender master version 11. The 36 × 8 ft roof includes two edge girders, 13 floor joists, eight roof diagonals, two transverse bearers, and four 130 mm square interior uprights. The modeled boxes supply outside dimensions only. The wall slider selects three solved, constant effective wall thicknesses across all modeled tube members, including posts and portal sills. Real HSS section properties, corner radii, tolerances and material specification still need verification.

The Python/NumPy model uses 3D Euler–Bernoulli beam elements, rigid internal joints and fixed rail seats. Steel inputs are E = 200 GPa, ν = 0.30 and density 7,850 kg/m³. Plywood is ¾ in at 600 kg/m³, with gravity assigned to joists and no composite stiffness credited. Touching coplanar roof intersections share nodes. Centerline offsets at seating faces are collapsed. The 7.5 mm clearance at diagonal-to-bearer crossings is left open. Checked bases can take tension. Actual joints and chassis compliance remain unknown.

The slider selects a solved integer headcount with dead load recalculated for each wall. The party-side sweep tests 4-ft-wide footprints of length 12 ft and 0.6 ft per person. The shorter footprint preserves the original 20-person cluster density of 93.75 lb/ft²; it is an explicit sensitivity assumption. The front edge of each footprint advances in 1 ft increments, plus exact centered and aftmost positions. Peaks cover sampled positions; individual footfall loads, alternative crowd shapes and every continuous position were not analyzed.

Download the Python solver and geometry to reproduce the interactive results. Each solve checks applied live load, static force and moment balance and element end forces. A 20-person reference replay checks the earlier study. Original cantilever and mesh checks validated numerical consistency. In the earlier 20-person centered ⅛-inch case, a geometric-stiffness update changed 6.02 in to 6.06 in. That check covers initial axial-force effects. The interactive results use first-order elastic service loads. Values beyond yield flag overload; their elastic deflections do not predict post-yield travel.

What still determines an operating rating

The analysis excludes rhythmic dancing, vehicle braking and turning, bumps, wind, fatigue, local tube-wall buckling, weld and bolt capacity, guardrails, and whole-vehicle stability. Rails, stair, art and unweighed equipment are omitted from the deck dead load. Modeled sound shelves attach to the body. Measure actual equipment weights and load paths. Agree on deflection, vibration and acceleration criteria for parked and moving use. Floor response to rhythmic loading depends on frequency and damping. AISC guidance on floor vibrations.

The next geometry comparison is a deeper longitudinal girder across the 21.09 ft interior span, evaluated against the same crowd cases, stiffness and mass targets. Survey the actual chassis interfaces and check connections, dynamic floor response, axle loads and vehicle stability with a qualified structural/vehicle engineer before establishing occupancy or issuing fabrication drawings.

Fabrication on hold. The next deliverable is a measured survey of the proposed supports on Earl, our 2007 Thomas HDX. The mirrored rear portal is a model gap and has no approved chassis attachments.

The 36 × 8 ft dance deck concept has two edge girders, 13 floor joists and two transverse bearers. Four interior posts and two front portal shoes show proposed attachment locations. A provisional mirrored rear portal is now positioned against the visible rear bus-body face, not the rear edge of the deck. This leaves any remaining deck overhang visible. Its two candidate rear feet and actual chassis interfaces remain unknown. Their positions determine girder spans and rear overhang. View the model in the panel workshop.

Support survey

Use one datum for all proposed support points, including the two speculative rear portal feet. Record fore-aft distance from the front axle centerline, offset from the bus centerline and height from a documented level reference. At each point, locate both the deck-to-post joint and the foot on the bus. Identify the supporting member, measure accessible dimensions and clearances, and photograph both sides with a scale.

Proposed location	Driver side	Passenger side
Front portal shoe	Awaiting survey	Awaiting survey
Forward interior post	Awaiting survey	Awaiting survey
Aft interior post	Awaiting survey	Awaiting survey
Mirrored rear portal foot · conceptual	Location / chassis load path unknown	Location / chassis load path unknown

Trace each front shoe’s load path to the chassis. At the rear, establish the actual main-rail, drop-frame, crossmember and engine-support geometry before locating any new foot. Measure the fully open engine-access envelope against the proposed rear portal, bearer, stairs and landing; record cooling, exhaust, and service clearances. The appearance of a rear foot in the model does not establish its load path or capacity.

Deliverables

Deliverable	Status	Evidence to record
As-built survey	Open	Dimensioned sketch, eight labeled photo sets including the provisional rear feet, engine and stair clearances, rear deck overhang.
Design loads	Open	Gravity screening now covers 0–20 people and three tube walls. Confirm full dead load and parked and moving occupancy. The comparison brief assumes 20 dancers at 225 lb each, moving up to 5 mph. This is a design target. Record deck, rail, stair and equipment weights and locations.
Vehicle assessment	Pending inputs	Front and rear axle scale tickets with weighing configuration, axle ratings, and review of stability under uneven loading, braking and turning.
Checked framing and connections	Pending survey and loads	Engineer-reviewed spans, sections, floor support, brace and splice details, welds, bolts, guardrail and stair connections, chassis interfaces, and operating limits.
Fabrication BOM and cut list	On hold	Record the released drawing revision and scope, then list stock grade, tube sizes and walls, lengths, cuts, plates, fasteners and quantities.

Progress

November 13, 2026 · Published the static load comparison for 12, 15 and 20 dancers and three trial tube walls. The ¼-inch, 20-person workshop sweep reaches 281 MPa and 3.71 in. The roof occupancy and fabrication decisions remain open. Explore the load cases.

November 8, 2026 · Added a mirrored rear portal as a visual-only model study, then aligned its aftmost face with the modeled rear bus body rather than the deck edge. Rear chassis attachment coordinates, body-to-chassis load path, drop-frame clearance, engine-service envelope, and dynamic torsion behavior are explicit unresolved gaps. No section sizes, joints, or fabrication release are implied.

November 8, 2026 · Reviewed the roof-attachment concept with its front portal X brace. Opened the six-point survey. All six interfaces await field measurements.