Flite Test reference reconstruction#

Reference review: 2026-09-05. These presets are independent procedural reconstructions, not manufacturing CAD or flight-test calibrated digital twins. Images establish assembly topology; full-size drawings establish dimensions where measurable. Cut outlines are not automatically assembled dimensions: foam folds, overlap, dihedral and bevels must be interpreted.

Inspect and reproduce a preset#

In Aircraft editor, select FT Bronco · V-tail, FT Bronco · Conventional, FT Tiny Trainer · Sport or FT-22 Raptor. Use Top and Side for orthographic plan comparisons, and Perspective or drag to inspect the assembly. Components shows the selected part's installation envelope. Existing local builds keep their earlier definitions: export any custom work before choosing Restore original aircraft → Apply to flight to use the revised bundled reconstruction.

Validate and exercise the FT-22 with the same core used by the browser:

npm run aircraft:validate -- ft-22-raptor
npm run simulate -- ft-22-raptor --scenario pitch-pulse --duration 5
npm run replay -- results/ft-22-raptor-pitch-pulse/recording.json
npm run physics:envelope -- ft-22-raptor

The other IDs are ft-bronco, ft-bronco-conventional and ft-tiny-trainer. For an external project, use Import JSON with the relevant file from aircraft/, inspect it, then apply it. Unapplied imports are session drafts; applying retains the custom source and edited definition locally. Export produces the standalone JSON needed by the CLI. Read the trim limitations below before interpreting a successful numerical run.

Sources inspected#

• FT Bronco build article and assembly images
• FT Bronco v1.0 full-size plan, five PDF pages: specifications and nose on sheet 1, booms on sheet 2, A-tail and conventional tails on sheet 3, wing on sheet 4.
• FT Tiny Trainer build article and assembly images
• Tiny Trainer two-sheet plan: fuselage and polyhedral option on sheet 1; powered nose, tail and sport wing on sheet 2.

The trainer preset selects the four-channel sport wing, not the three-channel polyhedral wing. The catalog’s Simple Trainer is the separate Vortex RC design; the old generic trainer is retained only as a numerical test fixture. Source images and PDF artwork are not bundled.

Bronco shared wing installation#

Both Bronco presets now seat the wing above the central fuselage and twin boom saddles, following the build photographs. The lower wing reference is raised 23 mm; fuselage and pod roof stations meet its lower skin at the authored 2° incidence. Wing mass centers, wing servos and tip contacts move with it. Crossed retaining bands and dowels are drawn over the wing; their mass is already included in the structural allocations. Vertical registration is an assembly estimate, not a dimension printed on the flat plan. The total mass and longitudinal CG remain 830 g and 51 mm aft of the leading edge.

Bronco A-tail#

Published references: 1,086 mm wingspan, 20.8 dm² wing area, 51 mm CG behind the leading edge, 640 g without battery, 830 g reference flying weight, 12° throws, and 30% recommended transmitter expo. Preset uses 203.2 mm (8 inch) props. Expo is a controller setting, not an aerodynamic coefficient; the aircraft preset does not overwrite your controller profile.

Ruler-based readings from the plan give approximately 192 mm wing chord, 254 mm flat span per A-tail panel, 102 mm tail chord, 658 mm boom length and approximately 68 mm fuselage core width (refined from the earlier 63 mm estimate). Reading precision is approximately a few millimetres, not manufacturing tolerance. At the assumed 45° assembly angle the A-tail yields approximately 359 mm boom spacing and 180 mm apex height. Those two assembled dimensions depend on the angle assumption. Outer-span ailerons occupy approximately 61% of each half-wing in the visualization; the aerodynamic surface still uses lumped control effectiveness.

Battery mass is assumed to be 190 g; its position is solved to meet the published CG. Four 9 g servo masses are accounted for. Remaining component allocations, folded nose cross-sections, thrust curves, aerodynamic polars and inertia distribution are estimated. The PDF length uses inconsistent imperial/metric values (36 in versus 927 mm); do not treat either as a precise assembled measurement.

The updated wing skin uses 5 mm board and an estimated 21 mm folded height, with a straight aileron hinge at 76.5% root chord. The sheet-3 tail hinge is approximately 192.42 / 288.06 = 66.8% chord; its inner trailing-edge relief is retained so the moving panels do not form an unbroken corner at the apex. Each nacelle/prop assembly has a 20 g allocation (14 g nacelle plus 6 g prop), split from the former 124 g fuselage allocation. The 60 g booms taper toward the tail; their estimated mass centers move 40 mm forward to represent this taper. Tail servos sit on the rear booms. These are assembly estimates, not weighed or scanned parts. Battery position is solved again so total mass remains 830 g and the longitudinal CG remains 51 mm aft of LE.

The central nose now also uses the original sheet-1 Nose / part 05 outline. The right-side cut path runs from Y = 139.420 to 712.960 pt; its inner fold is at X = 1344.86 pt. This gives a 202.33 mm projected nose length and a 131.66 mm rear height (371.64 pt). Sampled upper/lower curves replace the previous generic pointed profile. The approximately 68 mm parallel section follows the core fold separation, including foam overlap; the front retains its broad width rather than tapering to a narrow point.

The nose apex stays near body X = 327 mm. Joining it to the rear fuselage, vertical registration and the cockpit paint are photo-guided estimates. The 84 g fuselage allocation and its mass center are unchanged; its updated outer 68 × 132 mm cross-section changes the cuboid inertia estimate. Curved roof paint and side panes follow the skin, with thin white mullions. This is a procedural reconstruction of the build's silhouette, not a scan or a CAD fit.

Bronco conventional tail#

ft-bronco-conventional uses the sheet-3 H-tail / part 14 and two vertical stabilizers / part 13. The source build also shows an OV-10-style raised stabilizer; that third option is not this preset.

The traced stabilizer area is 0.04260 m²; each fin is 0.01682 m². Aerodynamic centers use area-weighted quarter-chords. Longitudinal registration to the boom remains photo-guided. Both fins are fixed: the drawing has no rudder hinges. The elevator has one 9 g servo, while yaw uses the twin motors’ differential thrust. A gliding conventional Bronco therefore has no powered yaw command; the fins still respond to sideslip. The V-tail retains its two mixed ruddervators.

Estimated conventional tail masses are 13 g per fin and 13 g for the stabilizer, with three 9 g servos overall. This retains the shared 830 g reference mass; it is not a weighed conventional build. The battery is placed longitudinally for 51 mm CG and approximately 8.1 mm to the right to balance the left-boom elevator servo. Tail lift slopes use 2π / (1 + 2 / (0.8 × AR)); polars, efficiency, control effectiveness, component inertias and motor curves remain estimates.

Default cruise and hand-release trim, lateral balance, control signs, fixed-fin behavior and model clearances have regression coverage. In the operating-point survey, the conventional version trims at 72/135 points and the V-tail at 81/135. Both fail at 6 m/s and at the power-limited 22 m/s point; the heavier conventional case also fails at 9 m/s with its authored elevator travel. These limits are reported rather than hidden by changing the aircraft’s throws. The survey uses analytical coefficients and is not flight validation.

Tiny Trainer Sport#

The plan specifies 193 g without battery, 16° throws, 30% expo and a minimum 5×3 propeller. The preset assumes a 60 g battery for 253 g total. The source recommends a 250-size 2200 kv motor, minimum 12 A ESC, 500–1000 mAh 3S battery and 5 g servos. The preset uses an estimated 3S 650 mAh pack and simplified battery discharge. Motor-current data remains estimated; ESC electronics are not simulated.

Vector/ruler readings put the flat sport half-wing around 478.5 mm with 139.5 mm root chord and approximately 32 mm aileron chord. The assembled approximation uses 5° dihedral and 953 mm projected span. Integrating the sampled tapered outline gives 0.12566 m² wing area and 131.3 mm equivalent chord. The 2.8 N motor thrust and aerodynamic coefficients remain estimates requiring bench and flight data.

The original sheet-2 lower wing skin has its CG marker at X = 523.187 pt and leading-edge fold at X = 649.187 pt. Their 126 pt separation is 1.75 in = 44.45 mm. The opposite half repeats the same offset at X = 1603.187 and 1477.187 pt. This replaces the earlier 35 mm estimate; the 60 g battery moves aft to approximately X = 124.3 mm while the airframe stays at 253 g. The later v1.1 plan cover independently lists 1.75 in / 44 mm. The current MKR2 product specification gives a 38–44 mm range. These references establish a starting balance point, not measured aerodynamic stability. Later references recommend a 6×3 prop; this reconstruction retains the original plan's 5×3 minimum configuration.

The aileron runs approximately 31–373 mm outboard on each half, leaving the swept, rounded tip fixed. Its hinge follows the plan's 90 pt trailing strip. The folded section assumes 5 mm board and 14 mm height. The horizontal tail and rudder outlines are sampled from the diagonal drawings on sheet 2, rotated into their assembled hinge axes; the resulting horizontal span is about 286 mm. Mounting tabs are omitted. Tail area and quarter-chord locations follow these outlines, but aerodynamic effectiveness remains estimated.

Main fuselage, removable powered nose and prop divide the existing 49 g allocation into 37 g, 10 g and 2 g. Horizontal tail and fin divide the 20 g tail allocation into 14 g and 6 g. Crossed retaining bands and dowels follow the assembled wing root; their mass is included in the structure. The assumed battery is placed beneath the nose to clear the motor envelope and balanced longitudinally to the nominal CG. Nose overlap, band placement, battery installation, folded heights and all individual mass centers remain estimates. The total is still 253 g. These presets do not simulate folding, fastener loads or structural flexibility.

The sport-wing servos face the lower skin, as shown in the original wing installation photograph. Their component frames follow panel dihedral/incidence, so their shafts and arms face out of the underside instead of up into the wing. Existing mass, station and case dimensions remain estimates; full skin cutouts, guide rods and insertion depth have not been reconstructed. Fuselage servo placement remains estimated.

FT-22 Raptor#

Added 2026-09-05 from the original FT-22 build and assembly photos and two-sheet full-size plan. This is the foamboard pusher park jet, not an EDF model or the Mighty Mini. The plan's 235 g dry reference plus an assumed 85 g 3S pack gives 320 g. Its two servos operate mixed elevons; the canted fins are fixed. There is no rudder command. Roll to turn and use elevator to manage the turn.

The manufacturer's listing gives both 25.5 inches and 635 mm for span; these disagree. The reconstruction uses the original PDF ruler: its one-inch rectangles span exactly 72 points (X = 231.474 to 303.474); centimetre bars independently span about 28.347 points. The traced wing is therefore approximately 649.3 mm wide at the printed scale, replacing the earlier 635 mm interpretation. This preserves the source drawing rather than forcing its outline to a conflicting retail specification. The project article contains an older F-22 photograph as well as the FT-22: the newer foam nose, intakes and side rails are the relevant assembly. The PDF's assembly inset was also inspected. No source artwork is bundled.

The current reconstruction uses these PDF vector stations (points, before scaling):

Body X points forward, so increasing sheet-1 X points aft. The main plate's front bevel lines are intake folds, not elevon hinges. A symmetric propeller opening is retained. The separate elevons join the main trailing edge with a 0.8 mm visualization gap; the two fixed fins lean outward by an assumed 18°. The nose side curve is sampled into a light faceted loft; widths, foam overlap, forward intake registration and canopy marking remain assembly estimates. The underside side plates and servo openings use the registration described below. No detailed glue beads, decals or photogrammetry are claimed.

The 400 mm nose-to-CG target is retained within the manufacturer’s 394–406 mm range. The independent body-sheet registration remains an assembly estimate. This places the CG approximately 8.9 mm behind the reconstructed main-wing root leading edge; the nose and wing leading edge are different references. The earlier 90 mm root reference was an assembly approximation and has been replaced. Wing/elevon mass positions follow polygon centroids; inertia remains a cuboid approximation. The intake and rail masses are split from the existing 60 g fuselage allocation; total mass stays 320 g. Structural dimensions, area, mass centers and inertia follow the ruler correction; 5 mm board thickness and physical hardware dimensions stay fixed. The 9-inch propeller now clears the reconstructed main-plate opening in a sampled 360° blade sweep; this does not establish physical build tolerances or flex. The battery is positioned to balance the mass ledger, not assumed to be at a measured installation station.

Aerodynamic centers use area-weighted quarter-chord estimates from spanwise polygon strips. The 9×4.7-inch prop follows the plan; 4.8 N thrust, speed falloff, current draw, servo response and all aerodynamic coefficients remain estimates. The two servos now provide the published ±40° surface limit through an assumed 50° servo command and 10/12.5 mm horn ratio. Smaller throws may leave insufficient trim authority at low speed. Delta vortex lift, separated flow and actual control response have not been measured.

Hand launches now solve a steady 8° climb at 8.5 m/s: body pitch, motor command and elevon trim follow that operating point, instead of imposing 65% power on every aircraft. This only prepares the release; controls remain fixed unless the pilot moves them. Battery discharge, weather and later throttle changes can move the aircraft away from trim. Ground mode adds the separate removable gear modification; it is not in the original FT-22 build.

Verification boundary#

Mass totals, CG references, prop diameter and elevon control signs have regression checks. All three designs trim at the default operating point, and the numerical physics report includes their definitions. These checks verify implementation consistency; they do not establish agreement with a real aircraft. Reproduce an actual build's component masses, CG, thrust-versus-command curves and recorded response before making that claim.

Ground mode adds the existing optional 45 g tricycle gear modification. It is not part of the original foamboard builds. Select hand launch or airborne mode to use the reference airframe without that modification.

Motor installation consistency#

Fixed-wing motor housings now use their linked component positions and dimensions instead of an unrelated renderer size/offset. Bronco housings use estimated 27 × 34 × 34 mm envelopes at X = 88 mm; Tiny Trainer uses 24 mm envelopes at X = 193 mm. These stations retain the existing reconstructed visual mount positions and are not measurements of a particular motor. The FT-22 uses its authored 28 × 28 × 26 mm envelope. Vents and adapters are cosmetic detail whose mass is included in the existing motor allocation. Motor thrust direction is still along the vehicle axis; component inertia-frame rotation is not motor cant.

The Bronco and Tiny motor mass allocations and total masses are unchanged. Moving their motor centers to the visible housings shifts the battery aft by 10.9 mm and 3.3 mm respectively to retain the same mass-weighted CG. This changes calculated inertia. Housing geometry is a simplified envelope, not an exact manufacturer CAD model; enter measured dimensions/positions for a known physical build.

Fixed-wing propellers use slender tapered, twisted blade meshes rather than scaled ellipsoids. The Bronco's light blade finish and vertical parked pose follow the inspected assembly photograph. Each blade uses 300 triangles and stays inside the specified rotor radius; geometry is shared between that propeller's blades. Chord, twist and section thickness remain visual estimates. They do not supply propeller polars, pitch-derived thrust or extra component mass. Specified CW/CCW spin now also controls fixed-wing propeller animation.

Separate propeller mass allocations#

All three FT presets now link a distinct prop mass component to each motor. Bronco allocates 6 g per prop from the former nacelle allocation; Tiny allocates 2 g from the powered nose; FT-22 allocates 8 g from its fuselage allocation. These are explicit estimates, not weighed propellers. Reference all-up masses remain 830, 253 and 320 g, respectively; the battery stations are rebalanced to retain the longitudinal CG references. The changed distribution updates inertia.

The prop's authored installation position drives its visible center, and motor package replacement updates both masses and the matching propulsion curve once. The component envelope represents the parked blade assembly approximately; it is not blade CAD or a spin-averaged inertia model. Catalog props can exceed the reconstructed opening. The baseline FT-22 clearance regression does not establish clearance for arbitrary replacements, flex or mounting tolerances.

FT-22 side plates and servo installation#

The original build sequence, side-plate installation photograph and servo close-up show sideways servos in the underside walls. Earlier separate rails above the wing and upright servo blocks have been replaced by that assembly.

The forward cheek folds follow the narrowing intake with an estimated fold angle. Servos face outboard at ±90° roll in their component frames. Their cases sit in the modeled openings, with estimated insertion depth and mass centers. Plastic color follows component data. The original 20 g folded assembly allocation is now floor 6 g plus two 7 g side plates; the battery is rebalanced to retain 320 g and the nose-to-CG reference. Inertia follows the revised installation.

Regression checks cover transformed servo bounds, registration, mass/CG and the side-wall distance from the full prop radius. Clearance is sub-millimetre in this reconstruction, so these checks do not prove physical tolerances. Pushrod routing, guide channels, glue and loaded linkage kinematics remain unresolved details.

FT-22 trim and power-response review#

The current geometry is closer to the drawing than its aerodynamic model is to measured flight. On the 320 g preset in calm Northfield conditions (18 °C, 120 m), the 8.5 m/s, 8° climb solution requires approximately +0.777 pitch command and 37.6% power. This corresponds to about 31.1° physical elevon trim under the current 40°/horn setup. Raising power to 100% while retaining that pitch command produces repeated large pitch excursions and an impact near 8.7 seconds in a deterministic run. The browser power-step trial showed the same pattern and an impact near 9 seconds. The trimmed-power run remained near its 8.5 m/s climb for 12 seconds.

These are observations of the estimated simulator, not a validated description of real FT-22 handling. The original plan's CG markers and the manufacturer's 394–406 mm nose reference were rechecked. They do not establish the pressure center, wing/elevon interaction or control effectiveness. The current independent surface/quarter-chord approximation needs measured trim and control-response data, and potentially a coupled wing/elevon model. Moving the sourced CG or changing coefficients solely to make the power-step trial look calmer would not validate it.

To inspect this behavior in the browser, restore the FT-22 original, select Northfield with calm wind and Hand throw, then use Chase. Start at calculated power, pause, change power and resume while leaving pitch neutral. Reset restores the calculated release controls. The setup now identifies high pitch trim; a failed solve is reported separately and does not claim a balanced release.