Inverted RC aero car
A model car that makes more downforce than it weighs.
- Period
- 2026 to present
- Revision
- 09 Nov 2026
- Status
- Research phase
- Stack
- AeroSandbox, NeuralFoil, CadQuery, Fusion 360, XFLR5, ESP32, MPU6050, HX711
- Inverts from
- 65km/h
- modelledExpected case. Range 58 to 77 across three scenarios.
- Downforce at 110 km/h
- 48N
- simulatedAbout 3.0 times the car's weight
- Ready-to-run mass
- 1.65kg
- modelledThe first version was 2.01 kg
Problem
A car drives upside down only while it generates enough downward force to hold on. The condition is simple: downforce has to exceed weight on both axles. Upside down, the car falls as soon as either axle runs out of grip.
Scale makes this harder. A model car runs at a low Reynolds number, where wings lose efficiency quickly. The speed that matters most is the slow one where the car just sticks, not top speed. The design target was set by a professor who previously worked at Porsche: 110 to 140 km/h.
Approach
- Wings: 2D polars from NeuralFoil at the real Reynolds number, with a Gurney flap increment, endplate effective aspect ratio, a Helmbold low-aspect-ratio correction and a ground-effect factor.
- Floor and diffuser: scenario-based average pressure coefficients (-0.35, -0.60 and -0.90). This is the biggest unknown, so every result is reported across all three scenarios.
- Aero balance: matched to the weight split, so both axles run out of downforce at the same speed. This gives the lowest inversion speed.
- Chord sizing: wing chords were sized for 60 km/h, the slow speed that decides whether the car sticks, not for the 110 km/h top target.
Design
- Chassis: Yokomo Rookie Speed RS2.0, a belt-drive 1:10 touring car with a 258 mm wheelbase.
- Body: closed-wheel prototype shell in 1 mm lightweight PLA, with louvres over the front wheel wells.
- Floor: sealed flat floor, a 9 degree rear diffuser with strakes, and TPU side skirts at 9 mm ride height. Diffusers stall below about 7 mm, which causes porpoising, so the bump stops are set at 7.5 mm.
- Wings: an inverted Selig S1223 front wing with a Gurney flap, and a single-element rear wing over the rear axle.
- Load path: wing loads go into the chassis, not the shell. Carbon swan necks carry the rear wing to the shock tower, and a carbon front mount carries the front wing to the bulkhead.
- Telemetry plan: ESP32 with an MPU6050, ride-height sensing, and an HX711 load-cell rig for static downforce checks.
Current status
The CAD model is at version 2 and the aero model is complete, but nothing has been built. The car is in the research phase. I plan to build it as my project lab next summer, once my supervisor has signed off the project.
Next steps
- CFD on the floor and diffuser, the biggest open item in the model.
- Build the wings and floor, then run the static downforce checks on the HX711 rig.
- Connect the ESP32 and MPU6050 telemetry and compare driving data with the predictions.
Figure 1
Downforce against weight
Each axle has to make more downforce than its share of the weight before the car can hang upside down. The chart shows where each curve crosses the weight line.
Figure 2
Where 48 N comes from at 110 km/h
| Source | Downforce |
|---|---|
| Floor and diffuser | 18.9 N |
| Rear wing | 17.6 N |
| Front wing | 13.2 N |
| Dive planes | 1.2 N |
| Upper body (lift) | −2.9 N |
| Total | 48.0 N |
Expected case, simulated. Car weight about 16 N.
Figure 3
Roadmap
- 01
Now
Aero design
Wings and floor modelled at the real Reynolds number, with three floor scenarios. Model at version 2.
- 02
Next
Fabrication
Wings and floor from 3D-printed parts, with carbon fibre once the process is proven on test panels.
- 03
Next
Integration
Fit the wings to the chassis, mount the ESP32 and MPU6050 telemetry.
- 04
Next
Validation
Static downforce on the HX711 load-cell rig, then driving data against the predictions.
The full roadmap has 10 phases. The four stages above are the main ones.
Figure 4