First foiling day (17 October 2026)

The first foiling day is the hydrofoil's first time on the water under autopilot control, planned for 17 October 2026. This page is the plan for the day: what to check at the dock, the four stages on the water in order, when to stop, and what to read in the logs between runs. Read the pilot guide first; this page assumes the driver knows the status light.

The goal of the day is modest on purpose: check that the height sensors hold up under way, then get the first real unstick and trim. Every number below marked sim is a prediction from the software-in-the-loop simulator, not a measurement. Replacing those predictions with real numbers is the most valuable result of the day, more than a long flight.

Stop rules

SOLID RED LIGHT = CUT THE THROTTLE
Come off the foils first, ask why afterwards. A self-clearing alarm still means: come in and read the log before the next run.

Stop the run and come off the foils on any of these:

• any alarm (steady red with a long blast), or steady red without beeps;
• a roll oscillation that grows from one swing to the next;
• ride height more than 0.15 m off target, or a foil ventilating;
• blinking red with counted beeps: follow the pilot guide. Do not foil on 3, 4, 5 or 6 beeps.

Before the day (at the dock)

• Upload hydrofoils.lua rev 27 and foil_tune.lua protocol v11, power-cycle for the reboot gate (2 beeps), then confirm HYD_SEED = 27, HYD_RSCALE = 1.0 and HYD_FRNTFF = 0.944. Update the keyboard tuner and the display at the same time, or the tuner will flag a version mismatch.
• Take a fresh parameter dump (tools/fc_backup.py). The backup file in the repository is from July.
• Sort out the compass / yaw. It has never been calibrated, and in the August logs the attitude estimate's heading was stuck. An unhealthy attitude estimate blocks foiling.
• Check directions with the boat on its trailer: a nose-up pitch demand gives the front foils more angle, and a roll demand moves the correct wing.
• The rear foil stepper (on the rudder controller) homes, and jogging it moves the foil the way the setpoint says.
• Measure the real rear foil angle at two or three setpoints. The rear-trim logic assumes the commanded angle is the true angle.
• The height sensors read about 0.114 m at rest. Transducers clean and dry (at the Hungary race 70–90 % of readings under way were missing).
• Weigh the driver and the boat. The rear-foil schedule assumes 180 kg all-up.
• Logging is on: the flight controller logs HYDC, HYDR, ATT and RCOU, and the datalogger records the CAN bus.
• The driver knows the light: blinking green = ready, steady green = foiling, solid red = cut the throttle.
• Pre-launch check: flip the enable switch off and on and see the light change between blinking and steady. Launch only on green.
• Coordinated-turn banking stays off all day (TRN_ENABLE = 0).

On the water

Do the stages in this order. Each one has a time box and a go / no-go check; do not start the next stage until the current one passes.

Stage 1: displacement, foil control off (about 30 min)

Speed steps at 11, 14, 18 and 22 km/h, about 30 s each, then one full-throttle run.

This is mainly a sensor check under way. On the Zaan the boat already reached 26–28 km/h with the struts mounted, so the top speed should match that. The sim needs 23–27 km/h to unstick, so the margin is real but thin.

Go / no-go: top speed close to the Zaan 26–28 km/h · height sensors give readings under way · attitude estimate healthy · no alarms.

Stage 2: height mode on, below 18 km/h (about 10 min)

Enable at the dock and run up to about 15–17 km/h. Nothing can lift at this speed, so this is a free check of the whole chain: the rear foil moves with speed, the alarms stay quiet and the light stays green.

Go / no-go: rear foil follows speed · no alarms, no yellow.

Stage 3: wheelie (about 20 min)

Set SCR_USER1 = 1 (roll-test / wheelie mode: height loop off, fixed nose-up pitch). Start around 15 km/h, add speed until the bow lifts, and stay below 22 km/h. Give small roll steps (SCR_USER3) and pitch-trim steps (SCR_USER2); see Tuning parameters.

This stage is short on purpose. It does not teach much about foiling, but it is the cheapest way to catch the errors most likely to wreck a first flight: a pitch or roll sign the wrong way round, an unstable roll loop, or pitch that does not follow its demand. All of it happens with the stern still in the water. Do not tune roll to its best wheelie value: the stern adds damping that disappears on lift-off, and the flying gains are about twice the wheelie optimum.

Go / no-go: bow goes up when asked · roll settles without growing swings. If roll oscillates, halve the roll gains.

Stage 4: first takeoffs (rest of the day)

Set SCR_USER1 = 0. Flat water, ideally in the morning, with a chase boat and enough depth for foils hanging about 0.6 m below the hull.

• Start with HYD_TARGET = 0.22 m (the default is 0.30). Further from the struts breaching, still clear of the 0.15 m floor.
• Push gently: about 1–2 km/h per second, up to just above unstick. Full throttle on the Zaan was about 6 km/h per second; a slow push gives the smallest fly-up.
• Hold a few seconds, then ease off.

Change one thing per run, in this order:

1. hold longer;
2. target 0.26 m;
3. target 0.30 m;
4. a harder push;
5. small throttle changes while flying;
6. a full-throttle takeoff (last).

Key numbers

Value	What it is
23–27 km/h	Unstick speed (sim)
26–28 km/h	Top speed with struts mounted (Zaan logs, August)
18 km/h	Below this the controller treats the boat as not foiling
0.22 → 0.30 m	Ride-height target ladder for the day
≈ 0.56 m	Ride height where the struts breach
0.15 m · 0.114 m	Lowest allowed target · resting height reading
1–2 km/h per second	Gentle push for the first takeoffs
≈ 6 km/h per second	Full-throttle push measured on the Zaan
2 °/s	Rear foil stepper speed
180 kg	All-up mass the rear-foil schedule assumes

Reading the logs between runs

Each of these checks a model assumption that cannot be verified any other way.

Look at	What it tells you
Unstick speed	Whether the lift model and HYD_RSCHED are right
HYDR.Bias after a steady stretch	How far the rear-foil schedule is off on the real boat
Pitch and front-foil servo at steady trim	Whether the front foils sit near their cruise trim, as the speed feed-forward assumes
HYDC.I and HYDC.Deg against the +2 / −3° clamp	Whether the height loop is running out of authority
HYDR.Lag	Whether the 2 °/s rear stepper keeps up
Rangefinders against the EKF height, in flight	Whether the height estimate holds up at speed
Motor current flying vs displacing at the same speed	Drag saved by flying, and the margin for takeoff

If it won't take off

• More speed, or less mass.
• Lower HYD_TARGET; it costs nothing.
• More rear lift at low speed: raise HYD_RSCHED. This brings back a bigger fly-up, so push gently.
• Do not raise the height gains. In the simulator, more gain never helped the fly-up.

If it flies badly

• Roll oscillates after lift-off: turn the roll gains down, and note the speed it starts at.
• Big fly-up: push more gently first, then give the height loop more nose-down authority (HYD_CMDMIN).
• Overreacts or lags on throttle changes: HYD_FRNTFF is the live knob.
• Rides steadily off target: let the rear-foil offload settle (about 20 s), then check HYDR.Bias.

Notes

Speeds are converted from m/s at ×3.6 and rounded. Tune values are from hydrofoils.lua rev 27 (6 October 2026). The tuning parameters themselves, with their ranges and hotkeys, are on Tuning parameters.