B747-8 FirstFlight Crash Replay

Public drone flight log, interactive replay and telemetry analysis.

Type
Quadcopter
Hardware
AET-H743-BasicV2
Software
ArduPlane V4.7.1 (dbe79216)
Log format
ardupilot-bin
Location
周村区, 中国
Flight started
2026-10-30T08:57:09Z
Duration
32s
Home coordinates
36.7502297, 117.9216788
Onboard video
Not included
Published
2026-10-31T06:34:19Z

Flight telemetry summary

Distance Meters
165.4 m
Maximum Altitude AGL
57.4 m
Maximum Ground Speed
18.58 m/s
Minimum Battery Percent
80%
Minimum Battery Voltage
18.34 V
Minimum GPS Satellites
13
Maximum GPS HDOP
0.79
Flight Modes
FBWA
Maximum Distance From Home Meters
137.7 m
Maximum Distance From Home Time
27s
Average Ground Speed
7.45 m/s
Maximum Climb Rate MPS
14.01 m/s
Maximum Descent Rate MPS
-28.95 m/s
Battery Consumed MAH
655.2 mAh
Battery Consumed Wh
12.63 Wh
Voltage Sag
6.25 V
Maximum Estimated Wind Speed MPS
13.49 m/s
Flight Mode Seconds
FBWA: 31s
Maximum GForce
2.25 g
Compass Healthy
Yes
EKF Healthy
No
Maximum GPS Accuracy Meters
0.83 m
Board Temp Max C
33.5 °C
CPU Load Max Percent
9.2%

Weather at the flight

Source
open-meteo-archive
Sampled At Utc
2026-10-30T09:00Z
Latitude
36.75
Longitude
117.92
Temperature C
17.7 °C
Wind Speed MPS
1.87 m/s
Wind Direction Degrees
290
Wind Direction Cardinal
WNW
Weather Code
0
Weather Description
Clear

AI flight report

This was a short flight — just 32 seconds — covering 165 m of track distance and reaching a maximum altitude of 57 m above home. The aircraft took off at 0:16 and climbed quickly, hitting a top ground speed of 18.6 m/s at 0:24 and its highest point at 0:29, only to show a steep descent by the end of the log: the last telemetry row at 0:30 already shows a vertical speed of -10.5 m/s, and the derived diagnostics record a peak descent rate of 28.9 m/s. Weather was mild and unlikely to be a factor, with light wind (1.87 m/s from the WNW) and clear skies. The flight ran entirely in FBWA mode (0:30 of the 0:32 total), with GPS quality good throughout (13+ satellites, HDOP never worse than 0.79).

The log shows trouble from the very start and escalating through the end. At 0:00, before the mode change, a motor imbalance warning flagged a 492 µs spread between motor 3 and motor 2 — a sign something was already asymmetric (weak motor, damaged prop, or CG issue) before the aircraft even left the ground. That imbalance seems to have worsened under load: by 0:20 the motor spread had grown to 950 µs, and at 0:22 a critical event recorded motor 1 pinned at 1900 µs, within 100 µs of its mechanical ceiling — meaning the flight controller had almost no authority left to correct attitude. Consistent with that, an extreme-attitude warning fired at 0:18 (47° from level), and the derived diagnostics show the steepest attitude of the whole flight, a full 180° from level, at 0:30 — the same moment a critical EKF_UNHEALTHY fault was logged, meaning the flight controller's own state estimate lost confidence. Battery voltage also sagged hard early on, dropping from a 24.6 V rest to 19.8 V at 0:14 (0.81 V/cell), which is a steep sag for a 6-cell pack and suggests the pack was already working hard before the climb even began. Peak acceleration of 2.23 G at 0:25 and a barometric-vs-fused altitude mismatch of 6.1 m at 0:31 both line up with this final few seconds of instability, consistent with a loss of control rather than a clean, intentional descent.

Taken together, the sequence reads as: a pre-existing motor asymmetry at idle, a hard voltage sag on throttle-up, motor 1 saturating near 0:20-22 as the mixer tried to compensate, and then an attitude/EKF breakdown in the final two seconds. The data doesn't let us say with certainty which of these was the root cause — that would need looking at raw IMU and individual ESC logs around 0:18-0:30 to see whether the motor saturation caused the attitude loss or was a symptom of it — but the pattern is clear enough to act on. Two things stand out as worth checking before the next flight: first, the 492 µs motor imbalance was present at 0:00, before takeoff, so a pre-flight motor/prop inspection (balance, damage, mount) is warranted rather than assuming it was flight-induced. Second, the voltage sag to 0.81 V/cell at 0:14 is steep for a resting-to-loaded transition this early in a 32-second flight, so the battery's internal resistance/health should be checked — a pack that sags this much this fast leaves little margin if a motor is already asking for extra current to compensate for imbalance.

Detected flight events

  1. 0s — Sustained motor asymmetry: motor 3 1491 µs vs motor 2 999 µs (492 µs spread)

    At 0:00, motor 3 is pushing 1491 µs against motor 2's 999 µs — a 492 µs spread while still on the ground, meaning something was already off-balance before liftoff.

  2. 1s — Flight mode changed to FBWA

    Right after that imbalance warning, FBWA engages at 0:01 and stays the only mode logged for the whole 32-second flight.

  3. 14s — Battery sagged to 19.8 V from 24.6 V resting (0.81 V/cell across 6 cells)

    Voltage drops from 24.6 V resting to 19.8 V by 0:14 — a steep sag that lines up with speed already climbing toward 12.7 m/s.

  4. 16s — Takeoff (altitude threshold)

    Takeoff registers at 0:16, but ground speed was already near 12.7 m/s moments earlier, suggesting a fast, running-start launch rather than a gentle climb.

  5. 18s — Steep attitude: 47° from level

    Just two seconds after takeoff, the airframe tips to 47° from level at 0:18 — an aggressive pitch for something barely off the ground.

  6. 22s — Motor 1 at 1900 µs — near maximum output, little authority left

    By 0:22 motor 1 is pinned at 1900 µs; this is near the same window (0:20) where spread hit 950 µs and satellites dipped to 13 — the mixer was maxed out trying to correct.

  7. 24s — Maximum ground speed: 18.6 m/s

    Top speed of 18.6 m/s hits at 0:24, right on the heels of motor saturation — consistent with an uncontrolled acceleration rather than a planned sprint.

  8. 29s — Maximum altitude: 57 m above home

    Altitude peaks at 57 m at 0:29, but the next sample (0:30) already shows a -10.5 m/s descent, so the climb reversed almost immediately.

  9. 30s — EKF filter fault or high innovation variance

    At 0:30 the EKF flags a fault exactly when attitude data shows 180° from level — the filter was losing track right as the aircraft flipped.

Open the interactive 3D flight replay

This is a ArduPilot flight log. ArduPilot log viewer and 3D flight replay — what RavnView reads from this format, and how to replay your own.