It flies again, falls — and the battery sings at ~15.8 Hz.
Serious rebuild, first real hovers, then a four-meter fall. In the logs, the real culprit: 840 grams of battery making the structure sing.
rebuild brass standoffs + threadlocker flights first hovers held crash #2 ~4 m · throttle cut in panic battery mass 840 g · amplitude ×18 (soft mount) resonance 15-19 Hz at rest → ~15.8 Hz in flight
Serious rebuild: brass standoffs, threadlocker everywhere. And there they are — the first real hovers. The drone holds.
That same evening, a throttle punch that’s too hard sends it to four meters. I panic, I cut the throttle, it falls. The printed GPS tail breaks.
In the logs, a shake keeps coming back. An A/B test, firm mount versus loose: the amplitude is eighteen times higher on the soft mount. The culprit this time isn’t the pilot — the battery (840 grams, the heaviest part) excites a mode of the structure. Rigid rebuild: the resonance moves from 15-19 Hz to ~21 Hz at rest, but in flight the mode climbs back toward ~15.8 Hz. So we document everything: the measurements, the decoupling hypotheses, what we try on each run.