# PhysicsFlight — Instrumented baseline 001

Source: Betaflight Blackbox logs btfl_001.bbl through btfl_003.bbl; processed with documented decoding and low-command screening.

## Provenance and configuration
This report publishes the owner's supplied independently decoded trial summaries. The raw .bbl files and decoder implementation are not included here; no new raw-log decoding is claimed by the website. Exact trial dates are unavailable. These instrumented recordings are distinct from the qualitative October 17, 2026 first flight.

Aircraft: BETAFPV Meteor75 Pro Matrix 1S. Firmware: Betaflight 2025.12.5. Angle Mode. Added payload: 0 g (not total aircraft mass). Approximately 407 Blackbox samples/sec. All three logs end cleanly; zero corrupted frames.

## Metrics and screening
- Overall gyro RMS: sqrt(mean(gyroADC_X^2 + gyroADC_Y^2 + gyroADC_Z^2)), in physical degrees/second (deg/s).
- Gyro p95: 95th percentile of combined three-axis angular speed, in deg/s.
- Low-command screen: throttle 1140–1400 microseconds inclusive AND absolute pitch, roll and yaw rate setpoints each <=10 deg/s. Low-command RMS uses the same RMS formula on screened samples.
- Low-command proxy duration is the duration of screened samples, not independently verified hover time. Low command does not prove hovering or stationary position.
- Recorded duration is arm-to-disarm recording length, NOT verified airborne time.
- Battery voltage and throttle are median values except the explicitly labeled p95 throttle.
- Total recorded duration: 141.43 s. Total samples: 57,613.

## CSV columns
trial: recording index; source: original log name; recorded_duration_s: arm-to-disarm recording seconds; samples: decoded sample count; gyro_rms_deg_s: overall three-axis RMS; gyro_p95_deg_s: angular-speed p95; low_command_gyro_rms_deg_s: screened three-axis RMS; low_command_proxy_duration_s: screened duration; median_battery_voltage_v: voltage median; median_throttle_us: throttle median; p95_throttle_us: throttle p95; added_payload_g: added mass only; corrupted_frames: corrupted frame count; clean_end: log ends cleanly.

## Interpretation limits
Trial 2 has the least angular rotational activity. Trial 3 has much greater rotational activity and larger throttle inputs; it cannot be labeled the worst hover stability under equal conditions. Median battery voltages differ. Conditions and pilot inputs were not matched. No statistical significance, stable-hover claim, or verified repeatable matched-hover reference is established.

GPS, position, altitude, wind and drift were not measured. Current is uncalibrated: do not infer precise current or efficiency. No unavailable raw-log download is provided.

## One-second time series
baseline-001-timeseries.csv preserves the 143 supplied row values: 58 bins for Trial 1 (0–57 s), 44 for Trial 2 (0–43 s), and 41 for Trial 3 (0–40 s). Columns: trial, second, gyro_mean_magnitude_dps, battery_mean_voltage_V, throttle_mean_command_us. These are one-second arithmetic means of angular gyro magnitude (deg/s), battery voltage (V), and throttle command (microseconds), not the all-trial RMS or medians above. Seconds are elapsed recording bins, not verified airborne windows; the last bin may represent less than a full second. Lines connect only supplied points with straight segments; no synthetic samples are added. Signal scales are shared across trials; battery voltage is plotted on a truncated axis.

One-second averages of recorded signals; changes reflect flight maneuvers and conditions and are not equivalent to hover-position accuracy.

Next: refine controlled hover protocol, document battery state and conditions, match pilot inputs, and independently verify position and airborne windows before variable comparisons (payload, CG, propellers).