DATA REPORT / 003 DESCRIPTIVE PLACEMENT EVIDENCE

Switch experiment

Weight Placement
& Flight Stability

Four Blackbox recordings: three with weight on side(s), and a final recording with +2.3 g at the front. Measured rotational activity—not a verified placement effect.

BETAFPV Meteor75 ProBetaflight 2025.12.5Angle Mode (owner-reported) · owner-reported, not independently time-stamped
Recordings
4
Side mass
Unspecified
Front added mass
2.3 g
Session 3
Side · confirmed

Four logs exist although the owner describes three tests. After initial processing, the owner confirmed session 3 had the same side-weight configuration as session 1. Side-mounted total mass is unspecified; matched mass and center of gravity are not established.

Summary CSVMethods & limitationsbtfl_all(2).bbl · raw BBL not bundled

Source: exact owner-supplied derived evidence, decoded and analyzed outside this website. Reported zero parsing continuity errors and recognizable end markers in all four sessions; approximately 407.2 Hz. Not independently re-decoded here.

01 / ONE-SECOND BIN MEANS

Flight signals over time

Arithmetic mean gyro vector magnitude—not RMS—with mean logged battery voltage and throttle. Angular velocity is a rate, not tilt angle or position.

Session 1 · Side(s) · mass unspecified · Angular velocity magnitude (deg/s) · 5 recorded bins

Lines connect supplied bin averages, not raw samples. Axes auto-fit each selection; voltage and throttle axes are truncated. Final second-4 bin: 149 samples. No higher-frequency samples reconstructed.

Time-series CSV
Exact recorded bin readings
Session 1 one-second arithmetic means
SecondMean magnitude deg/sMean battery VMean throttle µsSamplesLow-command samples
014.914.1581000.04080
120.163.9761279.9408258
29.644.0611220.5408408
38.894.0801196.3408339
421.153.9151327.6149149
02 / SCREENED ROTATIONAL ACTIVITY

Compare low-command samples

Front trial 4 low-command gyro RMS: 9.05 versus side trials 1, 2 and 3 at 17.83, 14.44 and 12.56 deg/s. A descriptive finding only—different battery state, inputs, mass and recording windows prevent a causal conclusion.

Low-command gyro RMS (deg/s)Recorded order · descriptive only · not a causal ranking

Sessions 1, 2 and 3 are side-weight; session 4 is front +2.3 g. Bars show different screened populations and unequal recording windows—not matched mass, verified hover or a placement effect.

03 / EXACT SESSION SUMMARY

Every recording, kept separate

Overall and low-command gyro statistics describe different sample populations. Recorded duration is not verified airborne time. Proxy seconds are screened sample count divided by approximately 407.2 Hz.

Experiment 003 exact per-session summary; horizontal comparison with units
Session / placementRecorded duration sSamplesOverall gyro RMS deg/sOverall gyro p95 deg/sLow-command samplesLow-command proxy sLow-command RMS deg/sLow-command p95 deg/sMedian battery VMedian throttle µsp95 throttle µsNotes
Session 1Side(s) · mass unspecified4.3651,78117.9632.301,1542.8317.8333.834.0712141392Very short window
Session 2Side(s) · mass unspecified13.0525,321241.44267.362,0565.0514.4432.534.0512151362Large angular spike near end; do not compare overall RMS to other flights as steadiness
Session 3Side(s) · mass unspecified12.8485,23661.20146.282,3715.8212.5626.244.0012091406Same side-weight configuration as session 1; owner-confirmed after initial processing
Session 4Front +2.3 g61.68125,12256.72122.4313,03932.029.0517.123.8712391316Longest recording

Source: btfl_all(2).bbl, sessions 1–4, owner-derived summaries. Session 3 is side-weight, with the same configuration as session 1 confirmed by the owner after initial processing.

04 / INTERPRETATION

Lower angular RMS, not proven stability

01 / SCREENED EVIDENCE

Front recording: lower screened RMS

Session 4 records 9.05 deg/s low-command RMS, versus 17.83, 14.44 and 12.56 deg/s in side sessions 1, 2 and 3. These are unequal screened samples, not matched-hover trials or proof that front mounting improves stability.

02 / LATE SPIKE

Session 2 needs context

A large angular spike occurs near the end (second 12 mean magnitude: 502.72 deg/s). A collision or maneuver is possible, not independently verified. Overall RMS 241.44 deg/s must not be ranked as flight steadiness.

03 / CONFOUNDERS

Battery and windows differ

Session 2 median battery is 4.05 V versus session 4 at 3.87 V. Side recordings last only 4.365, 13.052 and 12.848 s, versus 61.681 s at the front. Side mass, pilot inputs and conditions are not matched; no significant or causal conclusion.

05 / REPRODUCIBLE METHODS & LIMITS

Keep the evidence boundary explicit

Source & integrity
btfl_all(2).bbl, four separately decoded sessions. Owner reports zero parsing continuity errors and recognizable end markers. Full combined-file SHA-256, not a per-session hash:
14a6a462a64864c34aad54ecff0e84753aec78b572a752a380d3ac2260b44797
Raw BBL is not bundled. This site presents supplied derived evidence without independently verifying the decode.
RMS & p95
sqrt(mean(gyroADC[0]² + gyroADC[1]² + gyroADC[2]²))
Gyro p95 is the 95th percentile of vector magnitude. Overall RMS uses all session samples; screened RMS uses only low-command samples. Both are deg/s rates, not tilt angles. Neither is reconstructed from bin means.
Same baseline screen
Throttle rcCommand[3] 1140–1400 µs inclusive AND absolute roll, pitch and yaw setpoints each ≤10 deg/s simultaneously. Proxy seconds = screened samples / ~407.2 Hz. Not verified hover or stationary position.
One-second means
94 genuine supplied bins, in session order (5 / 14 / 13 / 62). Mean instantaneous gyro vector magnitude, logged voltage and throttle rounded to 2 / 3 / 1 decimals. Final bins are partial. Connecting lines are visual guides, not raw samples; axes auto-fit each selection.
Reproduction limits
Verify the original file hash, decode all four sessions in order, convert gyro axes to deg/s, then apply the formula, inclusive screen and one-second arithmetic means. Decoder/version, percentile interpolation and exact timestamp-boundary convention were not supplied. The methods download documents these gaps rather than inventing them.
Not established
No GPS position, velocity over ground, altitude, verified airborne time, hover, position holding, calibrated efficiency or significant stability result. Side mass and center of gravity are unverified. November 15 is owner-reported, not an independently time-stamped flight date. Experiments 001 and 002 remain distinct and unchanged.
06 / NEXT CONTROLLED CAPTURES

Document the mounting, then repeat

Document side mass and mounting, match total added mass and battery state, standardize pilot inputs and recording windows, and independently verify airborne state and position before attributing a stability effect.