One real flight, measured end to end

This is the maintainer's own Betaflight 4.2.9 (e097f4ab7) STM32F405 log — 273,093 frames at 1,719.5 Hz (8,000 Hz gyro, denom 2), 139.2 s of clean flight across 26 segments. Nothing on this page is illustrative arithmetic: it is the engine's output for that file.

1. What the file contained

Item	Value
Firmware	Betaflight 4.2.9 (e097f4ab7) STM32F405
Frames	273,093
Logging rate	1,719.5 Hz
Gyro / looptime	8,000 Hz / 125 µs
Clean flight time	139.2 s
Ignored samples	38,016
Debug mode	NONE

Ignored samples are disarmed time, ground idle and spikes. They are counted and dropped rather than averaged in, because a tune computed over a stationary quad on the bench is a tune computed from nothing.

2. Is the log good enough to tune from?

Score 64/100 — Acceptable (Missing Maneuvers). The score is about the evidence, not the aircraft: logging rate, debug mode, length, and whether the flight contains the manoeuvres that make a measurement meaningful.

Manoeuvre	Status
Smooth Throttle Sweep	detected
Step Response Stick Flicks	not found
Propwash Dives	not found
Full Throttle Punchouts	detected

This log has a clean throttle sweep and a punchout, but no held step input on pitch and roll. That is why the step response below is labelled partially usable: the curves are real, but a chart-grade step needs a deliberate, held stick flick.

3. Step response, per axis

Axis	Overshoot	Rise (10–90 %)	Settle (±5 %)	RMS error	Confidence
roll	8.6 %	54.9 ms	86.7 ms	0.146	good
pitch	28.9 %	47.5 ms	69.8 ms	0.197	good
yaw	22.4 %	100.6 ms	157.0 ms	0.171	good

Overshoot is measured against the commanded step size, not against the gyro's own maximum. Normalising each trace to its own peak — what the previous version of this page did — forces every axis to look perfectly damped, because the peak is defined as 100 %. The numbers above can therefore disagree with each other, which is the point: they are measurements.

Partial is a real answer. When the log has no held step, the response is marked weak and the engine is not allowed to raise P, I or feedforward on the strength of it. A tune that claims more certainty than the log supports is the failure mode this page exists to avoid.

4. Noise spectrum

Welch PSD over 256 bins, 3.36 Hz apart, up to 856 Hz. The peaks that survived the detector:

Frequency	Axis	Type	Amplitude
30.0 Hz	roll	low_frequency	-3.2

A peak is a fact; a filter setting is a decision. The engine names the peak and then checks whether a filter change would help or whether the peak is better fixed mechanically. That distinction is the difference between a spectrum and a tune.

5. Diagnostics

Health score 100/100 with no structural issues found. Diagnostics are the part pilots argue with most, which is why the paid report prints the rule, the measured value and the threshold next to every verdict.

6. What the three profiles would change

Profile	Values moved	PID / filter / misc	CLI size
Cine	17 of 27	11 / 4 / 2	45 `set` lines
Freestyle	13 of 27	12 / 0 / 1	45 `set` lines
Race	13 of 27	12 / 0 / 1	45 `set` lines

Counts only. The values themselves are the product and are not published here — a complete CLI for this airframe is a tune a stranger can paste without the log that justified it.

Read the whole report

The same flight laid out exactly as the paid report lays it out — every section, in order, including the shape of the CLI block for Cine, Freestyle and Race.

Open the sample report

Prefer the interactive screen? The in-app demo draws the same log's spectrum and step-response curves with the chart components the analyser uses.

Easyquadtune is not affiliated with Betaflight. Published 2026-10-21. Measurements come from the maintainer's own flight logs; an example is illustrative of the method, not a tune for your aircraft.