Alternative
PIDtoolbox is a serious analysis tool with more knobs than most pilots need. Easyquadtune makes the choices and shows the consequence. Here is the honest split.
| You want | Use |
|---|---|
| Full control over how the step response is computed | PIDtoolbox |
| A before/after table and a CLI script for three profiles | Easyquadtune |
| To inspect raw traces frame by frame | Blackbox Explorer |
| To know whether the log can support a tune at all | Easyquadtune's quality score, or the free guide |
A step response is only as good as its normalisation. If each trace is scaled to its own peak, overshoot becomes zero by definition and every axis looks critically damped. Easyquadtune normalises both the setpoint and the gyro by the commanded step size, measures rise from 10 % to 90 %, and settles at ±5 % — which is why its numbers can disagree with a plot normalised the other way. The worked example shows those numbers for a real flight.
A bent prop, a dry bearing, a cracked arm. No analysis of a log repairs hardware, and a tune that hides a mechanical fault usually makes it worse.
The sample report shows every section and every measurement, with the tune values withheld.
Open the sample reportEasyquadtune 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.