Rate and throttle curves describe how the quad responds to stick position. Blackbox logs can record how you actually moved the sticks as you flew, and this page applies some analytical tools to compare the quad's configuration versus how much of the input range you used. This can help you find rate and throttle curves that fit how you're actually flying. Rate and throttle tuning is very much a personal preference and subject to "feel" that only comes with experience, but hopefully these tools can give you some insight into how well your current configuration is working for you.

Recording a log

To record a blackbox log, you need the following settings in Betaflight:

Once you've recorded a log, download the .BBL file and you can process it on this page. Your blackbox logs are not uploaded or stored on a remote server, they're processed and stored entirely in your own browser.

Rate curve analysis

The rate curve analysis looks at your yaw, pitch, and roll axes, and measures hold and reach. Hold is time spent at a particular location, and reach is how far you moved the stick before moving it back, for faster or more extreme maneuvers. Both of these are drawn on the graph along with the recorded rate curve for that input axis. There are vertical lines for the p50 (50% of your time was spent below this mark) and p90 (90% of your time was spent below this mark), to give you a sense of the outer ranges of how you're flying.

You can use test new curve to see the estimated effect of a new rate curve on the recorded data. The provisional graph shows where the p50 and p90 hold points move relative to what was recorded. You can also compare the effect on the elbow point on the rate curve where sensitivity goes from sub-linear (less sensitive for more precise flying) to super-linear (more sensitive for faster response).

Throttle curve analysis

The throttle curve analysis is much simpler than the rates: it's a histogram of time spent at each location. The overlay on the graph shows the 10th to 90th percentile region where you spent the majority of your time. The distribution will necessarily be somewhat wide, since even a hover throttle will move upwards as the battery's voltage sags.

test new curve projects the recorded throttle data through the new curve to show you where your stick time would end up with a different curve.

Since min_check isn't logged in the blackbox recorder, it's here as a configurable option in case it's different from your quad's configuration.

Data analysis methods

What gets counted, and what does not

Not every datapoint recorded is considered, since much of the flight time is idle on one axis or another. The analysis only considers datapoints where the stick is not centered and the throttle is above idle. The quad must also be armed. This means the analysis focuses on when the sticks are actively being used to control the quad. The gate detail on the page shows how much of the flight log is being considered for the analysis.

How reach is measured

Reach is determined by measuring reversals, defined by a stick movement that reaches a measured point then reverses direction for more than a few percent of the stick range. This means movements in a single direction with pauses between are not counted as reach, only the end range of the movement. This is to identify how far the stick needed to travel, not how long it spent at any particular location.

How hold is measured

Hold is measured by counting periods of active time where the stick is held within a small range for more than 250ms, which is slow enough to measure intent without including more rapid movements.

How throttle time is measured

The throttle measurement is based solely on time spent at each location, subject to "is the quad armed", and excluding zero throttle. There's no reach metric because the maximum throttle is constrained already to 100%.

How exact these are

These are all approximations, tuned on the author's recorded flights. The goal isn't perfect accuracy, but rather a big picture view: how do the rates you've configured compare to how you're actually flying? Would some expo give you better control over the active range while leaving headroom for fast flips? Would a throttle limit give you more control if maximum throttle is never used?

Further reading

For more information about how to set your rates and throttle curves, see the following videos:

back to blackbox analysis built by zerowidth