In the older sensors increasing the timing budget by 4x gave you 2x improvement in accuracy. But when the sensors started using histograms, that statistical model broke down. I expect you either have enough signal for a good measurement, or you do not.
Phasecal does help. A little. It means it takes longer go get started as the sensor spends longer calculating the clock phase, and it works but it's not magical. But it does mean that if you do a stop/start, your answers will be more consistent.
The trick to all this is that the longer you give the sensor to range, the better the answer. But that also means the sensor won't return as many values and thus the latency goes up.
The sensor gives a pretty good answer with a 30ms timing budget, but some people use 100ms for a 'better' answer.
However, if your target is moving, then this extra time is counter productive.
You have a huge advantage - you've read the documentation. This is not something most people do.
Increasing the merge threshold was put in so people could play with it. I don't think there is one magic setting that solves everyone's issues.
Best advice I can give is double it and see if you like it better. Do a binary search after that and narrow in on a good value.
- john
If this or any post solves your issue, please mark them as "Accept as Solution". It really helps the next guy.
And if you notice anything wrong do not hesitate to "Report Inappropriate Content".
I am a recently retired ST Employee. My former username was John E KVAM.
