In Chris Rosser’s ArudPilot tuning tutorial, he mentions that spikes in RPM data from BDshot are caused by DMA conflicts. In his example he has a huge number of spikes - I only have on on each motor on a 60 second flight.
Rosser says the remedy is to move the motor to a different port. I’m using AUX 1-4, which I’m understand are the best for BDshot.
Are these few spikes something to worry about? Or should I abandon RPM notch filters and go with FFT?
It’s unlikely it’s DMA since for most modern flight controllers we give bdshot exclusive access to DMA. This is more likely because the motor wires are noisy and the reading process is imperfect - sometimes bad data is read and no value is procured.
Thanks. In my case, is there any danger in using this data? Or should I switch to a FFT notch filter?
As info, I’m using a Cube Orange.
There’s no danger, the filter code is written to expect this kind of dropout
