It is working really good beside for an error. I have, always in the same point of the rotation, a spike in the measurement.
Could you think to something reasonable which causes that?
That isn't a spike ... its a gap/delay in measurement data. See previous reply for a reasonable suggestion.