Work out robot repeatability estimate instantly with clear inputs, formula shown and shareable results.
Repeatability is dominated by transmission lost motion, not encoder resolution. Combining the independent error sources in quadrature (root sum of squares) is the standard approach, and it usually shows that halving the encoder count changes almost nothing while eliminating backlash changes everything.
Error budget
total = sqrt(encoder^2 + lost motion^2 + drift^2), each converted to a tool-point distance
Independent random errors rarely peak together, so root sum of squares gives a realistic figure while simple addition gives a worst case that is far too pessimistic.
No. Repeatability is returning to the same place; accuracy is reaching the commanded place. Accuracy is typically several times worse and can be improved by calibration.