Hi, I think you have no errors, how accurate is your counting of output revs, to the fraction of a degree?

The proof is as you increase the total of each rev sample your perceived error is less and less.

In fact no matter how many turns you count I would say you will always be 8 to 15 turns out, this is due to how accurate you measure the stopping point of the output shaft.

1 rev is 472 counts, are you stopping the output shaft to exactly 1 in 472 of a rev?

I doubt it mate.

So your output count is accurate, its the stopping point that isn't.

Tom....

(sunglasses cos its 30DegC)