@Erik,

"Thus the use of a CPLD that can keep up with the oscillations"


sorry, but I can't understand, how a CPLD can give me an error below one step ?

If I would use a CPLD I would do exact the same as my software. I would use four FFs which latch the actual and the previous state followed by a logic, which decode the state to an increment or decrement signal for a counter.

The only difference, the CPLD can be clocked with e.g. 50MHz instead 50kHz of the software sample interval.

But again the error was still the same: one digit.

Thus the step rate of an encoder was designed in such a way, that an error of one can be accepted.
And thus it plays no role, if on the point of the oscillation the last or the actual step value was accepted.

I have also seen on our device oscillations, e.g. of exact 100Hz, caused by mechanical oscillation of a big transformer nearby.
But this was only internal, since the displayed position accuracy was above one step.


I know there exist position encoders, which deliver two analog voltages and then also step rates below one can be measured.


Peter