Stefan Sczekalla-Waldschmidt wrote:

> Something which is not really clear to me is the difference between
> the "velocity" and the "positional" PID approach.

  PID is usually used to control mechanical devices, often motors.  I
believe this refers to two applications of PID algorithms to control in
the one case the position of an actuator, and on the other, the velocity
(often rotational) of an assembly.

  Controlling temperature or level is of course classed as "position".

  The complications for control of velocity are that you must count
events (from an encoder) over a time period, or time between events, in
order to determine velocity, so that it takes a certain time to make the
measurement, and by the end of the measurement period, the velocity may
well be different to that at the start.

  Timing between events is particularly troublesome as at low velocity,
the time interval may become very long and you also have to perform a
tedious division to get velocity.
--
  Cheers,
        Paul B.