2009/6/25  <a...@conceptmachinery.com>:

> Sequences 1-3 very clear.
> Sequence 4 – means increase D until graph will be represent with line and
> not with line that has noise (vibration type).
> Am I right here?
> Sequence 5. What is mean instability? Is this same as an oscillation?

Yes, exactly,

> I know what oscillation means, but what instability is and how it will be
> represented on Hal Scope.

In this context they are much the same thing, though perhaps different
in detail. An oscillation is probably meant to describe a periodic and
increasing failure to meet a  step change in the setpoint, and an
instability might be random variations in a position when there is no
change in setpoint.

You seem to be describing Ziegler- Nichols tuning, and a google search
on those words should help with more detail.
http://en.wikipedia.org/wiki/Ziegler–Nichols_method as a starting point

However, Z-N tuning is most effective for symmetric systems, you might
still get better performance with hand tuning from those base points.
My personal experience is "Beware of D" especially in any system with
backlash.

[Irrelevant detail] in my real job we run controllers where P I and D
are themselves functions of other things, including noise factors such
as fuel temperarature and simple factors such as distance from
setpoint. PI and D are not the be-all and end-all of control, and if
that is all you have then perfection (and indeed adequacy) might be
unobtainable)

-- 
atp

------------------------------------------------------------------------------
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to