Re: Digest Number 12
Posted by
Jon Elson
on 1999-05-17 23:43:53 UTC
Mike Romine wrote:
now when I tried it.
last sampling period. Real PID needs you to compute the integral term
from a history of some previous samples, at least 5 or 10. many also use
several previous samples to build a trend for the differential term. These
things usually require a rolling buffer of previous error samples.
Still, I'm cutting metal with accuracy approaching the machine's basic
accuracy, so I have no complaint about the CNC. Tool deflection
is probably my most significant error, now!
Jon
> From: Mike Romine <mromine@...>Yes, exactly. their URL is : http://www.servotogo.com/ but it didn't work just
>
> Jon wrote:
>
> > A dedicated
> > card (currently ISA) with quadrature encoder counters and servo DACs
> > interfaces to the motion hardware.
>
> Who makes this card? Is this the 'servo2go' card I have heard mentioned
> on this list?
now when I tried it.
>Yes, it does. Their PID doesn't keep history, it just uses error from the
> > I don't know what jerk control is - do you mean backlash compensation?
>
> Jerk control is limiting the rate of change of acceleration. I was not
> thinking of the right thing when I wrote that as jerk control is part of
> the feedback loop (you feedback 3rd derivative of delta position), not
> feed forward. Actually, it sounds like the EMC control system needs
> some work if the PID filter is not working and you are using feed
> forward to compensate (feed forward is an open loop compensator).
last sampling period. Real PID needs you to compute the integral term
from a history of some previous samples, at least 5 or 10. many also use
several previous samples to build a trend for the differential term. These
things usually require a rolling buffer of previous error samples.
Still, I'm cutting metal with accuracy approaching the machine's basic
accuracy, so I have no complaint about the CNC. Tool deflection
is probably my most significant error, now!
Jon
Discussion Thread
Andrew Werby
1999-05-17 02:59:40 UTC
Re: Digest Number 12
Dan Mauch
1999-05-17 14:27:04 UTC
Re: Digest Number 12
Mike Romine
1999-05-17 19:35:43 UTC
Re: Digest Number 12
Jon Elson
1999-05-17 23:43:53 UTC
Re: Digest Number 12
Don Hughes
1999-05-17 23:48:23 UTC
Re: Digest Number 12