Re: Hi, and 2 questions (kinematics and look-ahead)

[robin szemeti <robin-at-rszemeti.demon.co.uk>]

On Tuesday 10 September 2002 06:10, Jon Elson wrote:

> > Take my router for example .. it 'sags' in Z along the X axis by 20 thou'
> > due to flex in the beams, I'd really like to take that out of the Z axis
> > as the X axis moves along ... trivial really .. but sadly after reading
> > through the [blah]kins.c files I'm no wiser ... I'm familiar with C C++
> > and Perl (and a bit of lisp on occasion too) so its not the language its
> > just the core concepts of whats going on thats confusing me here.
>
> Now, this is an interesting idea, and I suppose it could be done.  I'm not
> sure the
> usual 4x4 matrix has a high enough order to perform the kind of correction
> you are considering.  Normally, these kinematics routines are used to
> translate the joint position of a robot to/from cartesian coordinates.

The idea is to use the software (which is cheap [1]) to compensate for lower 
cost components ... yielding an improved performance from cheap hardware 

so .. assuming I have a 48" X axis .. if I call 
G0 X0
G1 X48

will the kinematics get called hundreds of times as it crosses the X axis or 
just once to work out the position of x=48 in joint terms and then its up to 
emcmot to get it there?

The router was just one example ...

The main reason I need this is to provide compensation on a machine that has 
a guide rail made up in multiple sections ... because alignment is not 
perfect instead of being say 4 staright sections [ ---- ] it ends up as a [ 
\/\/ ]  ideally I'd like to map the zig-zag errors into a table and get 
straight lines out of a wobbly setup ... if however the kinematics is just 
called the once to calculate the end point of a line then if the start and 
end point have hte same offset in the table then no correction would occur .. 
it needs to be called continuously throught out the operation .. I haven;t 
quite figured out exactly when it is called in the overall scheme of things.

<snip look ahead which sounds like it will work perfectly>

[1] no it's not,  I'm a programmer .. I should know better :)

-- 
Robin Szemeti