Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - calibration
Posted by
Carl Mikkelsen
on 2003-06-15 17:13:42 UTC
Right,
Returning to "home" position does consist of retracking all the legs until
the home switches are satisfied.
The use of Simulated Annealing I talked about in my last email is for the
initial calibration.
The fundamental problem is that measurements are difficult, especially
when you need to accurately locate something in 3-space. It is easier to
position the milling head to defined positions -- known x, y, z, R, P, W
(yaw) values, and have the software report the degree of leg extension
required to position it there.
What I left out is that I use manual-mode through a six-axis pendant to jog
the head into the right position.
To capture the calibration data, the control software read a list of
desired coordinates, positions the head approximately at the right place,
and then I manually fly the head to the coordinate. The X,Y,and Z are
simple. If I keep the R and P at 0, I can use a level for reference
(although a better level would be better). For Yaw, I attach a long
pointer to the head, and use it to keep the head at 0 degrees of yaw. When
I press the "sample" button, the software records the coordinate, and the
leg actuator absolute length (in "steps"). This data is passed to the
Simulated Annealing code to find the best fit for the machine geometry.
I've looked at placing switches, or electrical contact points for the
tool-tip to touch, but the problem is that this only handles 3 axes. It is
necessary to also measure the angular axes. I have some automatic methods
under consideration, but haven't implemented anything yet.
-- Carl
Returning to "home" position does consist of retracking all the legs until
the home switches are satisfied.
The use of Simulated Annealing I talked about in my last email is for the
initial calibration.
The fundamental problem is that measurements are difficult, especially
when you need to accurately locate something in 3-space. It is easier to
position the milling head to defined positions -- known x, y, z, R, P, W
(yaw) values, and have the software report the degree of leg extension
required to position it there.
What I left out is that I use manual-mode through a six-axis pendant to jog
the head into the right position.
To capture the calibration data, the control software read a list of
desired coordinates, positions the head approximately at the right place,
and then I manually fly the head to the coordinate. The X,Y,and Z are
simple. If I keep the R and P at 0, I can use a level for reference
(although a better level would be better). For Yaw, I attach a long
pointer to the head, and use it to keep the head at 0 degrees of yaw. When
I press the "sample" button, the software records the coordinate, and the
leg actuator absolute length (in "steps"). This data is passed to the
Simulated Annealing code to find the best fit for the machine geometry.
I've looked at placing switches, or electrical contact points for the
tool-tip to touch, but the problem is that this only handles 3 axes. It is
necessary to also measure the angular axes. I have some automatic methods
under consideration, but haven't implemented anything yet.
-- Carl
>This is purely theory, but it seems as if it should work.
>
>On a traditional 3-axis machine, it's calibrated/zeroed by sending the
>units towards the home position, till they activate a limit switch. Why
>could something similar not work for a Hexapod?
>
>The traditional design (extensible legs) has 6 points to indentify home
>-- reel in the legs (or extend them) until they hit a limit switch built
>into the leg...then stop. That would give you an identifiable "home"
>position, with an excellent degree of accuracy, would it not?
>
>Something similar would be even easier to include on the glide-type
>designs -- simply raise or lower the leg assembly till it hits the
>switch...simplicity itself.
>
>As for the initial calibration, this is a MUCH more difficult
>proposition. Anyone have any practical ideas about how this could be
>done? Modelling it in software is one thing, but making theory into a
>physical model is something else.
>
> -- Chuck Knight
>
>P.S. A practical question. How are Hexapods interfaced with your
>computer? If my count is right, you need 12 lines to control the motors
>(6 step + 6 direction), plus a few miscellaneous connections. If limit
>switches for each leg are included, that's another 6 as well. We're
>really pushing the limits on a parallel port, since it has only a
>guaranteed 8 data lines, and a few more if we get creative with the
>pins. How is this handled, on a practical level?
>
>
>Addresses:
>FAQ: http://www.ktmarketing.com/faq.html
>FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
>Post Messages: CAD_CAM_EDM_DRO@yahoogroups.com
>
>Subscribe: CAD_CAM_EDM_DRO-subscribe@yahoogroups.com
>Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@yahoogroups.com
>List owner: CAD_CAM_EDM_DRO-owner@yahoogroups.com, wanliker@...,
>timg@...
>Moderator: pentam@... indigo_red@... [Moderators]
>URL to this group: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO
>
>OFF Topic POSTS: General Machining
>If you wish to post on unlimited OT subjects
>goto: aol://5863:126/rec.crafts.metalworking or go thru Google.com to
>reach it if you have trouble.
>http://www.metalworking.com/news_servers.html
>
>http://groups.yahoo.com/group/jobshophomeshop I consider this to be a
>sister site to the CCED group, as many of the same members are there, for
>OT subjects, that are not allowed on the CCED list.
>
>NOTICE: ALL POSTINGS TO THIS GROUP BECOME PUBLIC DOMAIN BY POSTING
>THEM. DON'T POST IF YOU CAN NOT ACCEPT THIS.....NO EXCEPTIONS........
>bill
>List Mom
>List Owner
>
>
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
Discussion Thread
Charles Knight
2003-06-15 11:53:52 UTC
Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - calibration
Graham Stabler
2003-06-15 16:30:12 UTC
Re: A "simple" Hexapod project - calibration
Carl Mikkelsen
2003-06-15 17:13:42 UTC
Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - calibration
Carl Mikkelsen
2003-06-15 17:13:43 UTC
Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - calibration