6-DOF Stewart Platform to be controlled in one month!!

[bvillersjr]

In short... Nima says he knows what he's doing, and that the profile does not conform to the Stewart Algorith because it would be HUGE.

[Nima]

ok
i have isolate the x y z and yaw moves in the profil with copy and combine.
now we have four sektions.z is not a problem all actuators make the same work.
yaw all actuators make the same way only 2,4,and 6 are invert.
on y actuator 1,2,5 and 6 are make 100 percent way.
3 and 4 make only 80% way.for example:with 100mm actuatoren.you are cut the power from the actuatoren 3 and 4 and go with 1 and 2 on 0 mm way and with 5 and 6 100mm way.than you look how many way have 3 and 4.and now messurge(in the best you messurges with the potis)for example 3 and 4 about 90mm on still the sim all actuatoren on 50mm and hi can go 50mm(100%) plus and minus.now you see 3 and 4 are move 40mm this is 80%.
in y you must cut the power from 1,2,5 and 6 and move 3 on 0 % and 4 on 100%
and now messurge 1,2,5and 6 and you have all.
you only have a problem with havy profils.only one actuators go on 100%
and want more the other becom a problem.but i mean with a littel work on the
profil i can find a answer on this problem.look the diagram for the actuatornumbers in any postes higher.

[Frakk]

Thanks Bernard and Nima!

That is a creative way to get the positions Nima! Unfortunately I don't think it is going to work. The actuator positions are not as clear-cut and linear, they cannot be calculated by simple math calculations and percentages. Their position have to be EXACT (+- the mechanical slack), otherwise the platform will fight itself just trying to hold positions. Every mechanical part will wear out and the motors will burn.

IMO the only way to do this is to do it properly. Even with the proper math algorithms, the geometry has to match perfectly. If you have a slightly elevated or shifted pivot point, it will be enough to cause problems. It will probably work within a range in the middle, but over stress parts when you move further from the "no stress" zone. We can spend weeks fine-tuning all the percentages for a platform and still end up with a sub-optimal performance and increased wear on the parts.

The only way I can see this working is with an intelligent controller that senses the position of the platform, and senses the current draw for each motor. Then it can correct itself by reducing the power to the required actuators while maintaining the position setpoint. But this is just running in circles, creating more problems than solving. Getting the algorithm done properly is much simpler and can be very easily adjusted for any odd geometry.

[Nima]

hi frakk,
i agree with you a correct plugin is the better way.but i dont see any guy were can make it.and ali have not many time.the idea with a controller is a good for this configuration.now a littel dampering over the actuators than hi have a testversion in the time were ali have.we must have a percentscaler with any numbers behind the point(80,1234565)

[Frakk]

It is impossible to get it done in the time frame. They don't even have the motor controllers to position each actuator AFAIK.

[Nima]

hi
i mean ali work on contollers.hi must have it in one or two weeks or the time is out.

[Nima]

hi frakk,
i have make thinks again and i mean my idea is for a home build dof better as the
steward.for the steward you must have a perfect geometry and the perfect size
from the geometry and all work over and under the actuatoren must perfect to and this is a big problem for a home user.my profil work with mesurge after all
work and geometry problems.and i mean i can make it with play in the actuatoren under 0,5 mm und this play can go in a rubber over or under the actuatoren

[Frakk]

I guess we have to try to see. I really hope your profile works without big problems, but the only way to see is to build it. Maybe when Ali gets to that point, he can try it on his platform.

[aliaero]

Hi to All,
First of all i want to Thank all the people who are really concerned about this project Especially Nima who is really helping me alot.
2nd i was quite busy in designing a PID controller for the model provided by Mathworks for my own configuration of stewart Platform. http://www.mathworks.es/access/helpdesk/help/toolbox/physmod/mech/ug/bq5q4dy.html

The problem i am facing currently is with the stability of the controller, i have edited the model file for my configueration and using SISO tools but unfortunately the system is not staballizing. can anybody help me regarding this.
here are some basic specs of my platform.
Height : 0.35m
Top Radius : 0.165m
Base radius: 0.26m
Anybody having event the remotest info ca be helpful
regards
Ali

[akina]

Hi, there is this demo about Simulink :
http://www.mathworks.com/products/simmechanics/demos.html?file=/products/demos/shipping/mech/mech_stewart_vis.html

You can may be refound equation and Matrix with this model, and try your control may be.

My tests are begun for 3DOF http://www.dailymotion.com/video/xdgat6_3-dof-perso-type-frex_videogames
It is a good solution to look force and loop of control.