New 6dof Stewart platform for a school project

Hi glad it's working!

@Thanos

Kalimera, you can use the auto limit button to adjust the limits to 70% of its maximum. Support for cheaper servo controler will come in its time, I don't own any right now. But I can share the source with anyone who wants to do it.

@Raceray

L1 gives the lenght of the servo arm, L2 the lenght of the rods (I don't know if it's the right word). I shall add absolut measurements for size of platform and base soon.

Thanks for your appreciation!

Robin this is really great.
I am currently trying to determine the cinematic equations to calculate the position of the servos for a given position of the platform.
Would it be possible for you to describe a bit the method you use to do these calculations? Or may you have any kind of documentation on this subject?

Thanks in advance! And keep up the good work!

Hi Lancelot. As you may not be the only one interested by these equations I'll give you the method to calculate the servo angles. First read planmix's post here:
http://www.x-simulator.de/forum/post8383.html#p8383

First step: calculate platform translated position

Let's say PF[] are the platform points at position 0 (Pm in planmix's post). You calculate the translated points PFT[] like this:

PFT= (PF + T) * [R]

T is this vector: Tx = surge
Ty = sway
Tz = heave

[R] is the rotation matrix with pitch yaw and roll inside: see here the coefficients of the matrix:
http://en.wikipedia.org/wiki/Euler_angles

Second step: calculate position of PL

Let's say PL[] are the points where the KneeCap is between the servo arm and the rod (attachement of the two). PB[] is the center of rotation of the servo (constant). L1 is the lenght of the servo arm (constant). beta[] is the orientation of the servo around z axis (constant). Servo[] are the angular position of the servos (variable we want). You can now calculate the coordinates of PL like this:

PL[0]=PB[0] + L1*cos(Servo)*cos(beta); PLx
PL[1]=PB[1] + L1*cos(Servo)*sin(beta); PLy
PL[2]=PB[2] + L1*sin(Servo); PLz

in 6dof master we have:
beta[] = {pi/3, -2*pi/3, pi, 0, 5*pi/3, 2*pi/3}

Third step: ||PFTPL|| = L2
L2 lenght of the rod (constant).

The lenght of the vector PFT to PL is constant and equal to L2. You'll have your equation from there, try to solve it! If you ask I'll give complete solution.

Hi Robin Thnaks for the updated Program. have to report it works great and cant wait to interface it with my 6 DOF motion platform model ... built all the mechanical and eletrical elements months ago have the contoller rc servos fitted and even bought Matlab + Simulink but software was never my strong point so that project kinda stalled waiting for Plannmix to realease his 6DOF Motion software , so i have to say your work will be a godsend if i can interface it through X-sim My Cad drawings for the 6DOF Servo model can be found in the Cad Section of the forum and hopefully if i can get your software to work with it and prove it out mechanically i'll release the Fully dimensioned and easy to make parts list and drawings . So On Behalf of all the other sim dreamers out there thanks for sharing the work you have done and are doing

Hi wannabeaflyer, I have to say it's a pleasure that you appreciate the software. Why? Cause Arnaud and I used your cad drawing as a basis for our platform dimensions! ahahah.
We'd like to take this occasion to thank everyone who contribute:
* egoexpress for his support.
* Thanos whose work has been an inspiration: come on making a badass simulator out of pvc tubes and wiper motors! (among other things)
* planmix for his post on the reverse kinetics of the platform.
* and of course sirnoname for x-sim without whom all of this wouldn't have been possible.

As we have to finish our platform before the 17th june, the link to xsim will be made as soon as possible and we intend to build a dynamic model of the car moving the gcenter to follow the transfer of masses.

A question for wannabeaflyer: which servocontroler did you use with your 6 dof platform?

Hi Robbin glad to hear some of my mad professor stuff was used by a fellow member :yippiee: i got the servos off e-bay from hong kong, i belive i got the Hitec ones with decent Torque but will have a look at them when i get home and post the details hear.. Cant really say thanks enough cos the work you guys released will make my baby Fly ( ok ok onlt 8 off the deck ) but as Doctor Frankiensiten Said of his monster HE LIVES HE LIVES :lol: in My case it'll be SHE MOVES SHE MOVES :

Hi everyone, tks for for your answers but let's try a little less so enthousiast hahaha! A lot of work still have to be done!

@wannabeaflyer

your baby should be able to move if we are able to implement your servo controler. For what I can see on the picture you I think you have the Pololu mini servo8 controler. Does anyone have some source code using this servo controler who could share it with me (C++ would be best)? I should be able to learn from this code and adapt it for 6 dof master.

@Raceray

I have no problem sharing the code for servo angles. A 6dof platform is really kick-ass too bad it's a miniature one! But I really think you should build one as it's not too expensive.

Some updates of the project: Arnaud and I build the very first prototype of the platform yesterday, we have some few mechanicals problems which will be resolved asap, but for what I saw, the moves are correct and are the same ones as the 3D model (well with less accuracy and some shaking allright :lol: ). So it's rolling. Have to link it to xsim this week-end, hope I won't have to much trouble doing this.

See ya

Hi, Tks Raceray for the source code, I'll try to have a look at it when I'll finish my project. Arnaud and I made the first prototype of the platform this week-end and we linked it to X-Sim :yippiee: . Here are the first results:

http://www.youtube.com/watch?v=3lDv-r9mcOQ

No dynamic model of the car what so ever, just
longitudinal accel -> pitch
lateral accel -> roll
vertical accel -> heave
and yaw accel -> yaw
with no smoothness no washout no nothing. Pretty raw config.
Music is Joy Division.

See ya

Congrats! Great job! Cool video :thbup:

How did you guys manage it?

Regards
Christian

superb :clap:


... the next step is a 6dof-simulator for the human usage 8)

Hi everyone tks a lot for your comments! To get the data out of X-Sim we simply used the visual export interface as descripted here (Big tks to sirnoname):
http://www.x-simulator.de/wiki/Visual_Export
and here http://www.x-simulator.de/forum/post14804.html?hilit=Visual_Export#p14804
I maintain a refresh rate of 50ms for the datas as in Visualierung interface here:
http://www.x-simulator.de/main/en/programmieren/visualisierungs-interface.html

I will release very soon a new version of 6dof master where you can animate the 3D model using the values of X-Sim (axis set in math setup). It's already working but I need to do some interface job.

:thbup: yes but that'll wait until I have money and a big home!

So for now the next step will to give some more accuracy to the moves, building a dynamic model of the car and so on. I would like to have the g-values of X-Sim converted to real force in m.s-2 (one G is 9.81m.s-2) so I think I will make a small program which derivate the speed of the car to have the longitudinal accel and compare it to X-Sim output.

some greek for the greetings:
Yassas filoi

@RaceRay
I watched the video of your sim yesterday, very cool simulator it's like it came out of a Star Wars or Mad Max movie! Love your speakers too, what are those: B&W, Cerwin Vega?

Hi René,

Sorry for my mistake, I thought it was your simulator :lol: ...

Hi Robin Have only just gotten a chance to see your Video Bloody hell man thats Brilliant : Now all i have to do is Copy It LOL Ps now you can teach me something , on my Model i have the Legs adjustable for setting up, What length did you use and also the servo arm length < may as well use your figures as you have proven that they work

I would like to make a suggestion. Since you are at the very beginning of writing your motion algorithms, you may want to consider the following observation I made when your model undergoes braking. The platform tilts forward (good), but it also moves forward (bad). Please read my post about accurate motions as it relates to forces generated on the body and head...

my post at bottom of page;
cycraft-t1645.html

This appears to be a function of how the pivot arms for each leg are arranged. Study this video to see an accurate movement for braking...

the first motion (platform going to the left 'back'), closest side to sofa being front of 'car') after the 2:00min mark, this would be identical to what is needed under braking;
http://www.youtube.com/watch?v=MxXEOcYY1M0

What is being applied is the surge motion in conjunction with tilt. Under braking the platform nose would tilt down and also move (surge) rearward. This would apply the accurate 'G' forces on the body for the braking effect.

Similarly,
Accelerate; surge to the front and tilt the platform down at the back
Left turn; sway to the left and tilt the platform down on the right
Right turn; sway to the right and tilt the platform down on the left

I am not going to say that the motor/arm alignments are perfect in this video example. But at that point, the braking motion/forces achieved would be correct for a car.

You are making great progress, and it would be wonderful if it achieves accuracy also.

Check out this controller to see if building a similar device to move a full/model simulator would be helpful in understanding the movements required of each leg to achieve proper motion. Using linear potentiometers for the 'legs', and recording the measures for the applied motions, would probably help in developing your algorithms.
http://www.youtube.com/watch?v=fnRBX51U ... re=related

Regards,
R-eng

6dof Software new version

As promised a new version of the 6 dof software, click on Link X-Sim to do so. The visual axis one set in math setup will go to the yaw input, the 2 to the pitch etc until the 6 goes to the heave. The 3D simulation animates itself when receiving differents inputs from the profiler.
All dimensions in millimeter, click on setup points->Change to have the values of the differents coordinates of the points of the platform and base. CAD drawings will come shortly.

There's a new version, go to page 5 of this topic to download it!

@R-eng

Thanks for your comments. Unfortunately I have a problem with SD84 Servo Controler I'm looking at for the moment. There's a dead band causing the servos to jump sometime from one position to another and which causes the platform to make huge deplacement or shaking. You can see it in the video maybe around 5 times. I wouldn't advise anyone to use this servocontroler.

About mass transfer of the car, open 6dof software, open 3d simulation, move platform, move gcenter, observe what happens. that's the answer I can give you for the moment, I'll look forward in this soon.