1. Do not share user accounts! Any account that is shared by another person will be blocked and closed. This means: we will close not only the account that is shared, but also the main account of the user who uses another person's account. We have the ability to detect account sharing, so please do not try to cheat the system. This action will take place on 04/18/2023. Read all forum rules.
    Dismiss Notice
  2. For downloading SimTools plugins you need a Download Package. Get it with virtual coins that you receive for forum activity or Buy Download Package - We have a zero Spam tolerance so read our forum rules first.

    Buy Now a Download Plan!
  3. Do not try to cheat our system and do not post an unnecessary amount of useless posts only to earn credits here. We have a zero spam tolerance policy and this will cause a ban of your user account. Otherwise we wish you a pleasant stay here! Read the forum rules
  4. We have a few rules which you need to read and accept before posting anything here! Following these rules will keep the forum clean and your stay pleasant. Do not follow these rules can lead to permanent exclusion from this website: Read the forum rules.
    Are you a company? Read our company rules

6dof inverted stewart platform

Discussion in 'DIY Motion Simulator Projects' started by harwoodr, Mar 24, 2019.

  1. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    Hey all,

    Long time lurker, second time poster. For my masters project, I'm building an inverted (hanging) stewart platform, using pneumatic artificial muscles (PAMs).

    I've built my own PAMs using hydraulic hose fittings, latex tubing and kevlar mesh sleeving and servo operated ball valves... because the commercially available offerings were really expensive.

    My PAMs are about 72 inches long and get about 18 inches of contraction at 60PSI. Tests with a mediocre compressor gave about 9 inches/second travel with a 75 lbs load. I've load tested them with static load of 200 lbs (roughly my body weight) and a dynamic pulse of me bouncing my weight on them harshly. :) The break strength of the combined kevlar strands in a PAM is over 12000 lbs - I'm more concerned about the hose/mesh pulling free of the fittings.

    The pneumatics consist of solenoid valve (directional control), ball valve (flow/speed control) and pressure sensor (position feedback) for each PAM.

    The servo valves are off the shelf ball valves with hobby servos attached to them. My original prototype valve had a 13 Kg-cm rated servo and worked quite nicely. I've just recently discovered that the (rated at) 11 Kg-cm servos I'm trying to use (because I cheaped out) are not "working as advertised" and are struggling. I've got some 21 Kg-cm servos on order from Hobby King, and expect them this week.

    I had tried a three-way ball valve but found the lag from "full fill" to "full exhaust" (or vice-versa) was just under one second - which is too slow. The solenoid in series with a two-way ball valve operates at a much greater speed, but sacrifices some of the fine control provided by the three-way ball valve.

    In a perfect world, I'd use high-speed and high-flow (Cv in pneumatic-speak) solenoids that could give me three-way direction control while being able to control flow speed (either via rapid on/off or PWM). I could find reasonably priced high-speed valves and reasonably priced high-flow valves... but not a combination of both that was reasonable. :p

    The control circuit is pretty simple - 6 atmega328p microcontrollers (running arduino firmware and using the 8MHz internal oscillator) are networked together (via i2c). Each atmega controls a solenoid (via mosfet), a servo valve and reads a pressure sensor.

    A master controller (either an arduino or a raspberry pi) sends commands to each microcontroller (or can send a broadcast to all). I'm in the midst of writing code for the master controller to take in the force data and do the inverse kinematic magic... inverse kinematic positioning is easy - inverse kinematic mapping of forces is a little trickier. Yes, I could do a lot of this with simtools, but then I'd be giving up a lot of the meat of what I need in my masters project.

    I've designed the whole thing to be portable - it takes about an hour for two people to dismantle it and about the same to assemble it... and I'll be doing a whole "howto" or "instructable" on making your own when I'm done with it.

    Let me know what you think. Beyond "are you insane?" because I get that one a lot...

    IMG_0153.jpeg IMG_0444.jpg IMG_0445.jpg
    • Like Like x 1
  2. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    Oh, additional note - I picked up a used oculus rift dk2 that works quite nicely with No Limits 2...
  3. chauffeur

    chauffeur New Member

    Joined:
    Feb 23, 2019
    Messages:
    11
    Balance:
    67Coins
    Ratings:
    +0 / 0 / -0
    My Motion Simulator:
    6DOF
    Very nice project, quite different from all the others I have seen on these forums. Looking forward to learning more about it.

    Those DK2 VR setups are a bargain and very much usable. I too picked up a pair recently.

    Thanks for sharing your project.
  4. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF


    Initial twist test of the platform
    • Like Like x 3
  5. PaulB

    PaulB Titanium2dof

    Joined:
    Apr 15, 2017
    Messages:
    121
    Location:
    Canada
    Balance:
    723Coins
    Ratings:
    +55 / 1 / -0
    My Motion Simulator:
    DC motor, Motion platform, 6DOF
  6. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    Oh, that's cool. Since I'm doing mine for my masters project (out of pocket, I may add) I decided to make mine portable.

    The whole contraption started in the basement, moved it to the garage last year (my son had a lot of friends over for his birthday) - so portability helped. It takes me about 2h to tear it down and about 3h to assemble - if I'm doing it by myself.

    The slotted angle steel makes it like a giant meccano set (I used to say erector set, but my wife giggles too much when I do - that and when I talk about air muscles getting "shorter and thicker" when in use... sigh.) so I'll be able to throw it in the SUV and bring it down to the university for presentation.
  7. PaulB

    PaulB Titanium2dof

    Joined:
    Apr 15, 2017
    Messages:
    121
    Location:
    Canada
    Balance:
    723Coins
    Ratings:
    +55 / 1 / -0
    My Motion Simulator:
    DC motor, Motion platform, 6DOF
    I can take mine out just by disconnecting the cables and unplugging the computer. All my drives are attached to the wall. I can get most of the room back just by tilting it forward. : )

    Making the big decision on 3 monitors or vr headset.

    Birthday! It's tough to get my friends out. must be a blast with kids.

    I'm also thinking about putting 3 or 4 of these on a trailer and renting it out at parties!
  8. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    VR vs monitors - depends on the game... monitors will always work and (from what I understand) are less likely to cause motion sickness... Dunno, never had motion sickness myself.
  9. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    So, I've got my motion cueing software started now. Using python, numpy and scipy - the built in filters make life so much easier. See the step response plot:
    bilinear.png
    Just need to tie a few things together and the basic motion cueing will be done - next step will be to integrate it with the inverse kinematics...

    My plan here is that I'll have my python code (on a raspberry pi) listening on UDP for (relatively) raw forces from simtools for each of surge/sway/heave/yaw/pitch/roll. The code will feed that into motion cueing which will in turn feed it into the inverse kinematics... then that will be spit out to my master valve controller.

    I could use the pi to control the valves directly (each valve controller communicates as an i2c device) but the pi has issues with i2c and clock stretching - and my master valve controller talks over serial, so, when the project is done, I can simply use the master valve controller to talk to simtools for game play. ;)
  10. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF


    It was going really well until the hose end pulled out... nearly have that problem licked though.

    May have to introduce some smoothing to the signal or hysteresis on the valve control logic.
  11. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    Latest test run... replaced my two way ball valves with three way ball valves... and now the solenoids only operate as part of a enable/"kill switch"...

  12. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    C76A78A8-7E04-45F0-95AB-E43F47217F2C.jpeg Rig passed over an hour and a half usage with 260lbs of bricks on it... so, I’ve mounted the seat now - I guess the next test is with me riding it.

    Good thing I put a seatbelt in too...
    • Like Like x 1
  13. harwoodr

    harwoodr New Member

    Joined:
    Dec 9, 2013
    Messages:
    28
    Balance:
    441Coins
    Ratings:
    +11 / 0 / -0
    My Motion Simulator:
    Motion platform, 6DOF
    My first actual ride:



    Now I have to get the kartkraft plug in working... grumble