1. 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!
  2. 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
  3. 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

Showroom Belt driven linear actuator

Discussion in 'DIY Motion Simulator Projects' started by drumlight, Dec 27, 2021.

  1. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    I've already come across a couple of problems which mean I'll want to start the design over from scratch but before that I thought I'd see what other peoples thoughts are.

    LayoutAttempt_2021-Dec-27_05-43-40PM-000_CustomizedView32628438028.png

    My basic concept is to bolt two end mounts and some 12mm linear rail to a 2"x2" aluminum extrusion. A carriage would be pulled up and down these rails with a 4:1 pulley system driven by a BLDC motor. The actuator rod can be aligned with the centre of the belt so most of the forces align well and the whole assembly can be fairly low profile so the rod isn't centred very far above the aluminum extrusion either.
    LayoutAttempt_2021-Dec-27_06-16-12PM-000_CustomizedView15610321884.png

    Using Gates MGT3 belt (3mm pitch 15mm wide) with 16 teeth pulleys seems to work within their specifications for toque and belt tension whilst producing decent actuator forces and speeds with available motors.

    index.png

    I was trying to use lots of standard and repeated parts like 1" long dowel pins and 3/8" machine key to transfer loads from the rod and pulleys into the machined aluminum plates. I think the machining can be kept pretty simple and 3d printed parts can probably be used anywhere machining parts from multiple sides would otherwise be required.

    LayoutAttempt_2021-Dec-27_05-46-31PM-000_CustomizedView14181630083.png LayoutAttempt_2021-Dec-27_06-03-22PM-000_CustomizedView9664615445.png

    I haven't really decided if I should mount the motor directly to the end plate or to try having it mounted to the extrusion and connected through another belt to offer some more gearing flexibility and higher motor speeds. The fixed end of the belt should also be secured in the centre of the carriage instead of at an edge which would also remove the need for two of the pulleys one the carriage required to clear the rod.

    ps I'm unemployed and broke so not likely to be doing anything physical with this idea, it's just for practice on Fusion and because I saw PTFlyer and some hexapods on YouTube which were awesome.
  2. Ads Master

    Ads Master

    Balance:
    Coins
    Ratings:
    +0 / 0 / -0
  3. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Here is the same graph without the transparent background which looked fine until I saw it in my browser. index1.png
  4. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
    I think belt drives are superior to ballscrews in terms of noise

    You can use a gearmotor to drive the belt and changing pulley size can adjust the torque and acceleration

    The belt will act like a type of shock absorber reducing vibration.

    Have you worked out tje cost to build one

    Consider a chain instead of a belt, might be easier to affix the output
  5. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0

    I think the minimum diameter of a chain cog is going to be much larger; I don't think I've seen cogs much below 10 teeth and I don't think chain is efficient or long lasting as belt at very small radius.

    I looked into using 5mm pitch belt because it allows 10 teeth pulleys as the smallest and it works out very close to the 3m pitch with 16 teeth. The extra size of the belt would still have required a wider design (I think ~5" is possible with 3mm pitch belt) unless you went to a 2:1 pulley system at which point the extra strength of the 5mm belt was required because the loads are twice as high.


    The belt should be pretty easy to secure in the centre with a scaled up version of the clamps they use on XY 3d printers (its also what will allow the belts to be tensioned). If other things hadn't already depended on the pulley location I would already have removed two diagonally opposite pulleys in the middle of the carriage, added two clamps for each end of the belt and then have the belt path run between the two clamps from one side of the carriage to the other. The location of smooth and toothed pulleys would end up reversed on each side the system.

    I'm still new to Fusion and learning the best ways to set up dimensions and reference points to avoid trapping yourself down the road. If I started the model over I'll be able to make my life much easier but it also felt like a chore trying to finish this far.


    I haven't worked out the exact cost but I've seen bags of 50 suitable idler pulleys for ~$300 dollars and it would require roughly that number of both toothed and smooth varieties to build six actuators. BLDC motors in the range of 140-200kV capable of 30A are something like for $50-120. 5m lengths of belt (should be good for one actuator using 2ft extrusion) are available for $10-20 but I'm not convinced they would meet the same specs as real Gates belts. A motor controller like the ODrive would be ~$60 a motor, position encoders might be required although just hall sensors should still have positional control <1mm..

    I guess things would add up to a minimum of around $200 per actuator even with careful shopping and the ability to make some parts yourself. It was really just a hypothetical exercise imagining what I could do if I had a small CNC machine and 3d printer.
  6. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Actually that cost is definitely on the low side, 15mm wide belts were more like $40 each. I'm not sure what a reasonable price for the 12mm shaft would be, custom ones on McMaster started at something like $100 but I'd try local machinists first. Blanks shafts from china are really cheap but wouldn't be easy to work with at home.
    Once I have a slightly better design I'll get the flat machined parts quoted online, I know there is a plugin for Fusion that automates it all.
  7. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
  8. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    This is more practical belt path and it should accurately represent a 3GT belt pitch line for 16 tooth pulleys. I think this is about as narrow as it could be made and is basically 120mm wide. The carriage that pushes the rod doesn't need to be so big that it limits the available travel range significantly.
    Belt Plan Drawing v1.png


    Edit: This pulley layout is wrong in a couple of ways. Firstly I never drew the 4:1 ratio I was aiming for and then when I relocated the fixed ends of the belt I split the pulley system into a second and unnecessary parallel system.


    For the motor I had something much more like this in mind https://flipsky.net/collections/e-skateboard/products/bldc-motor-h6355-160kv-1620w . At a 30A peak it should produce actuator forces close to 200lbs if directly driving one of the pulleys. I think on average the currents will be much lower so maybe the peak could be pushed higher still.
    Last edited: Jan 2, 2022
  9. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
    why are you using so much belt and a pulley arrangement ?

    i think you can make a simpler design with a gearmotor and a simple belt loop
  10. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
  11. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
    Last edited: Dec 29, 2021
  12. Gadget999

    Gadget999 Well-Known Member

    Joined:
    Dec 27, 2015
    Messages:
    1,666
    Location:
    London
    Balance:
    10,360Coins
    Ratings:
    +402 / 8 / -0
    My Motion Simulator:
    2DOF, DC motor, Arduino, 6DOF
  13. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Those are definitely neat designs but I still see some advantages; the first chain driven example seems to be using a 16 tooth sprocket on the motor and if it is a 6m pitch chain that would give a pitch radius of 15.3mm. The belt pulleys are exactly half this size so would produce twice as much force from the same motor torque.

    The chain is directly carrying all the force in those actuator so it is under greater stress (especially where the rod connects to the chain in the first case) and then also require a motor with much more torque or a gearbox. In my design very few parts ever see the full force of the actuator and connecting the rod to the carriage is the area I'm most concerned about.

    The pulley system quadruples the forces from the actuator but also importantly keeps the belt tensions low enough to be viable, a 9mm wide belt would have been right on it's specification limits even with the 4:1 pulley system but the 15mm wide belts seem more than strong enough.
  14. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Basically the tiny pulleys and the 4:1 pulley system are my approach to avoid the need for a gearbox and big motor. I think this system allows small brushless motors to spin fast enough they can be used without needing a gearbox or even an especially strong motor; something with a peak toque of <2Nm should suffice.
  15. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0

    Attached Files:

    • Agree Agree x 1
  16. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
  17. adgun

    adgun Active Member

    Joined:
    Jan 28, 2008
    Messages:
    457
    Occupation:
    mechanic
    Location:
    Netherlands
    Balance:
    5,279Coins
    Ratings:
    +112 / 3 / -0
    Hi can you tell me where you build this for.
    Youre 2nm peak force gives 20kg/cm no need to make a 1to4 pulley system, thats only diveded this little force.
    There is no reduction to give more force and lower speed with this motor.
    regards Ad
  18. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Without the pulley system dividing the forces by 1/4 the belt I used wouldn't be strong enough; other bigger belt options exist. I don't think actuator speeds are being limited by the pulleys; 3000 RPM should be acceptable for the belt and pulley combination and results in 600mm/s speed.

    I might have overshot how much force is required in an actuator as I kind of guessed at that. My basic thought process was a sim platform might weigh from 200-lbs to 350lbs. I think a hexapod design is essentially connected by equilateral triangles so depending on the exact position and direction the forces from the actuators sum somewhere between 6 * 0.5 = 3 and 6 * 1/2 * root(3) = 5.2.

    I assumed to be responsive the platform would want to accelerate at a few m/s^2 so having the peak actuator forces approach 200lbs made some sense to me. I'm sure you wouldn't want/need that all the time and I'm pretty sure a smaller motor could be used but I figured I'd start with plenty in reserve.
  19. adgun

    adgun Active Member

    Joined:
    Jan 28, 2008
    Messages:
    457
    Occupation:
    mechanic
    Location:
    Netherlands
    Balance:
    5,279Coins
    Ratings:
    +112 / 3 / -0
    You make a mistake
    You get 600+mm/sec at 3000rpm with a 38mm Diameter pulley,and then you come at 10kg peak force with 2nm motor peak force
  20. drumlight

    drumlight New Member

    Joined:
    Dec 26, 2021
    Messages:
    25
    Balance:
    217Coins
    Ratings:
    +3 / 0 / -0
    Another advantage from the pulleys is it reduces the holding torque required across the board. Assuming a platform weighs 300lbs I think there is a static load required from each actuator of 300/(6*1/2*root(3)) = 57lbs. With the pulley system this will require between 5-10A steady current but without the pulleys it would be close to the maximum current I wanted to run the motors at.

    ps. To your earlier question, unfortunately I'm not planning on using this, I'm currently unemployed so don't have the money or even space to be building a project like this in real life.
  21. adgun

    adgun Active Member

    Joined:
    Jan 28, 2008
    Messages:
    457
    Occupation:
    mechanic
    Location:
    Netherlands
    Balance:
    5,279Coins
    Ratings:
    +112 / 3 / -0
    i made a mistake youre talking about a 3,8mm D pulley for that speed and you come at 100kg force.
    you need a lot more belts every tooth from a belt can handle some force ,with a 10 tooth pulley you get 6 tooth contact with 90kg its 15kg every tooth. If a tooth can handle 3kg max you need 5 belt and pulleys on motor shaft