Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Building a "slider crank"-style lifting platform. Need help.

  1. Sep 27, 2015 #1
    Hi, I am trying to design a lifting platform for light duty loads (3d printing build platform) that only uses rotation for its actuation. I have a design that uses a mechanism similar to the slider crank and I want to maximize the total height the platform can travel.

    This is a sketch of my design: addon.php?185,module=embed_images,file_id=62957.png

    The rotation is going to be executed by a motor at the red dot and my dilemma is about the behavior of the two arms when they reach the position at the middle of above image. I think this is called a singularity or dead point.

    If the platform is going down, the way I have suggested in the sketch, at that point I see a weak mechanical position when the platform will tend to fall more than it should because of its weight. This is where I would loose precision in movement and I need a solution to overcome this.

    A preloaded spring was suggested to me but I don't exactly know how should that be setup, to which direction is better? Is there any other (simple/cheap) solution employed by other such mechanisms/robot arms?

    I am also thinking about having gears at the green dot ends of both arms but that would introduce backlash when the rotation would be reversed.

    Thank you for your suggestions.
     
  2. jcsd
  3. Sep 27, 2015 #2

    mfb

    User Avatar
    2016 Award

    Staff: Mentor

    Do you need the crank? Rack and pinion sounds so much easier. You just have to shift the motor a bit (or use a gear in between).
     
  4. Sep 27, 2015 #3

    billy_joule

    User Avatar
    Science Advisor

    Are you hoping to achieve the >0.2mm resolution even the cheapest 3D printers achieve?
    Over what range of bed motion?If it's comparable to repraps (~200mm) you'll need some high precision parts ($$$) to do that.

    There's good reason why threaded rod (or lead screw in more expensive machines) is used pretty much universally, the resolution per $ is much higher than any other option.
     
  5. Sep 27, 2015 #4
    This slider-crank is very easy to build compared to a rack and pinion and the way I am going to move it requires this kind of design.
    I would really want to find a solution based on the proposed design.

    My only solution to avoid the singularity would drastically decrease my Z height capability. Please see the sketch below:
    file.php?185,file=63011.png

    I don't really have much engineering experience so I need some suggestions based on my specific situation.

    I try to avoid rods/screws, however stupid that might sound. Leadscrews cost a lot more than threaded rods and threaded rods aren't that good either. So I am trying to lift a pretty large bed, that would need at least 3 leadscrews and many smooth rods to slide vertically. I would need 3 of these wheels and a bucket of bearings to achieve the same.

    I want to go first for achieving the design and later i'll fine tune precision.


    Thanks.
     
    Last edited: Sep 27, 2015
  6. Sep 27, 2015 #5

    Nidum

    User Avatar
    Science Advisor
    Gold Member

    Mechanism as drawn will have numerous problems :

    Very poor positional control - basically rattling loose at some crank angles and jamming at others .

    Even if it was ok mechanically actually controlling it to make predictable incremental moves would be very difficult . Motion is not even basic SHM - there are big distortion terms to deal with as well and variable loads as seen by motor during an operating cycle .

    Best to take advice of others above and use conventional drives .

    Alternatively reconfigure machine so that heavy table is fixed in height and the much lighter work head moves up and down instead .

    If you really need a bargain basement arrangement then use toothed belts .
     
    Last edited: Sep 27, 2015
  7. Sep 27, 2015 #6
    Exactly about these issues I have opened the thread. I need a way to avoid these situations. Many systems use rotational components and surely they have these weak points too. And they use some solutions to work around those.

    The variable resolution in motion can be quite easily handled in software. The incremental moves are not exactly the same around the circumference of the pulley but the calculated precision can be achieved.

    I am not sure what distortion terms are you referring to and how a stepper would react to variable loads.
     
  8. Sep 27, 2015 #7

    Randy Beikmann

    User Avatar
    Gold Member

    If you want to let the slider move through the rotation center, as shown, I'm not sure how you get around its "indeterminate" motion.
    Let's say the slider is exactly at the rotation center, and stationary. Now you turn the crank; which way will the slider go? It could go up or down, per your diagram. But the slider could also stay at the rotation center, since the crank arm and slider rod are the same length. How do you make it go the way you want? You really can't.
    To make the slider go where you want, it needs to stay to one side of the rotation center.
     
  9. Sep 28, 2015 #8
    Hi, I have found this video of a slider crank-style mechanism that shows a way to work around the singularity. This is quite the same I've seen about Robot Arm on How it's Made but when I saw that video it didn't occur to me that one of those belts was for this.

    I was really hoping for something even simpler to overcome the position of infinite resistance but I don't know if it is possible.

    One more question: if the initial position of the platform was up and it only descended, wouldn't the weight of the platform "help" it decide to go down at the singularity? This is what I was imagining at the beginning. The only problem would then be when I raise the platform all the way up, at that middle point it wouldn't choose to go above the center by itself, more likely it will go back down because of this weight bias.
     
  10. Sep 28, 2015 #9

    Nidum

    User Avatar
    Science Advisor
    Gold Member

    For a crank and con rod to work in a stable and predictable manner the con rod has to be longer than the crank radius by a significant amount . Rule of thumb is that con rod length >= 2 crank radius . Or put it another way point of application of lifting force from con rod has to be outside the swept circle of the crank pin .

    To expand on points mentioned last post the end motion of con rod in mechanism as drawn will be totally unpredictable at some crank angles and force acting on platform will be indeterminate for quite a large part of the full turn of crank .

    The Video is nonsensical .

    The only variation of your mechanism which might work within same amount of space is the Scotch crank .
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Building a "slider crank"-style lifting platform. Need help.
  1. Slider Crank Problem (Replies: 3)

  2. Slider-Crank Using NX (Replies: 0)

Loading...