Hi, I am trying to design a lifting mechanism for a quite large (in area not weight) 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.

Below you can see a sketch of the design.

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. Is this true? What are the weak points of this design? Just a note: there is going to be direction changes in rotation at arbitrary positions.

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.

PS: I don't want to change the slider-crank-based design to a leadscrew or something else, dut to other factors it has to be rotational.

Thank you for your suggestions.

 

If you extend the length of the arm/arms so that they never have to reach a vertical orientation, wouldn't that be the easiest solution? Also, your drawing indicates that you want to achieve 180 degrees of rotation, and that's a very bad idea for this type of mechanism, particularly if you wish it to be reciprocal in motion, as you've indicated.
Have you given any thought to the forces and frictions that would affect your design if the direction of rotation was reversed when the mechanism was in a state similar to the middle of the five images? If this design had to act against any (or even no form of) load in that state, it would probably tear itself apart... Essentially, there would be an ejection force applied to the black dot (pin etc) on your image, but over 50 or so degrees, the friction it works against can be looked at as increasing exponentially per 10 degrees (close enough..) As the black dot reaches ever closer to the center line, this is more accurately described as 'sticktion', and the graph goes off the page.... for this type of linkage, 5 to 45 degrees of angular travel, and no more...

 

This is the exact reason for opening this thread, to get some discussion going around those issues and learn from it.
I have found the following solution, also employed by industrial robots:

It just seem too complex for what I want.

In search for a solution to avoid such defavourable positions I have sketched another linkage, link here.
If my aapproach would work, I could obtain roughly 80% of the diameter at the black dot/pin. But the full range of motion of the solution in the video would be just huge. However, the complexity it requires is beyond what I want.

I wish I could envision what you wanted to say. Does my previous sketch bear any similarity with what you had in mind?

Can you please elaborate on this a bot more?

Thanks.

 

Thorq, for you to go this route, I suspect that you have power available but wish to stay away from hydraulics or pneumatic's.
However, this design is going to give you even more headaches once the dynamics come into play.
Is there any reason that you do not make use of a screw-jack mechanism. It is smooth and you can create a sturdy support with far less unknowns and may-be's.
To run this type of mechanism in a Vertical orientation is horse of another color than using it in a horizontal orientation.
Good Luck

 

I sent you an email, so check your inbox.
Basically, I could fix your "dead spot" problem without gears, chains, springs, etc.; but I would need to know your underlying goals (why you don't want to use something more conventional) to see whether or not my design would even be of value to your project. I have a simple linked solution worked out in SolidWorks, but I am not sure that it would be the best design for your application.

 

Thorq,
I have decided on a compromise: I will provide you the slider solution (and leave out the linear linkage without sliders mechanism we discussed offline). Follow this link and you will see how I eliminated your dead spot with just links and sliders; no gears, no chains, no springs, etc.:
https://www.youtube.com/watch?v=xhnK5-_R0OM
(If you have trouble viewing the movie you might need to switch browsers, but I think this one will play in Explorer or Chrome.)

 

Thanks Jeff, that's a really neat solution. It introduces one more arm in the assembly and this will work for your design. In my situation, as I am trying to get rid of the linear slide too, this will be a bit more complicated and I'll try to explain.

I am thinking of a solution where at least 3 such cranks, each working in its own vertical plane, constrain the lift of the platform on the vertical axis without any slide. In theory this would work but another arm to the already existing two would increase the inherent play in the shafts for each arm which will translate in tremble/wobble of the whole platform.

Here's a link to my 4-cranks sketch:



If everything would be super tight and play would be minimal, this would work.

PS: don't mind the holes in the orange arm, i've reused some of my other designs.

My other idea is to have two cranks in the upmost position while the other two are at the colinearity position at the centre of rotation but didn't give that much thought as I am involved at the moment in another hobby project.

 

Firstly, I am assuming you see the need to keep both the blue radius equal to the orange radius if you want to pass across the input axis. Secondly, I assume you see that the speed/distance of the extent of the orange arm is twice that of the blue arm. With these two points in mind, how do you feel about extending the blue arm out two double the radius while keeping its same radius where it connects to the orange link and using the extended part of the blue arm to hold up the other linkage on the perpendicular side as it passes through the dead spot? This way the first linkage would hold up the second linkage at the dead spot and force it up or allow it to go down as the first motor determines. Each linkage could this be used to hold up the next perpendicular linkage through this problem point.

 

Actually, looking at the way you have your linkages set up, you are already constrained to linear motion. All you have to do is keep everything as you have it, except for making sure blue and orange radii are equal, and move your motors. Keep the anchor point for the linkage the same, but move your motors to the point where blue is tied to orange, fix stator to blue and fix motor shaft to orange. The motors will orbit at the end of the blue link, which means you may need a rotary electrical connection, but everything else is perfect.
Jeff

 

Yes Jeff I already have a constrained vertical motion with 3 cranks, there are the two points I am worried about: one is when the arms become co-linear and the second is play, which will happen how ever tight the rotary joints are. Oh and the length of the arms should be equal so they pass one another, the image is from a version where I kind of settled with the limited-range solution.

I get that putting the motors at the elbow of the rotating arms would actually make the singularity a non-issue and that's a very good suggestion but how is the blue arm going to rotate around its own axis. I think I am missing something here.

Now, the problem is that I am using the large blue wheels as a gearing down mechanism to get some precision from the stepper motors. I would be more than happy to be able to use cycloid gearboxes but that is not an option so my idea is to gear it down via an at least 8:1 gear and pinion approach.

I wonder if I would use a 2 stage transmission, where I would gear down at only one crank and use a timing belt to serially rotate all other elbows as if the motors would be there... this might work. I'll have to think a bit about this, it's just an un-edited thought.

Edit: it seems I can't modify my previous post where I linked the image. I've replaced it with a more recent one i've found...here it is:

https://www.dropbox.com/s/76k1egunj5i9s3q/screenshot.408.png?dl=0