can this be done?

[james.lynch]

sorry for the simplified drawing .. I cannot share the original model


View attachment 1004
So, I have a 3 part assembly, Part 1 fully constrained, part 2 is connected by a pit joint(can only move say 90 degrees clockwise from the horizontal) and the same applies for part 3. <?:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" />

Now say I apply a servo motor to joint

 

[james.lynch]

Anybody have any idea at all?

 

[jrobi]

good luck

 

[Speling]

From what material is that green thing?

If that is something like rubber - I don

 

[Speling]

I forget to ask

 

[james.lynch]

the "green thing" is like a rope, (a tether) a pulley ropewhich is routed around A,B,C & D wheels.


no sheetmetal parts.


simply 3 rigid parts connected together buy pin constraints. I should be finished modeling tomorrow evening.. you're right, I should knwo more then, but I do have a simplified example (similar to the picture above) modeled that I've been playing around with..


James

 

[AHA-D]

If you ask me this doesn't even need "mechanism" functionality to make it work. Forgetting the springs for a moment, the only thing varying is the amount of rope pressed against (and around) A and D. This is dependent on the angle between the parts. All other lines and arcs are fixed because the tangents between the circles are fixed. And if the radius of the pulleys is the same then there is a one-to-one relation between the length of the arc and the angle of movement. Meaning that the same angular displacement will occur between part 1 and 2 as between 2 and 3.


But looking further at the sketch I think I see something conceptual wrong. If part 3 can rotate around D, there's nothing holding it up. The green rope should be under tension to hold up part 3. Since the rope goes underneath the pulley it will be "compressed", something ropes don't do. So IMHO part 3 will fall vertical under center D. This will take away tension between part 1 and 2. So part 2 will start to fall also until all the slack is taken up.


Alex

 

[james.lynch]

well it's a simplified sketch.. it's sprung return.. sorry should have stated that!


again, you're right about the simple relations - very basic if you want to siply relate the movememt of A and D without B and C.. but they are the who point of the concept.. I want to find ot what kind of forced are going to be on rollers B and C so I can specify the spring stiffness..


James

 

[phoxeoy]

Looks like a robotice finger to me. I think the tried and true method would apply here. Build a prototype and start trying springs.

 

[james.lynch]

yeah prototypes are quite expensive though.. I really would prefer to have a go a some simulation and optomisation...

 

[phoxeoy]

Well, being that the Spring is most likely an off the shelf item and if everything alse has been spec'd it wouldn't be a big dealmaking one for testing. Your going to have to make on sooner or later.


But I understand your position and I wish I could be more help to ya. Call PTC, they might help

 

[james.lynch]

but I'd liek to optomise the positioning of B and C and the diameters of A and D also... cant really do that with a physical model!


bummer! I may contact PTC, but I;ve neverdonethat before, we have a maintenance contract, (educational license) but I'm not sure about the costs involved..

 

[kvision]

I usually don't guess at things but I don't have the time to work on this right now so I'm gonna through out an idea and you can try it if you want.


I think you could make a curve feature and create a perimeter dimension for it. In the BMX optimization make sure to constrain the length to a fixed number. This would help drive you to only those solutions of the length you are interested in.

 

[james.lynch]

Kvision,


yeah, I was playing around with that alright, I must be doing somehing a little silly tho'.. I can't seem to get it to work.. I'll keep trying tho'

thanks for the tip!


James

 

[killingmesoftly]

james - for the elementary idea of the design, I don't see a need for two springs, only one if all it does is put tension on the tether, and allow the 3rd part some movement disconjunct of the first part (something like that )

 

[james.lynch]

for the elementary idea of the design, I don't see a need for two springs, only one if all it does is put tension on the tether, and allow the 3rd part some movement disconjunct of the first part (something like that )

The angle with the tether leaves the first part, and also connects with the third partare quite important, also the distance betwen the two primary axisis quite small, so to get the maximum possible increase in length, two springs werechosen..


James

 

[MartinH]

I really don't know anything about Pro/E's tools for working with mechanisms, but there are many software packages dedicated to that, like MSC.ADAMS. A much simpler package that would probably still work is Working Model 2D. But of course you can't use the software if you don't have it.

Even without the packages, it looks like a fairly easy system to analyse using any multibody dynamics technique (Langrangian, Neutonian, Virtual Work). If you just want the kinematics (ie, positions of everything) it is very simple, but even a full dynamic analysis (forces and accelerations too) should be possible by hand. If you know someone who can generate and solve these equations, no problem, but if not, I guess it doesn't help.

 

[james.lynch]

MartinH, I can work it out by hand, ( I havent actually done it but you're right, it's simple enough) the reason I want to do it in Pro is that I want to be able to take advantage of MDO and BMX ro optomise the entire assembley..

 

[slashct]

James,


Looking at this problem in ProE world there is one extra degree of freedom in the belt. I understand what you are trying to accomplish, the problem is that you can hinge the joints and make the shape follow the points of where the springs would be and it will work making the belt follow some fixed points. Now when you add the springs into the equation, they can move based on the tension of your belt. Once you do this ProE does not in the mechanism package have a way to simulate this (that I know of). Maybe in the advanced package there might be something... I'll look at it further when I get back to work. If you wanted to simulate this in mechanism it can be done by using displacements (from calculations) on the springs....