Multi-body dynamics sim software

[jimminy121212]

Hi all,

I've been looking around for a simple software to test mechanical assemblies, typically ones using a motor to drive some gear or pulley system. This is for garage scale projects, nothing really big or really small.

Ideally it would be something that would be able to give me the load required to do some work. Say, if I wanted to use a motor to turn a gear that would move something that ultimately elevated something with a given weight, it would be smart enough to calculate the load on the motor.

It could be very diagrammatic for my needs, I don''t need to see representations of the actual parts, but I think would need to be able to operate in 3D. A rigid bar could just be a line, a gear could be a circle of a certain diameter, a cable could be a line, joints between parts could be represented by a hinge element that constrains the rotation to a given axis, etc.

I guess it would basically be inverse kinematics that could analyze and report the forces acting on each part of an assembly and show how it all moves.

It seems like something must be out there like this, but everything I see is way more than I want to bother with. Solidworks, MCS Adams, etc.

I doubt something that simple but so complex exists, but any help appreciated.

 

[Baluncore]

For small systems it is quicker to compute an immediate answer than to fully specify the system.

If you know enough engineering to apply a physics engine, then you also have enough information to compute, from the fundamental relationships, the answers you actually require.

The physics engine you are describing is what actually goes on in an engineers head.

 

[serbring]

I agree with Baluncore, but you can try openmodelica...it's free and I think it does anything you need.

 

[jimminy121212]

I agree too. The problem is I'm not that smart.

I thought I would try creating a spreadsheet that would allow me to sort of arrange different "modules" (gears, pullys) with input and output calcs, but I got stuck along the way. maybe I'll pick that up again, it shouldn't be THAT hard.

In the meantime I'll check out that software, probably done by someone smarter.

Thanks!

 

[jimminy121212]

Yup, it was definitely done by someone smarter. Too smart perhaps. I would need a graduate level course to figure that out on my own. I guess I'll be looking for tutorials.

 

[Baluncore]

Don't put yourself down. A physics engine can be very clever and require every possible detail, but that is way more complex than you need for a first analysis. If you understand the concept of a physics engine then you are bright enough to solve your problems from first principles.

Give it a try, write out the numbers you know and a box for the answer you want. Then relevant equations that relate to the energy flow through the mechanism you are considering. If it works, automate it by plugging it all into your spreadsheet.

A physics engine is not an alternative for experience and the understanding of engineering principles.

 

[jimminy121212]

ok, I'll give myself a little more credit then. Here's what I'm trying to solve currently, and this is the typical type of project I do.

I'm trying to come up with a fun way to do a murphy bed. I've gone back and forth on which direction to take it, the standard way where it's hinged at the head against the wall and the foot goes up, or having the head slide up while the foot slides back. The second way is more fun but in the end is less practical for various reasons.

The wrinkle is that I want to do it with a motor and make it a push button affair, and I want to do it so it's a little more interesting than pully's and cables.

So, this entails:
1. Coming up with the a novel mechanism.
2. Figuring out what the loads will be, closing it, and opening it.
3. Sizing a motor for the load.
4. Determining an operating speed for the movement.
5. Converting motor speed/RPM down to movement direction/movement.
6. It'll have to operate like a garage door, in that each push of the button either raises or lowers and knows when to stop, and if something goes wrong will just stop

After much thought, I'm thinking of a system where the motor turns a ball screw or a lead screw that either pulls or pushes at the edge of a wheel attached at its center to a heavy hinge that will rotated the whole bed 90 deg. at it's head.

I'm assuming the bed itself is about 100 lbs. of uniform load.

I figure a ball screw @ 1/8" per turn will move the nut about 1" per 8 turns. If it has to move 12" or so to do the job that means 96 turns, over about 30 seconds, or about 192 rpm.

I come up with a lot of little issues to tackle, such as how to stabilize the thing when it's upright, or once it is, how to bring it back down without crashing down, etc.

There it is.

 

[Baluncore]

Ball screws with a 5mm pitch are available at low cost on eBay.
Also, short actuators that use ball screws are available, 12V powered.

 

[jimminy121212]

Great idea! I think that's the way to go. I obviously am new to the world of linear actuators, they are a no-brainer for this application. The question I have is how they are controlled, start stop and emergency halt, etc. Seems they

From there I have some more thoughts:

I think to lift something like this requires probably about 400-500lbs, assuming a uniform load of 100lbs, and a lever arm of 10" opposite the load. I can easily counterbalance this and use a lighter duty actuator I suppose.

Everything else is pretty well sorted, sizing and sourcing the actuator is really the only concern other than fabrication.

I'm going to have to come up with a way to complicate this more...