So when I finally got the pic uploaded, it was a fair template to go by, Fletcher? ... Yes .. I had actually roughed out the first box sim based on your verbal description (not always easy to do). I can tell those who use WM from their logical verbal descriptions usually - the pic confirmed I got it right - just in case I misinterpreted something you described. Then I'd be in do-over territory.
It seems to me I didn't change any of the bodies mass. At horizontal it didn't move. Them I picked up one side maybe 15° and it began to swing like a pendulum, except one side finished higher, each swing had more gpe. Then it flipped and got stuck at vertical.
I take it you just created the objects of the dimensions required and they automatically had the proportional mass as per the program default. And that you didn't change any prescribed masses.
It should have begun moving immediately, horizontal or not. Because its torquing COM is to one side of center of rotation. If it didn't move at horizontal then something was not right in the sim build. The next clue is when you rotated it (I used 20 degrees) and then it did have offset torque COM position and immediately swung like a pendulum - just what we'd expect.
One side finishing with a higher up-swing (increasing in system GPE is aberrant behaviour imo). I've seen it done a few times when parts are overlapped and asked to collide instead of joining at corners for instance (i.e. builder error). The program attempts to 'spit' them apart from their overlap and the sim appears to gain in energy as it moves. Other than that nothing else comes immediately to mind as to why it should gain GPE - if you think you didn't make any construction an checking errors then a build is required imo.
ETA
Some points
I'm not sure which side I picked up, at horizontal it was balanced. It might have been mass1.
No frictions set.
The frame of the roberval was proportional, of like density.
Yes, no pin or pivot frictions were set in my sims either giving most optimistic outcome. I did change all structural bodies in the 3 types to have a mass of just 1 gram each regardless of dimensions. Otherwise the system MOI is adjusted, tho it's no big deal if it is - force of habit !
** As you've probably gathered by now to have a robust sim I simplify and build it as many ways as I can think of (e.g. add truss designs). That way I can check against another type, and avoid builder errors which might creep into one type for instance. Then if I get an abnormal behaviour you really have to start scratching the bonce and digging a little deeper, imo.
Last edited by Fletcher on Tue Jan 24, 2023 12:05 am, edited 1 time in total.
I take it you just created the objects of the dimensions required and they automatically had the proportional mass as per the program default. And that you didn't change any prescribed masses.
One side finishing with a higher up-swing (increasing in system GPE is aberrant behaviour imo). I've seen it done a few times when parts are overlapped and asked to collide instead of joining at corners for instance (i.e. builder error). The program attempts to 'spit' them apart from their overlap and the sim appears to gain in energy as it moves. Other than that nothing else comes immediately to mind as to why it should gain GPE - if you think you didn't make any construction an checking errors then a build is required imo.
I've exploded a few SIMs, I thought I need to tie a rope to that explosion! :)
That final emboldened point is where any working simulation leads. That's not just an opinion.
Another very common reason sims can go into explosive mode is accuracy is set too low i.e. default is 20 animation steps per second. That's fine for small scale and small speeds, usually.
Out of routine of habit I change it to 1,000 or 2,000 steps before even the first run. Especially if the scale is around a meter etc and collision velocities could be high e.g. a pendulum or lever-weight hitting a stop. Then I back it off to between 200 to 50 animations per second if it doesn't explode or gain energy from nowhere.
The reason they explode is because at small animation steps per sec (i.e. large gaps between individual frames individually calculated in sequence) two fast moving parts can overlap into each other (just like I can't push my finger thru a stone wall) before the next frame comes up. So the program goes into "let's force those parts apart with energy (giving them velocity)" and it explodes off your screen.
Solution >> increase animation steps per second until it stops flying apart (not buried into each other), and use Kutta-Merson fast calculation.
A sim program is an analogue for the real-world - but care must still be taken with its build and analysis of results. Tempered by a good working knowledge of Newton's Mechanics and Laws imo.
"If it looks too good to be true it probably needs investigation to see why it is so good" lol.
Last edited by Fletcher on Tue Jan 24, 2023 1:15 am, edited 1 time in total.
What would be interesting for me would be how much energy does a particular mech/sim need to be continuously ‘over the top’.
Logically it would seem the fewest moving parts would need the least.
For example a balanced wheel (no parts) would need the least of all, correct?
From that starting point we could work towards how much B’s wheels needed to continuously make it over the top.
Long test.
Leafy wrote: ↑Tue Jan 24, 2023 12:05 am
I think the mechanism can be simplified into a Roberval pendulum.
The fixed mass 1 have no dynamic impact.
The question is why would we extend mass M out instead of keeping it inline as in the picture.
375C2639-C1F5-4774-A370-313EA165989E.jpeg
true there are many many ways to connect them or turn them in to shapes of all sorts , levers dont have to be bars , what makes it what it is is how the pivots work and how they connect.
Last edited by johannesbender on Tue Jan 24, 2023 10:30 am, edited 1 time in total.
