Oke this brain of mine just doesn't stop brain storming ;D. I stumbled on yet another rather simple concept that the math of seems logic at first hand. The idea is simple. You have a pendulum and a wheel. The pendulum gets only attached to the wheel in a favorable place to make it turn.
(https://overunityarchives.com/proxy.php?request=http%3A%2F%2Fbroli.dommel.be%2Fpenduwheel.PNG&hash=d41e92c214ea6ae83c0f09878b24c9ceb045d50a)
Keep in mind that friction is omitted here. But that shouldn't matter since it should accelerate.
Hi broli .. just my thoughts so far on what you have presented [assuming I understand the concept correctly].
When the pendulum falls [under gravity force] it engages the balanced flywheel & they both move, locked together - IMO, the pendulum will fall slower than if it were not attached to the wheel - that's because to accelerate the wheel by direct coupling the pendulum has to overcome the inertia of the flywheel - the result is that although the pendulum has velocity at 3 o'cl, which once it detaches should take it past 9 o'cl, it would not swing up as high as if it had been detached from the get go - then the rotating flywheel [with momentum & angular velocity] re-engages the pendulum & both continue to rise a bit further.
My estimation is that the wheel & pendulum will come to a stop well before 12 o'cl as the total energy in the system has not changed & after ordinary system losses it will be short of enough to get it over TDC.
What does WM show you ?
I just figured a way to simulate it but wm doesn't show what I want :P. The following sim shows a second method I tried to simulate it. The weird thing is that an earlier sim actually had it made go all back to 12 o'clock and keep the big wheel spinning but then the second turn it started to go back. So that was weird.
http://broli.dommel.be/penduwheel.avi (http://broli.dommel.be/penduwheel.avi)
I mean from a simple energy transfer point of view it makes sense in my head. You poor energy in the wheel from the pendulum. The speed of the pendulum doesn''t matter as long as it has some speed at 3o'clock so it can exceed 9 o'clock. Then you use back some of the energy you stored in the wheel but not all of it. Since you stored a quarter turn and using less to get the pendulum back up.
Edit:Before I go to sleep I want to add something else. I really want to convince people that this is supposed to work.
Instead of the pendulum imagen a weight at the top. For discussions sake lets say it's 1? to the right of 12 o'clock so it can fall.
Now while it starts to fall the wheel gains indeed momentum since the weight is causing an imbalance. Now at 3 o'clock this weight vanishes. the wheel now will spin at a constant speed. Then all of a sudden the weight reappears past 9 o'clock. Now the wheel says "fine by me I can use some of my stored energy" But when it reaches 12 o'clock it realizes it actually has some leftover energy. And so on. This makes completely sense in my head whether wm accepts it or not. I just want people to see this as well. It's really a very easy concept.
I should think that the second pendulum in the vid [dark blue] shouldn't get above where an ordinary pendulum let fall would rise too, without engaging any wheel ?!
Another possible method of simulating the action is to place two rigid pint joints on the pendulum arm - then double click each one in turn & in the active when box [bottom] put a time function in to make it active or not e.g. t<1.0 [time less than 1 second] or t>3.0, for example - that way you can do away with the catch [light blue block] that you have [it will have some mass] - alternatively you could place a second latch on the wheel exactly 180 degrees apart to balance out any weight imbalance discrepancy in the flywheel.
Quote from: fletcher on June 29, 2008, 07:07:30 PM
I should think that the second pendulum in the vid [dark blue] shouldn't get above where an ordinary pendulum let fall would rise too, without engaging any wheel ?!
Another possible method of simulating the action is to place two rigid pint joints on the pendulum arm - then double click each one in turn & in the active when box [bottom] put a time function in to make it active or not e.g. t<1.0 [time less than 1 second] or t>3.0, for example - that way you can do away with the catch [light blue block] that you have [it will have some mass] - alternatively you could place a second latch on the wheel exactly 180 degrees apart to balance out any weight imbalance discrepancy in the flywheel.
I actually did use some complex formula to activate the rigid joint but wm doesn't handle it nicely.
if(and(Body[2].p.x >= 0 ,Body[2].p.y > 0),1,if(and(Body[2].p.x < 0 ,Body[2].p.y > 0.27),1,0))
Basically there are two cases where the rigid joint is active.
One is where both the x and y position of the weight are postive.
And the other where the x pos is negative AND the y pos is larger than 0.27 (the wheel has a radius of 1m for size reference). I just did a test run and saw how far past 9 o'clock the pendulum went. The y position of the weight got up to 0.275. So I just activated the joint little sooner namely at 0.27.
But if the rigid joint is not active it will follow the wheel instead of the pendulum. So I don't know what kind of effect that has on it. Here you can see it..
http://broli.dommel.be/penduwheel2.avi (http://broli.dommel.be/penduwheel2.avi)
hmmm .. as far as I know the If & AND functions should work ok - I just prefer to put a measure > time box up & then do the simple greater of lesser in the Active When etc from eye balling the sim each run - don't know off-hand whether the rigid joints will stay with the pendulum arm or the wheel - if you place them on the pendulum first then drag it across onto the wheel it might give more satisfactory results for you - ? - simple is better but you gotta do what works for you.
The other way is to use a center one-way rachet on the wheel [rigid joint placed over the pivot joint] - then when the pendulum gets to its max swing height the rigid joint turns on, effectively connecting the pendulum to the wheel from thereon ?!
broli,
A few weeks back I had a similar idea... I had posted it in the archer wheel thread but here it is again.
The light blue is a heavy wheel. The dark blue gear is like on a bike, you can crank it forward then spin it backwards. The hammer weight should be able to be lifted with minimum effort do to lever action. Its not really shown on the picture but the yellow end of the lever will retract so on the down swing of the weight it does not hit any of the pegs attached to the outside of the wheel. When back to starting positing the lever will extend and another peg will lift the weight back up ans so on.... the animation got a little squirrelly on me while I was making it thats why the dark blue gear starts to float around..Its not meant to look like that. Tongue I would imagine the heavy wheel would need a good spin to get it going then like riding a bike it takes less effort to (or so it seems) to maintain the same speed. I think the wheel in reality would need to be bigger as opposed to the lever mechanism but to show the concept it is what it is here.
Cheers Grin
(https://overunityarchives.com/proxy.php?request=http%3A%2F%2Fhome.epix.net%2F%7Eamiller%2Fopenwheel.gif&hash=b3639c0533696d643fdf16d629cb8efda322d77a)
Oke folks I realized the mistake. When the pendulum falls and drags the wheel it doesn't transfer all of it energy in the wheel. Or else the pendulum wouldn't have an initial velocity at 3 o'clock. And that is the same energy that gets it past 9 o'clock. Now when it gets attached to the wheel it has pretty much has 0 energy and so the wheel has to poor lots of energy in it to get it back up.
@am1ll3r: That's an interesting concept I'll try to simulate it as well.