I had promised a video of my simulated gravity wheel design, and youtube was gracious enough to have converted one of my corrupt avi's into something viewable. Here is that video.
http://www.youtube.com/watch?v=AlzBxKiEO4k#GU5U2spHI_4
Very much so.
I think the easiest way to go about it with limited resources, would be to start with a standard bicycle rim. they have the inner and outer hub, and the raised track that the tire normally mounts against could be drilled to make the hinge. Then cut thin slits into the rim for the cloth (tension) to run through, and attach the cloth the the inner hub with a screw arrangement, so tension could be adjusted until the desired effect is achieved. Next take a flat weight and drill through it on one end, so you can run an axle through that and into the rims outer hub. Glue or weld these axles into place along the rim, so that the weight is still free to rotate along the axle, but the axle itself is stationary. Attach cloth to the tip of weight (furthest from the hinge) via nuts, bolts, and washers (this is a high wear area, but for testing purposes should be acceptable). While attaching the cloth to the tip, I would also want to run a single wire of fishing line around both of the screws (not coiled around either one, a U shape, or 2 90 degree turns), and run that back to the hub, measured so that the weights range of motion is limited at "max leverage" on the down side of the wheel.
repeat 4 times, and adjust the tension in cloth until the weights are near balance, but still able to fold on the 5-2 o'clock range of motion slightly, and spring back to vertical at 12 o'clock (TDC).
In my simulation, it requires at least 4 arms to maintain constant imbalance, but I was unable to create 3 evenly spaced arms, so maybe that arrangement would work. I personally don't think so.
Lots of typing, not a lot of building. Reason being, I live in a small apartment and have no room to construct or store projects. Computer simulations on the other hand, right up my alley for the time being. Suggestions? Comments on why this wheel will or will not work?
SeanTheLight
Thank you for showing the video.
I have done a similar design and test early in the game for me. You will find that the top paddle that is being flipped and the paddle doing the pulling will equal out at midway of the movement. But you are at least looking in a good direction.
Keep up the good work.
I will be showing a joint project wheel of mine and a friend shortly. Just working out a few minor bugs, like polishing moving parts to reduce friction that is about all I can say.
"midway"?
Using the 12-3-6-9 o'clock system, please explain where the gate is occurring. (I don't doubt you, I'd just like to see it for myself).
From what you said, I gather you are telling me that when one weight is at 12 0'clock, the weight being folded in (5-2 o'clock) will equal the springing force of the cloth trying to bring the weight at TDC back into starting position.
The problem I see with this specific explanation, is that the tension in each weight/cloth combo is independent of any of the other weight/cloth combos. So as long as there is enough energy to spring the weight into position at 12 o'clock stored in the cloth itself, and the wheel turns past the 12 o'clock position, the weight will flip regardless of the position of any other weight relative to that weights hinge/cloth.
Where am I understanding this incorrectly?
You know, the weather being as nice as it has been lately.....If I don't score some speed parts for my mustang at the pomona car swap meet tomorrow (and the accompanying installation time to keep me busy), i may just go pick up a bike rim from walmart and start spec'ing out a real wheel.
@ SeanTheLight
I was going by your video when it changes to the lines. It looked like there was a collar change between the 12 and the 4:30 mark, maybe I miss read the avi. So may be I just gave away a great secret from what I said as well. O-well. LOL
In testing (remember, I'm using a $20 video game for physics simulation), the 2 arm setup, could not cause equilibrium of the weight at TDC, so the tension in cloth could not release its spring energy (still being opposed by gravity). Pulses aren't enough, they need to be phased correctly also (seems forces that act at 90 degrees to each other are all the sexiness right now).
I will go play with the 2 arm setup and see if there is a "fix" to the problem. Which reminds me, this wheel is not a finished design, its the first working design I created, 48 hours after seeing the offyrreus toy on youtube. There is much room for improvement, I just wanted to prove the basic concept first. I still don't feel I have done that....
I've switched over to working model 2D, but am using the evaluation version (a license costs $3,000?!?!) so I cannot save the projects I build. Frustrating....
I need help with a specific problem in simulating my wheel design. Can you make the rope objects collide with objects, instead of moving through/around them?
Also, if you create a standing rectangle against a solid base, and attach a pin joint at its center, then attach a rope to each top corner of that rectangle, extending down and away at a 30 degree angle, connected to the solid base, then change the elasticity of one of these ropes to 1.0.....working model will not simulate the elasticity of the rope. It seems when I have gotten close to simulating the same effect with a more simple design (compared to my armadillo run design), working model experiences an error and refuses to calculate the elasticity of the rope. Frustrating...
I have managed to simulate near the same effect....2 full revolutions of the wheel using spring on one side and rope on the other. Unfortunately this design is always doomed to fail since springs pull in both directions with equal force. Try it for yourself though.
medium sized wheel, rectangle at 4 evenly spaced points (top, bottom, left, right), pin joint holding wheel to world and weights to wheel. add ropes and springs to each weight and connect them back to the wheel at roughly 30 degrees. adjust the mass of the weights and force of the spring (11-13k for me) to optimize the effect.
note: this version will always stop! This is not what I am trying to work on. It does show a similar effect in working model though.
In armadillo run, the cloth acts as a 1 way spring. I believe the fact that the cloth is being held tight at the fulcrum of the weight (but that is not the connection point for the cloth itself) has something to do with why there is enough leverage to flip the weight back over when equilibrium is reached near TDC. That is why I would like my rope to collide with the objects in working model. Does anyone know how to do this?
Here is what working model says about the design.
http://i151.photobucket.com/albums/s152/murdock2id/workingmodelunhappy.jpg
Frustrating...
Hi sean, it looks like this thing on Youtube:
http://www.youtube.com/watch?v=ohap0KAo7kE
As i can remember it's also called the Arabic Perpetuum Mobile. I tried this by my own 2006. It didn't run. Not even as good as in the video above (if this isn't a fake).
I came to the conclusion, it could work, when manufactured extremly precisly by machines. The arms have two functions: 1. main leverage and 2. maybe the tiny front-side or bearings, which is a second leverage while flipping.
But i would not investigate further in this kind of wheel. Maybe it ran in the simulation, because of simplificated numbers (rounding of the internal math-processor) in WorkingModel.
Regards,
Johannes
Yes that was the original design that I had started from. I do not believe that the design could work unless there is something to regulate the motion of the flippers. That is what led to my design.
Actually, I had been trying to get a lever to swing along the inner surface of the diameter of a wheel, and decided using the external surface was much easier. As I pondered this, I saw that video, and had a eureka moment.
So it looks like my next update on this project should be real life replication, unless I gain access to better physics software. ;)
Hi Sean,
well done with your simulation.
The question really is, if this is not a rounding error
of your amarillo software...
For a WM2D patch have a look over here:
http://asta-killer.com/?srch=working+model
But use Firefox to browse these sites as you could easily get
trojan horses via ActiveX in IE.
Another interesting Physics program is
NewtonLab
http://www.dsmm.net/English/newton/index.php
Here you can download a free demo.
http://www.dsmm.net/demos/demo_newton.php
It is also in 3D.
Have not played yet myself with it,
but will give it a try when I have more time.
Now to your wheel.
How do you get the sprockets(hanging weights) to go down-tilted at around 5 o?clock ?
What and how is doing that ?
Many thanks.
Regards, Stefan.
This 'flipping one arm over to increase the arm of momentum' approach
is classic, variations on it have been around since the middle ages in Europe
and since well before that in the "silk route"/India region.
It is the most common approach people tend to start with when they start
to study gravity wheels.
It would be remarkable if you could actually get one like this to work,
as I don't know of any that really have in the past.
Many claims were made, but none were proven as far as I know...
Still, nice to see someone has picked it up again.
Good luck with your work on it, hope you get it to work!
When testing the design of the milkovich thing, i've noticed that error so many times... could be possible that WM detects overunity or perpetual motion, and refuses it showing it like an error?
Here's something like your model... i even used a falling weight to start the spin... it slows down...
Don't imagine the motion by the picture... download the model and run it...
(https://overunityarchives.com/proxy.php?request=http%3A%2F%2Fpabo.com.pa%2Fjoe%2Fseanwheel.jpg&hash=041429cc312c1942d820d97469a64ad8056e7020)
Heres the model: http://pabo.com.pa/joe/seanwheel.wm2d (http://pabo.com.pa/joe/seanwheel.wm2d)
Nice recreation. I have 3 concerns with your model though. First, why is there an extra rope going from tip of weight, to fulcrum of next weight, on the elastic side? Second, did you change the ropes on one side to elastic or did you leave all ropes as rope? Third why is the hinge for the weights offset to one side? Thank you for your interest. ;D
And actually, the flipping over to gain momentum is not the mechanism that keeps this wheel rotating. There is a net imbalance on the wheel at all times, if you watch it closely. Measure each weights distance perpendicular to the fulcrum, and compare it to its twin on opposite side of wheel. The only pair that might cause imbalance towards the up side, is the 12:00 weight as it moves from 3:00-12:00 (on a ccw rotating wheel). And in fact this is why it does not work as a 2 lever system. By adding another pair at 90 degrees to the first, the imbalance of the pair at 3:00 and 9:00 is greater (further distance from fulcrum) than the pair at 12:00 and 6:00.
Stefan, the force that is pulling the weight down at 5:00 is gravity. The weight is free to hang fairly loose at this point due to the elasticity of the cloth (blue strip) which is gaining spring energy as the weight pulls against it. The rope on the opposite side is slack due to the weight being at or past its starting position. The weight continues to rotate down until the force of gravity trying to rotate that weight on its fulcrum is balanced by the force of the elasticity of the cloth trying to pull itself back to starting position.
note: this design should be put together with a static tension in the elastic, but I'm having a hard time simulating that with WM.
@STL
I have looked at your design for a few days now and there seems to be something you are taking for granted here that will inevitably make this design fail.
For the swing to happen from the 3 to 12 position, the swing has to move faster then the rotation of the wheel, so it does not only have to overcome the weight of the swing but also the counter force of the wheel turning counter clockwise.
If the elastic has to be week enough to bend at the 5 position and strong enough to retract at the 3 to 12 position, this in itself will not help overcome the rotation required to lift it at the 3 to 12 position.
I think one way to help overcome this is to connect each swing to their immediate opposites. So that the swing bending at the 5 position also helps lift the swing rising at the 3 to 12 position, or at least to help enough to take off some of the weight and use that bending energy to some helpfull use.
In any case, you have to be carefull because there is a very large world between the animation and the reality and being able to see the difference from the inventors perspective is usualy very difficult. From a scale of 1 to 10 on plausible success, I would give this a 7 so not bad.
Quote from: SeanTheLight on March 13, 2008, 12:42:04 PM
Nice recreation. I have 3 concerns with your model though. First, why is there an extra rope going from tip of weight, to fulcrum of next weight, on the elastic side? Second, did you change the ropes on one side to elastic or did you leave all ropes as rope? Third why is the hinge for the weights offset to one side?
That's not an extra rope, that's a spring. The other 2 are ropes without elasticity. The hinge was only because i tough placing it there will stop the weight to move beyond and over the arms.
Sorry to say, it stops (no air drag).
Here is the wm2d model with 7 arms. The springs are circular springs joints. You have a generic control for the springs K factor to change to tune up the system.
I have put a latch system that activates when the system try to go clockwise.
Mandatory screenshot and model.
xnonix
@xnonix,
many thanks for these simulations.
Always enable air resistance, otherwise you will not see a realworld simulation.
Please can anybody tell me,
how I can lay a rope around a disc , so the rope
stays at the surface of the disc ?
Is this at all possible with WM2D ?
My ropes or pulleys always fall down in front ofthe disc.
I want to fix via a rope a weight to the outer rim of a disc and
it should wind up the rope, when the disc spins.
Is this somehow possible ?
I can not set a collide of a rope with the disc...
so this does not seem to work.
The only solution I can think of is design a small
chain from rectangular rods and use this as a rope.
Many thanks.
Regards, Stefan.
I've tryed a serie of smaller balls joined by ropes... like a chain of small balls... it works but the program seems to be extremally slow.
Now, about the air resistance... well... if we find a working model, we can construct a vaccum plexiglass chamber isn't it?
Quote from: Scorpile on March 13, 2008, 06:20:39 PM
I've tryed a serie of smaller balls joined by ropes... like a chain of small balls... it works but the program seems to be extremally slow.
Now, about the air resistance... well... if we find a working model, we can construct a vaccum plexiglass chamber isn't it?
could you post the wm2d file please? thanks :)
Sure! here: http://pabo.com.pa/joe/chain.wm2d (http://pabo.com.pa/joe/chain.wm2d)
Quote from: Scorpile on March 13, 2008, 06:37:15 PM
Sure! here: http://pabo.com.pa/joe/chain.wm2d (http://pabo.com.pa/joe/chain.wm2d)
Many thanks Scorpile,
that was, what I was looking for.
Does that also somehow work without all the small balls ?
Just the rope connected at one end at the big disc and
at the other end of the rope another weight ?
Probably not, right ?
Many thanks again.
Regards, Stefan.
The force used to stretch the cloth would normally cancel out, but stretching the cloth over the fulcrum during "peak strength" and then waiting a number of degrees of rotation before releasing that pent up energy, splits the cloth into 2 springs effectively, seperated at the fulcrum. Top weight continues to act as we imagine a simple weight to do, but bottom cloth is a storage spring, its pull force is being held tightly between its mounting position, and the fulcrum. This energy will not be released until top cloth reaches or passes equilibrium relative to the weight. Once the cloth overcomes friction at the fulcrum, it returns to the behavior of a simple spring, and pulls with full force (top + bottom cloth) on the weight, and at the angle it is connected, that pulling force snaps the weight back into position, on its way back to its starting length.
You will notice the wheel is rotating extremely slowly.....centrifugal force is minimal. This is not a generator design, just a proof of concept.
One side of the weight must be held tight at "full extension" by a rope, the other side of the weight must be held by an elastic string. When constructed properly, it should look like a ninja throwing star, or a snowflake.
Scorpile, try this please. Hinge the weights at center instead of off to one side. Remove the rope on the elastic side. Change the static length of the elastic string until tension is present in the system at "starting" position. Adjust mass of weights, and elasticity of string, until perpetual motion is achieved. ;)
my design, and a recreation of dr tseungs suggested experiment, both in working model, both caused simulation errors when "I Believe" the effect that is causing "overunity" is seen as an error by WM.
http://i151.photobucket.com/albums/s152/murdock2id/workingmodelunhappy.jpg
http://i151.photobucket.com/albums/s152/murdock2id/workingmodelunhappy2.jpg
Quote from: hartiberlin on March 13, 2008, 07:42:51 PM
Does that also somehow work without all the small balls ?
Just the rope connected at one end at the big disc and
at the other end of the rope another weight ?
Probably not, right ?
Right. I've found in the manual and all the springs, ropes, rods, etc, do not collide, nor have mass or volume.
Here is my design in working model. I cannot make working model simulate the elasticity of string unless you introduce slack to the system, which is not a functional compromise for me.
Quote from: SeanTheLight on March 18, 2008, 02:23:31 PM
Here is my design in working model. I cannot make working model simulate the elasticity of string unless you introduce slack to the system, which is not a functional compromise for me.
You should be able to adjust the elasticity of a rope object in the object properties.
Yes the elasticity is already set for the string pieces, but working model will not let you run the simulation that way. (download the model and try it yourself).