Perpetual motion machine examination rules, please.

[AgingYoung]

Nathan,

I don't really know. I do see your point and I think it's an excellent question. The idea of whether the simulation is approximating motion in larger than reality steps or if it's basing the motion on what would actually be the case is a very good idea; I can see how step approximations would give results that could be very misleading.

I didn't want to take GzuzKryzt's suggestion of starting a thread in the science section because it slows me down in what I really want to look at. It might be that I do need to slow down and think about what I'm seeing. Your question is an excellent example of a reason to slow down. If I don't know where I'm at and I move like crazy somewhere else I probably won't know where I end up at.

For that reason I think I'm going to take GzuzKryzt's suggestion. There's no reason not to benefit from the very sharp people on this forum to help me look at some of the ideas I have. It might take me a day or two. Thanks for asking the question, Nathan.

Gene

 

[AgingYoung]

Nathan,

It's been a couple of days and I haven't started a new thread. I've looked at some details of wm2d and it uses ....non-uniform rational b-splines (NURBS) geometry, and it may be approximating but the method it uses is definitely Kutta-Merson. I know there's a concept of single and double precision but from what I understand you need a cray to have double precision. It might be that all cad models approximate. It seems to me that I read there are times when wm2d can do calculations rather quickly based on the idea that the motion is easily defined. I just rotated a beam 10 to the minus 22 radians. I couldn't see it move. The other day I made two beams 100 meters long and pined them together and to the background then rotated one beam 1/1000 of a degree. I could see the movement at the end of 100 meters.

I just made two beams 1000 meters long and moved one 10 to the minus 23 radians and couldn't see the movement at the other end so it could be that amount of precision isn't factored into the calculations. I'll try 2000 meters.... no difference there either. That amount of rotation is minuscule and the tangent seems to me to be 5.55 repeating to the minus 26th. I might have to make the beam end in bfe before I could see any movement.

I've been trying to make a wheel turn by gravity for the last 3 years or so; I'd have to look at some notes to pinpoint it. I know self proclaimed mechanical geniuses that are in their 70's that have been looking for the same idea most of their lives. So far I've done as well as any genius I've ever met.

A point I'd like to make is that I don't think the idea of approximation vs. real representation of reality is a factor in trying to model with a simulation. I had thought some time ago that if you could increase the rate of acceleration (moving a mass faster than gravity would move the mass) to reposition mass on the wheel that it might work. By increasing that acceleration the product of it and the mass would give you an increased force or torque. The major problem with that idea is that if you could manage that acceleration of the force of gravity you'd put the mass further from the center of rotation and would be faced with the task of accelerating it closer to that center. I think that task is insurmountable.

Another idea that I've had is the idea of movement that is opposite and equal. The flaw in that reasoning is that no matter how many weights you have moving relative to one another your center of gravity remains unchanged or you end up gaining some torque more quickly by giving up lesser moments of torque. Basically you eliminate the top part of the wheel or rotations from 0°-30°. A real problem with that is that you're giving up time. One particular self proclaimed mechanical genius I read has the idea of 'center of gyration'. He's looking at the idea of causing equal and opposite movements that increase or decrease the density of portions of the wheel thereby changing the rate they spin. By pulsing the wheel in that manner he's hoping to cause rotation but I think that idea is a waste of effort. He needs to supply some power to cause those equal and opposite movements to obtain a difference in the centers of gyration and I doubt seriously any power that idea could generate wouldn't be sufficient to cause it.

I've shared some of the ideas I've had for a couple of reasons. If it is possible to cause gravity to rotate mass continously, before anyone could hope to find a solution they need to have some very specific ideas. The method of trying this or that configuration without an ability to verbalize what you're attempting is pretty futile. Attempting to model what would be a description of an out of balanced wheel is equally fuitle in my opinion (ie causing the cog to be mostly on one side of the center of rotation).

I am currently looking at several ideas that I'm going to attempt to model. I've tried to simulate them with wm2d but they're too complex for it. I have a method that I look at curves that I call clay-alculus. When I'm looking at 2 dimensions I use clay of equal height to fill a curve so I can compare different curves. If I'm looking at 3 dimensions I put the appropriate relief in the curves (like little hills); when I'm looking at 4 dimensions I take the hills and throw them at the wall.

Gene

 

[Thing]

A point I'd like to make is that I don't think the idea of approximation vs. real representation of reality is a factor in trying to model with a simulation.

I don't speak for Nathan but here's my point. It's exceptionally unlikely that you will ever simulate a perpetual motion machine of either the first or second kind. But if you do, it will only be because of the finite accuracy of the program. Simulation may be a quicker and cheaper way of convincing yourself that your designs don't work than making physical models, but your chances of designing a working perpetual motion machine this way is zero. Not tiny small, but zero. Asking for a system that obeys Newtons laws, that can be physically built, and that produces perpetual motion is, as Leonard Feather said in a different context, like asking for a piece of green chalk that's red.

 

[nathan]

It's been a couple of days and I haven't started a new thread. I've looked at some details of wm2d and it uses ....non-uniform rational b-splines (NURBS) geometry, and it may be approximating but the method it uses is definitely Kutta-Merson.

I am only familiar with NURBS being a geometry representation - namely surfaces in 3D space. I've no idea whether they can be pressed into physics simulation -- do you have a reference?

I know there's a concept of single and double precision but from what I understand you need a cray to have double precision. It might be that all cad models approximate. It seems to me that I read there are times when wm2d can do calculations rather quickly based on the idea that the motion is easily defined. I just rotated a beam 10 to the minus 22 radians. I couldn't see it move.

hmm, it appears you are unfamiliar with floating point. I don't know whether your CAD software uses single or double precision floating point (IEEE754 32 and 64 bit). I don't know whether your using an i86 or not, and if you are, I don't know whether your using the i87 math's engine or the sse instructions. The former uses long double intermediate precision. Anyway, 'rotating by 10^-22 radians' is going to tax those representations. Take the long double format, that has 80 bits of mantissa. Rotation around an axis involves sin & cos, in particular 1 - sin (or 1 - sin^2, I can't recall exactly). 10^-22 is about 2^-73, so there are only 7 significant binary digits in '1 - sin' thus your accuracy is going to be around 1%. For the other formats (which I think are much more likely to be involved), the rotation is going to be indistinguishable from zero.

I just made two beams 1000 meters long and moved one 10 to the minus 23 radians and couldn't see the movement at the other end so it could be that amount of precision isn't factored into the calculations. I'll try 2000 meters.... no difference there either. That amount of rotation is minuscule and the tangent seems to me to be 5.55 repeating to the minus 26th. I might have to make the beam end in bfe before I could see any movement.

This is ridiculous. To see a rotation on your computer, you need the end of the beam to move at least one pixel. For a rotation 10^-22 radiansto produce a 1 pixel movement at the end of the beam, requires a beam of 10^22 pixels. Thus you need to zoom in on your 100m beam so that the pivot point is some 10^19 monitor's worth to the side (approximating 1000 pixel wide monitor). If your monitor is about 33cm wide that means the pivot is 3*10^18m to one side. This is 1/10 of the way to the center of the milky way! (milky way radius 6000Ly, ly= 10^16m, sun is 2/3rds from center). What magnification level are you using?

A point I'd like to make is that I don't think the idea of approximation vs. real representation of reality is a factor in trying to model with a simulation.

I disagree. Numerical analysis is difficult and many approximations come into play. What you are attempting to do is overturn the entire experience of human experiments and over a century of robustly understood physical laws. It is far more likely that any perpetual motion that your simulation exhibits is far more likely to be an error in your simulation than a previously undiscovered feature of reality.

I'm sorry, but I could not understand any of your remaining comments. Perhaps you could start by defining 'clay-alculus' in rigorous terms familiar to a mathematician or physicist.

 

[AgingYoung]

Nathan,

• I am only familiar with NURBS being a geometry representation - namely surfaces in 3D space. I've no idea whether they can be pressed into physics simulation -- do you have a reference?

• This is ridiculous. To see a rotation on your computer, you need the end of the beam to move at least one pixel.

As you can see in that graphic body two is selected. That's what is shown in the properties box and also you can see the small square on the body on it's upper right corner. You can see from the property box (rad) that there is a very slight rotation. Also when you look at the body you can see it's upper edge is jagged indicating that it's not horizontal. You can't see the exact length of it at the bottom; it's truncated. It's 3.000 e +004 meters long.


When I preview this post sometimes the images show and sometimes they don't but I'm going to assume they will show up. In the above image you should see the selection of the first body with no degrees of rotation. You should also see that the bodies are 5.510 e +010 times normal size. You do need to magnify them to see any movement; that's a big 10-4.

The beautiful thing about a simulation is you can go ...

• ... 1/10 of the way to the center of the milky way!

and back all in an evening.

• It is far more likely that any perpetual motion that your simulation exhibits is far more likely to be an error in your simulation than a previously undiscovered feature of reality.

In a gravity wheel if the masses in the wheel aren't moving then the wheel won't move other than to move such that the collective centers of gravity of the masses and the mechanisms are at their lowest possible point. It's obvious that the masses move within some range. In the above examples with the bodies it's clear that you can define them to quite a degree of accuracy. I'm not sure if the simulation is that exact or not yet my point is that any inaccuracies gained when the masses move from one end of their extreme will be given back when they move to their other extreme. That's why I don't think that any inaccuracies in the simulation will matter. Also those inaccuracies don't amount to nil.

• What you are attempting to do is overturn the entire experience of human experiments and over a century of robustly understood physical laws.

This is inaccurate. A gravity wheel won't violate the idea of the conservation of energy. Energy won't develop then expend in a linear fashion. It will exhibit cycles of entropy. I would disagree that there is a robust understanding of the laws of physics. I think it's more like rehashing and force feeding what is understood. What I am actually attempting is to take what is known and look at it from a different angle. I've thought about Newton's third law of motion....

• For every action, there is an equal and opposite reaction

I need to think about it a little more but it's probably not a factor in a gravity wheel.

Clay-alculus is a method of approximating curves with clay to determine the area under them and to compare different curves (ie the various curves of a dampened oscillation) in an effort to see the rate of change between a sequence of curves. Sometimes it's good to know if the rate of change is linear or if it's accelerating, etc. Probably not the sort of rigorous terms familiar to a mathematician or physicist but Edison might have been impressed.

Gene

 

[AgingYoung]

Nathan,
When I try to add more than one graphic per post it won't show....

 

[AgingYoung]

Nathan,

Body one...

You can see that the selected 'body one' is behind the other one by the little black square above view size (indicating it's the one selected).

Gene

 

[AgingYoung]

Thing,

• Asking for a system that obeys Newtons laws, that can be physically built, and that produces perpetual motion is, as Leonard Feather said in a different context, like asking for a piece of green chalk that's red.

I do agree. I don't expect that a gravity wheel will be subject to Newtonian physics. There are two ways I can defend that statement. One is to explain the other ideas that I have that I haven't mentioned or the other is to build it. I'm in the habit of looking at an idea rather thoroughly before I share it so I think I'd better start building.

Gene

 

[AgingYoung]

Nathan,

The graphics sometimes show then sometimes don't. When I edit the message the image link is there but when I view the thread they're not there. Initially they are.

I posted them at....
http://community.webshots.com/album/551853109AiFhqf

Gene

 

[nathan]

Nathan,

• I am only familiar with NURBS being a geometry representation - namely surfaces in 3D space. I've no idea whether they can be pressed into physics simulation -- do you have a reference?

[qimg]http://image64.webshots.com/164/4/99/27/2627499270093730477MovAkt_ph.jpg[/qimg]

Umm, that's not what I meant. That appears be describing the use of NURBS for solid body collision detection. Not the use of NURBS to do the equations of motion. You should realize that NURBS are evaluated approximately and used to render whatever geometry they are describing. For a high fidelity rendering, you want the approximation to be at the sub-pixel level, thus it becomes negligible wrt to the granularity of your display. However, it is still an approximation and you are not solving things analytically.

• This is ridiculous. To see a rotation on your computer, you need the end of the beam to move at least one pixel.

As you can see in that graphic body two is selected. That's what is shown in the properties box and also you can see the small square on the body on it's upper right corner. You can see from the property box (rad) that there is a very slight rotation. Also when you look at the body you can see it's upper edge is jagged indicating that it's not horizontal. You can't see the exact length of it at the bottom; it's truncated. It's 3.000 e +004 meters long.

I see that image, it appears to have a slope of 2 pixels across half your screen. If your screen is about 1000 pixels wide, that means it's slope is 4e-3 radians.

• What you are attempting to do is overturn the entire experience of human experiments and over a century of robustly understood physical laws.

This is inaccurate. A gravity wheel won't violate the idea of the conservation of energy. Energy won't develop then expend in a linear fashion. It will exhibit cycles

Are you or are you not attempting to make a wheel that will rotate indefinitely in the face of friction without a conventional power source? I thought that's what you were trying to do.

 

[AgingYoung]

Nathan,

• I see that image, it appears to have a slope of 2 pixels across half your screen. If your screen is about 1000 pixels wide, that means it's slope is 4e-3 radians.

It might be that the apparent slope isn't what is actually in the model. I'd think the property box would indicate the actual slope or degree of rotation in radians. In that box 7.579 e-18 is indicated. Those are very large beams and the magnification is kind of up there.

• Are you or are you not attempting to make a wheel that will rotate indefinitely in the face of friction without a conventional power source? I thought that's what you were trying to do.

This is definitely what I'm attempting. It will prove that gravity isn't a conservative force as I understand it. I'm of the opinion that normal mechanics or mechanisms (made from simple machines, ie levers, pulleys, inclined planes, etc.) used in unusual manners (at times in ways they aren't normally thought of to be used) can model materials and ideas that although are commonly known become a very uncommon construct. By modeling the ideas I have it's my hope to prove that gravity isn't a conservative force.

I could be more specific but I'd prefer to make the model. Mr. Randi said that he'd give me a million bucks if I succeed. I wouildn't want someone as sharp as you to beat me to the punch.

Gene

 

[nathan]

Nathan,

• I see that image, it appears to have a slope of 2 pixels across half your screen. If your screen is about 1000 pixels wide, that means it's slope is 4e-3 radians.

It might be that the apparent slope isn't what is actually in the model. I'd think the property box would indicate the actual slope or degree of rotation in radians. In that box 7.579 e-18 is indicated. Those are very large beams and the magnification is kind of up there.

How does magnifying a beam make it appear differently sloped? Orientation is invariant wrt scalar magnification. Your hypothesis that the display is not accurate would invalidate your whole simulation approach, wouldn't it?

If that beam was originally horizontal and has been rotated by the amount you say, then there is a bug, for plainly 7.5-e-19 radians will not give that slope. It's quite obvious to prove that -- the rotation shows about 2pixels at a radius of 500 pixels, the circumference is going to be 500 * 2 * pi ~= 3141 pixels, so it'll take about 1500 repeats of that rotation to make a full circle. If it really was ~10^-19 radians, then one would expect of the order of 10^19 repetitions to complete a rotation. We're out by 10^16 orders of magnitude.

• Are you or are you not attempting to make a wheel that will rotate indefinitely in the face of friction without a conventional power source? I thought that's what you were trying to do.

This is definitely what I'm attempting.

Then you are trying to overturn eons of experiments and centuries of physical laws, all which show that gravity is a conservative force. Dress it up in all the fancy words you like, but that's what you're attempting. I therefore stand by my assertion that anything your simulation shows that disagrees with that is more than likely a fault in the simulation (either by approximation, or by bug, or by you operating it outside of its design constraints).

Only a physical manifestation of your wheel will convince me.

 

[nathan]

I could be more specific but I'd prefer to make the model. Mr. Randi said that he'd give me a million bucks if I succeed. I wouildn't want someone as sharp as you to beat me to the punch.

Thanks for the compliment. You've nothing to fear from me trying to beat you in making a spinning wheel -- I've got better things to do with my time

 

[Pup]

Are you or are you not attempting to make a wheel that will rotate indefinitely in the face of friction without a conventional power source? I thought that's what you were trying to do.


This is definitely what I'm attempting. It will prove that gravity isn't a conservative force as I understand it.

This is way out of my field, so forgive a dumb question. But you're designing a mechanical device that works by the force of gravity. Does that mean that it would work only if gravity were the force operating it?

For example, I'm picturing an old fashioned wind-up clock, that operates by gravity pulling the pendulum down. Put it in a non-gravity environment, and tug the weight down by any other means including just pulling on it, and put a magnetic attraction beneath the pendulum and any other parts necessary (can't think of any right now) and it would still work. Is that right, or am I already so far off base it's not worth continuing?

If that's correct, then are you saying there will be something special about your mechanical device, which will make it run indefinitely if gravity is the force acting on it, but if some other force is substituted, it'll obey Newton's laws and run down?

 

[AgingYoung]

Nathan,

At the far left the two beams are lined up yet at the far right where I showed them there is that difference between their tops (as seen by the black square on top and slightly right of center of the view box) . That's the actual slope. What an inexpensive low end monitor shows is another matter. I don't think the bug is in wm2d. The first beam is in the background and is horizontal.

Pup,

I've had several ideas that I've been pretty excited about. After I built them I could see that they weren't going to make a wheel turn. Most of the ideas I've had weren't as much an idea of how to cause gravity turn a shaft as they were a description of that wheel. The idea of increasing force by increasing the rate of acceleration is a description. What I'm presently considering is a means to actually cause that acceleration of the mass (or faster than gravity is moving it) and the means to cause the necessary deceleration and change of direction once the wheel is at 180° or pi rotation.

If you want to visualize what I imagine this might help. Think of a wheel where someone is sitting beside it at around 2 o'clock and someone is sitting beside the wheel at 5 o'clock. The person's job at 2 is to move the weight out a little and the person's job at 5 is to push it back in where it was. At 30 rpm with a 12 foot wheel those people have to move real fast but imagine they can. That wheel would turn.

An obvious question (since I'm not going to hire really fast people to sit beside the wheel) is where is that force coming from? I intend to mechanically simulate in reality a material that doesn't exist. That's one of the ideas in this model. There are other ideas that I plan on incorporating into this model.

I'm kind of excited about these ideas. You probably know that creativity isn't on tap. For me it's a slow process and I'm still considering the details. I'm practically certain if I can model these ideas that it will produce a wheel that turns. For these ideas to work gravity is essential.

Gene

 

[nathan]

Nathan,

At the far left the two beams are lined up yet at the far right where I showed them there is that difference between their tops (as seen by the black square on top and slightly right of center of the view box) . That's the actual slope. What an inexpensive low end monitor shows is another matter. I don't think the bug is in wm2d. The first beam is in the background and is horizontal.

Are you simulating the internal stresses and flexing of your beam under its own weight?

You have no evidence that my monitor is either inexpensive or low end (and you'd be wrong in both assumptions)

I intend to mechanically simulate in reality a material that doesn't exist.

Do you have any idea how stupid that sounds? And furthermore, if you do succeed in simulating such a material, what would you prove?

 

[Thing]

I do agree. I don't expect that a gravity wheel will be subject to Newtonian physics. There are two ways I can defend that statement. One is to explain the other ideas that I have that I haven't mentioned or the other is to build it. I'm in the habit of looking at an idea rather thoroughly before I share it so I think I'd better start building.

Well your simulations are going to (approximately) obey Newtonian physics so you can stop wasting time on them. The job of building something that isn't subject to Newton's laws is going to be plenty hard enough (i.e. impossible).

 

[russingram]

animated gif of Gene's wheel

this is the first time I've tried this, so I don't know if it will work. Here's an animation of Gene's wheel, based on his description from an earlier post.

 

[russingram]

cool, it worked!

 

[AgingYoung]

russingram,

That is cool! That's exactly what I expect!! There is the minor part of what is causing that motion not being shown. Thanks for not letting the cat out of the bag.

Gene