Design Status Update

[Silvertiger]

One more thing. Did you build the simulation to an appropriate size and weight? Wm2d can have problems with huge or extremely heavy objects. You have sparked my interest on this one. Let me guess: you are using a spring in place of the weight. The spring pulls your two frames together causing them to rotate.

I have six foot designs with over 80 pounds all the way down to 12 inch designs with 3 to 5 pounds. No springs are used. I have used them in some designs, but not as accumulators. Most times springs work against the design. The frames do not rotate - they are mounted to the floor. Just to clarify: the frames are the support structure to which the machine is mounted.

[Silvertiger]

Thanks Silver.
I am sure you are accounting for the weight of the mechanism itself and its position.

No accounting is really needed in this design, with exception to its starting position, since gravity is both the primer and the fuel. But, there's a lot of accounting I did anyway, earlier in the thread.

Is it balanced?

Hell no. I HAVE balanced it, like a rim on a car. It can have both at the same time, and in reality it needs the balance or it will rip itself (and whatever it's mounted to) to pieces, but in sim it doesn't matter.

[Silvertiger]

I haven't been able to get a hold of Mr Vibrating. If he reads this I was able to design a successful runner sim WITHOUT using interacting polygons...only joints this time which removes the possibility of overlap errors yielding false results, as was his primary concern.

[Silvertiger]

Video of a 50 Lb weight driving a 160 Lb piston: https://www.youtube.com/watch?v=lvpjvjMr0TY

Here's an update on my working lever (at least in sim). (Towards the end the time step couldn't keep up with integration at high accelerations so I had to stop the sim.) Just a quick rundown: it's an alternating impulse engine, similar in operation to what an AC generator does electrically. DC dissipates over distance and so does momentum and impulse. Yes, they transfer like current across the matter through which they propagate, like what can be observed in a Newton's Cradle, but they too diminish over time and distance. So, instead of having a power station every mile or so, they elected AC since the back and forth action yields the same results as a Newton's Cradle does: instantaneous transfer of current with minimal loss since electrons are being traded between a few atoms rather than along an entire line of thousands of miles of cable. This engine does the same thing mechanically, but with impulse instead of current. And, just like AC, it can be amplified through the manipulation of its frequency counterpart in mechanics, which I cannot state, but I will say that amplification is allowing the storage of a virtual mass, much like amplifying voltage, allowing for the continuous operation of the machine. As the energy is amplified, virtual mass is stored on the other end of the lever until it becomes "heavier" than the driving weight of 50 pounds.

Specifically, however, this simulation eradicates any possible "megabounce" errors that have come to be associated with WM2D as sort of an excuse to explain away why one might be getting OU out of a design. Does it work? Or does it not? So in this sim, all interacting polygons have been removed and thus it is solely being operated through the use of joints. There can't be collision issues if there are no polygons colliding, right? And the law of levers and laws of conservation remain uncontested, as they are definitely always in effect.

Edit: This was MrV's primary concern after looking it. I would really like to build a test model for this thing. But first I'm getting a physicist from work to review the design specifications and do a mathematical breakdown and hopefully some proofs to better harness the design's capabitlities. I showed him the actual device and he needed no convincing beyond that point which is encouraging. :)

[Silvertiger]

Alternating Impulse Engine Two-Weight Prototype: https://youtu.be/COUx1L4w90o

[Silvertiger]

Some good news. I have been trying for over a year (off and on) to get an inversion of the design to work in simulation to prove the motion and help deliver the mathematical framework. Today I finally got it to work. I successfully inverted the design and consequently inverted the dynamic and static functions of the fundamental components.

I have been trying to model the design based on parameters given by the First Postulate of Special Relativity in a specific congruity with the Second Law of Motion. This model may possibly be easier to build. If not, it at least allows me to test the design in the real world with little or no cost me. :)

[Silvertiger]

Here it is: Two-Weight Alternating Impulse Engine - Inverted Drive Prototype

The motion type should look to be slightly different from the motion in the last video. Like the others, the engine runs until WM2D runs out of memory. The fact that this inverted design works bodes very well for my confidence level lol.

[Silvertiger]

Ok, there is officially MUCH to be discussed, as I have a very important update involving the possibility of Jim_Mich's "Motion Wheel." Has he come back to to haunt us? It would appear that way. I just drew this current design in SimWise 4D and tested it with tremendous success. So it works in both WM2D and SimWise. I find myself forced to share more detail than I am comfortable with, but it became necessary as you will see in the video.

Here is a video of the wheel and flywheel in operation. This is the latest redesign of an engine I have been working on for a few years now, of which I now have three different types, this one being the most recent.

Thoughts are most welcome.

[cloud camper]

I've encountered anomalous results such as this in WM2D.

As soon as you place a load on the system, there is no anomaly.

Probably due to rounding errors that progressively compound.

Try the simulation with smaller and larger step sizes. It greatly affects the results!

[Silvertiger]

Ok. It works in Simwise 4d as well...a completely different platform. I have over twenty different variations of the main design and they all work. I have ones with polygons, and ones without polygons and only linkages. I have ones with springs, ones without springs, ones with loads, ones without loads, etc. They all work...on two different design programs. I don't know what else to say about it really lol, except a LOT of work was involved. I have run them down to insanely slow speeds, where I take the timestep down to 0.00005 and overlap to the same distance. They still work. I even did an inverted design, whereby the prime mover no longer moved, but the system was arranged in such a way that it was still moved BY it. Here is video of that, even though much is hidden, and only the lever is visible.

Let us lead with the assumption (even if you are not convinced...even I have my doubts) that it really does work and discuss it from there?

[eccentrically1]

It looks simple enough to build. I'm looking forward to seeing it. When do you anticipate construction to finish?

[Silvertiger]

I'm not sure. I just did an early cash-out on my retirement, so I finally have some spending money!

[Fletcher]

Hi Silvertiger .. that looks impressive. The ratchet build must have been a b.t.h. I do like the last sole lever model better as as I can see the reciprocating/oscillating action more clearly.

I'm sure jim_mich would approve, if it were his concept you have also discovered.

I'm sure that I have mentioned this before so if I have bear with me a while longer. Probably you should treat these items as things to check against.

1. Although you have air resistance on Low the reciprocating action (which maintains the same amplitude) should show an eventual slowdown of the period, after it reaches top speed. Generally small builds with low parts surface areas and not many parts hardly show any effect at all re: system energy losses like air friction so you need to exaggerate it to act as a Load.

Try increasing the radius of the red weights to much larger and/or increasing air frictions to HIGH. If it still does not slow down then there are other things to look into (this applies to any of your sim formats/programs I should think).

2. Increase the Accuracy because as things get faster there is a possibility for parts overlap jumping to each frame. Then things get weird as parts are expelled and sometimes apparent energy is given to the system. You don't want that confusing the results.

3. Don't start the system moving by the method you are using i.e. changing the x or y velocity value while retaining all other motions. Do something more realistic like pinning a small weight to the system (activation weight) that drops off after a short period of time having given motion to the system to be tested.

4. Most importantly - If you are using Spring Elements (lossless) then they are ok for quick and dirty testing of actions etc. If you get a result like yours then you need to change them out for Spring Dampener Elements which have losses akin to real springs built in and are much more realistic in behaviour.

Ok, that's all for now.

P.S. Reproducing effects in different programs is good and can give some degree of confidence. But basic overhauling of systems (change-outs etc) and simplification to most basic of essential elements is the best approach to identify an anomaly and give confidence that you have something extraordinary IMO.

[cloud camper]

Hey ST - really not trying to give you a hard time but long time users of sim programs have regularly produced anomalous results and the steps being proposed are just to try and eliminate known issues with the program so don't take it personal.

Basically we just have to reduce the variables that are known to create problems before assigning any credibility to the behavior.
And even then its only "virtual" credibility!

[Silvertiger]

Cloud Camper:

I completely agree. A sim is best to thought of as a time-saver for engineering on the low-ball end. However, verification across two design platforms makes it VERY promising.



Fletcher:

1. I have set the air resistance to high speed many times. No effect.

2. As a default, I set my accuracy to a time step of 0.005 seconds, and the integrator error to 0.005 inches, leaving virtually no room for error. For more complexities, I move the decimal place to left another zero to 0.0005. Still works, just slower because of the time step.

3. The original design is started by gravity alone. The motion wheel, because it IS balanced, must be given an initial push.

4. I have designs with and without springs...they all work.

5. I challenged claims of "megabounce" due to colliding/overlapping polygons, not only by setting my tolerance anywhere 5/1000 to 5/10000 of an inch AND reducing the time step, but I also removed the polygons completely and used only basic linkages like a rod and rope.

6. I also inverted the design itself to see if it would work. See my post three posts above yours for the info.

7. Applying the load makes the system resets come even faster, as each cycle now expends more energy.

8. I can't tell you just how few parts actually comprise the mechanism. Whatever Bessler's arrangement was no one will ever know, most likely...but I do know that this, just from what I've read in Bessler's papers, is far simpler.