Poss. Symmetry Break?

[MrVibrating]

You're both right i think, i've made a fallacious conclusion somewhere...

The principle i'm trying to apply is that we can generate more energy from a given torque by sinking it into a smaller mass.

But my mistake was thinking that this can be shared back with the wheel via a collision - which in reality will only exchange momentums, which remain equal and opposite, rather than energy, the difference in which will just make noise...? Or something..?

So what, instead, if we just throw the boosted mass straight upwards, to convert its RKE into GPE?



The fundamental issue here is whether a directional asymmetry of KE is useful or not.

A directional asymmetry of momentum exchange would clearly be useful, and fairly straightforward to harness.

But we don't have one, and there's no suggestion of one here.

A directional asymmetry of KE however is trivial. If unequal masses use each other as reaction mass then a given input of energy between them will be asymmetrically distributed - the momentums will be equal and opposite but the smaller mass will be faster and have more energy.

And an easy way to do this with inertial torques is to sink the positive inbound torque into a smaller orbiting mass, and the corresponding negative outbound torques into the net system. Voila, equal opposite angular momentums but unequal CW to CCW RKE.

But if that's not even useful in the first place then 'harnessing' it is kind of moot..

So the question is, can this asymmetry of RKE be converted to an up / down GPE asymmetry?

[daxwc]

One of the biggest perplexities I have is why the wheel didn’t vary its speed under varying of load. We see a different speed just under very heavy load (water screw) which probably means it was at the end of its power band.

[MrVibrating]

Yeah, same here. And i don't think my musings above can speak to that problem - a simple GPE advantage doesn't cut it, in terms of explaining these very unique performance characterstics.



The starting point here, i believe, is that the feat you describe is all the more impressive for the machine being a statorless peritrochium.

Another demonstration of its RPM constancy was when lowering a load - ie. under negative torque.

And yet another facet on this same characteristic is Bessler's assertion that it could gain yet further advantage by being coupled to conventional machines such as water pumps.

The sense i'm getting is that some internal apparatus is generating positive and negative inertial torques, but that this mechanism is somehow partially floating with respect to the main wheel, in such a way that the torque asymmetry is dynamically variable.

Under over-speed or direction-reversal conditions, the torque asymmetry tips the other way.

If it lowers a load at the same constant rate it raises it, then the degree of asymmetry is fully reversible - generating the same excess of positive or negative torque depending on the main wheel's resistance to accelerations in its preferred direction.

So this has to be useful info...



There must be some arrangement wherein the balance of +/- inertial torques applied to the axle is a function of its resistance to accelerations - ie. MoI - since, from inside a statorless wheel, what is an applied load but a resistance to acceleration of the housing of which the axle is part?

Hence we can deduce that in the first moments of startup, an unloaded wheel is accelerating in that particular direction, at least in part, because of its resistance to accelerations in that direction.

So push-starting a bi-directional wheel must initiate some kind of divergence of directional MoI's between the inner and outer parts, or something...

Should probably just stick to the one-way wheels in trying to suss this tho..

A load-matching one-way wheel...

Statorless...

Under static torque per OB, yet by all powers of reckoning, must've been employing inertial torques.

The amount of excess torque (the degree of asymmetry) sunk into a given acceleration is a function if its own resistance to acceleration.

Therefore, less inertial torque is applied when a mass is falling, but more is applied when the mass is relifted.

Likewise, then, we could surmise that if we overspeed the wheel, the inerital torque asymmetry backs off, possibly reversing if we turn it even faster.

But when we brake it, the inertial torque asymmetry increases, and presumably peaks when the wheel is forcibly halted.

We also know that the vertical requirement implies that the inertial torque asymmetry is gravity dependent.

So, it's a floating (dynamically variable) asymmetry of inertial torques, dependent upon the wheel's resistance to angular accelerations in one direction, and gravity.

So the scrapheap challenge is, how can we make something that does that, with the available junk?

Is it an asymmetry in the balance of +/- inertial torques actually generated, or rather, how equal opposite torques are distributed?

When the one-way wheels are under static torque, is the OB mass a discrete weight, as in a system of peripheral rising and falling masses separate from the prime mover... or else somehow part of the prime mover apparatus itself? Because this static torque kinda matches the profile of 'resistance to angular acceleration'... IOW, the internal apparatus sees the wheel's MoI as effectively infinite when it's tied off stationary, so is applying maximal torque. Once it starts rotating, less torque is applied as its net momentum rises, lowering its resistance to further acceleration against something moving at constant (ie. RPM invariant) speed in relation to the main system.

This in turn however would imply that the MoI-induced torques are never applied directly into the main wheel, but rather via a system of gravitating weights which in turn interact with the wheel..

My head hurts.

It's a jigsaw. The parts are all here, it's solvable... just gotta piece it together..

[ME]

I'm not quite sure if you covered it with this "We also know that the vertical requirement implies that the inertial torque asymmetry is gravity dependent.", but I think an overbalanced mechanism under gravity will just keep accelerating (or decelerate) until it reaches the maximum velocity (and centrifugal) where it doesn't change the MoI:
Basically, as a starting point, I expect something as basic as: a[centrifugal]=g --> v[max]=Sqrt(g*r) or w[max]=Sqrt(g/r), right?

[Trevor Lyn Whatford]

Hi ME,

Basically, as a starting point, I expect something as basic as: a[centrifugal]=g --> v[max]=Sqrt(g*r) or w[max]=Sqrt(g/r), right?

I would say that, in a out of balance wheel Centrifugal Force (CF) can never be = to the force of gravity, because in that scenario gravity is creating the CF (gravity is the dominant force), all CF can do is effect the force of gravity acting on the weights by adding additional force vectors to the weights, thus can only govern the wheels top speed.

I do not see there being a formula to fit all scenarios, you have to know the mechanism to work out all the force vectors, and they are all variables to sizes, numbers, speeds, designs, to name just a few, and that is before you start looking at CF effects on torque forces if your design is using leverage.

Rule number one (the old meaning), first get something that works.

I look forward to the day when CF become a problem to one of my builds.

[ME]

I do not see there being a formula to fit all scenarios, you have to know the mechanism to work out all the force vectors

True, I used the basic formula for a mass on a string swung in a vertical plane. For a circular motion it has the biggest issue at the top, that formula applies there: at such speed there's is no way some weight is able to positively affect the wheel.
I suspect a mechanism of any kind is not able to beat this maximum required velocity; so at least there is a hypothetical limit which is better than knowing nothing, unless I made a very wrong assumption here...

[Trevor Lyn Whatford]

Hi ME,

one of the ways to use CF, is to time the falling levers to aline the gravity force vectors with the CF vector to increase the leverage torque of the levers.

My current build is using similar to this, to increase its efficiency, by turning some of the negative forces of CF back into positive forces. Ever little helps.

I also think a gravity wheel can never reach Maximum velocity, but 26 RPM would be good enough for me. Depending on the design, I would think that faster wheel speeds might just be a trade off from wheel torque to wheel speed, so there would be a point were there is little or no advantage looking for extra speed.

[ME]

At the descending side some lever approaches weightlessness, while on the ascending side it approaches 2*G: so that's almost the same limit.
When a vertical lever falls to a horizontal position (I think in the most optimistic situation) then via E[pe]=E[ke] --> m*g*r=0.5*m*v² --> v[max]=Sqrt(2*g*r)
When the wheel moves faster than such lever can fall it will actually moves backwards instead of causing an impact.
I'm not yet sure why I got that other factor (2*), but at least things seem to indicate a relation between the RPM's and the square-root of the size.
I hope you can verify it soon with your design.

[MrVibrating]

I'm not quite sure if you covered it with this "We also know that the vertical requirement implies that the inertial torque asymmetry is gravity dependent.", but I think an overbalanced mechanism under gravity will just keep accelerating (or decelerate) until it reaches the maximum velocity (and centrifugal) where it doesn't change the MoI:
Basically, as a starting point, I expect something as basic as: a[centrifugal]=g --> v[max]=Sqrt(g*r) or w[max]=Sqrt(g/r), right?

Yes, if an inertial output requires a gravitational input, or a gravitational output requires an inertial or gravitational input, then the peak operating velocity of the gravitating part of the cycle is limited by the relative load balancing between input and output.

In the example below, the longer weight levers have a minor advantage over the smaller jack-mounted masses, at roughly 4:1 leverage, with some additional assistance from CF (the system wants to maximise its MoI):




(gake and fay anim for illustration purposes only)

These diametric weight levers appear to be good for fairly consistent activation angles across a wide RPM range, however when driving a significant load this performance is reduced - here, the mechanism only maintains consistent vertical orientation up to around 30 RPM.

At around 55 RPM the mechanism is still activating, but in a resonant mode with no consistent vertical sync.

At around 60 RPM the system hits a narrow band of stability wherein the predominant force acting on the large levers is CF, which wants to push them apart, one up, the other down, so they move outwards symmetrically, the smaller masses are drawn in symetrically, and the system perches into a minimum MoI state, with both pairs of masses at equidistant radii.

Beyond that it locks into a max-MoI state, one half of the mechanism stuck in, the other out, with no more internal motion regardless of rising RPM.



So we can draw certain conclusion from these dynamics with respect to what must've been occuring inside Bessler's wheels.

The most striking characteristic, aside from OU, is their astonishing acceleration.

I'm fairly certain that if any of us saw a ~ 300 KG 3 meter-wide wheel accelerate up to just shy of 60 RPM within a couple of revolutions we'd take a good few steps back, if not clear the room.

That startup performance alone is simply awesome. That is a lot of energy being gobbled up, very, very quickly.

Extrapolating these characterstics back to the mockup above, we have conflicting requirements - the long weight levers need massive advantage over the smaller jack-mounted masses, to the extent that the latter have almost no significant impact on weight balance or inertial momentum.

So this practicality absolutely precludes any possibilty of the approach ever causing useful OB. If the large masses can raise the smaller ones at all, then peak RPM is going to be under 10 RPM, and it'd take dozens of cycles at paltry acceleration to get there.

But the same limitation applies no matter what form of GPE input mechanism we use, and no matter what form of load is applied.

So for instance if we want a GPE input to retract some mass, causing an RKE output from inertial acceleration, then CF causes the same loading constraint - in other words, in order for the wheel to accelerate straight up to high speed, the work involved in inducing the inertial torque has to be almost negligible compared to the GPE input workload.


So in summary, simply lowering a drop-weight in order to pull some mass inwards - trading GPE directly for MoI reduction - is a dead end. Not. Going. To. Work. Ever.

Something else is required, and something pretty significant, at that..

Multi-staged inertial torque feedback loops or something, some form of super-charging, or else a wholly-free output that presents as no load at all upon the input, and also happens near-instantly...

Tall order eh..

[AB Hammer]

MrVibrating

Nice animation, of what we would like to see a wheel do. What program did you do this animation on. I have several designs I would like to do the same line drawing gif. I just need to know what to study during my downtime from my sinus operation in Aug.

[MrVibrating]

I just used WM2D, and ScreenToGif for capture.

FWIW the anim is driven at constant RPM by a hidden motor, and unpowered, generates very little OB torque, keeling within 180°.

[daxwc]

MrVibrating:

The most striking characteristic, aside from OU, is their astonishing acceleration. I'm fairly certain that if any of us saw a ~ 300 KG 3 meter-wide wheel accelerate up to just shy of 60 RPM within a couple of revolutions we'd take a good few steps back, if not clear the room.
That startup performance alone is simply awesome. That is a lot of energy being gobbled up, very, very quickly.
Multi-staged inertial torque feedback loops or something, some form of super-charging, or else a wholly-free output that presents as no load at all upon the input, and also happens near-instantly...

Have another look at MT 48 if none of the weights leave the quarter then fast acceleration will occur.
“ MT 48 The principle is good, but the figure as it is will not give birth to any motion until completely different structures bless this marriage.�

[MrVibrating]

Acceleration is excess torque divided by net angular inertia (F=MA inverted & rotating), in other words a function of whatever degree of lift advantage afforded by the linear conveyor.

His annotation acknowledges its utility as a closed GPE cycle - as a float of PE to KE conversion to play with - but the thing that plays with it, and the game to be played, is the prime mover, and its freely-varying MoI, that takes a lumbering input momentum, reshuffles its distribution of inertia to velocity and spits it back out leaner, swifter and at higher energy.

The other thing you notice in MT 48 is that the weight-balls never traverse the wheel radius while in contact with it - whether incidental or not to this particular model, the apparatus shown serves to isolate the wheel from any inertial torques associated with changing radius.

As such, the implication is that whatever the nature of these additional "structures", their function is capable of generating a linear to angular energy asymmetry, ie. a path dependent interaction.

[MrVibrating]

At the descending side some lever approaches weightlessness, while on the ascending side it approaches 2*G: so that's almost the same limit.
When a vertical lever falls to a horizontal position (I think in the most optimistic situation) then via E[pe]=E[ke] --> m*g*r=0.5*m*v² --> v[max]=Sqrt(2*g*r)
When the wheel moves faster than such lever can fall it will actually moves backwards instead of causing an impact.
I'm not yet sure why I got that other factor (2*), but at least things seem to indicate a relation between the RPM's and the square-root of the size.
I hope you can verify it soon with your design.

Was gonna give this reply earlier but deleted it before posting, as these kinds of nebulous fundamentals tend to cause attentions to glaze over..

But between you and me, i think the potential source and sink is implicit in the alternate scaling factors of time-invariant gravitational energies and speed (and thus time) dependent inertial energies, and some kind of get-out clause or snakes'n'ladders game to flit spontaneously between min and max energy states in alternate fields.

Trace a vertical line from axle to rim - is energy changing linearly per GMH or exponentially as a function of MoI? Could be one or other or some combo. And while inertial and gravitational interactions are symmetrical in their own right, when one is a function of the other we have analogues of hysteresis, remanance and coercivity coming into effect. Add in multiple axes and inertial torque induction and any preconceptions i may've had that mechanical OU had to be simpler than EM have been well and truly laid to rest..

But we've got a time-dependent energy field, and a time-invariant one. Noether's theorem says CoE won't apply if we can find an interaction using these as source and sink. Inputs in one field, outputs in another, different scaling dimensions yielding diverging energies as a function of changing velocities. Momentum in equals momentum out, but we'd be constantly using the energy gain to buy more low-value momentum, buffing it up into the high value premium stuff.

So... theoretically, we have the ladders. Tall ladders, and short ones. We just need some snakes. Shortcuts between energy peaks in alternate fields, or counter-balancing, or some other form of decoupling, that has to be as simple, and swift, as falling off a log.. somehow.

[Fletcher]

Have another look at MT 48, if none of the weights leave the quarter then fast acceleration will occur.

“ MT 48 The principle is good, but the figure as it is will not give birth to any motion until completely different structures bless this marriage.�

Acceleration is excess torque divided by net angular inertia (F=MA inverted & rotating), in other words a function of whatever degree of lift advantage afforded by the linear conveyor.

His annotation acknowledges its utility as a closed GPE cycle - as a float of PE to KE conversion to play with - but the thing that plays with it, and the game to be played, is the prime mover, and its freely-varying MoI, that takes a lumbering input momentum, reshuffles its distribution of inertia to velocity and spits it back out leaner, swifter and at higher energy.

The other thing you notice in MT 48 is that the weight-balls never traverse the wheel radius while in contact with it - whether incidental or not to this particular model, the apparatus shown serves to isolate the wheel from any inertial torques associated with changing radius.

As such, the implication is that whatever the nature of these additional "structures", their function is capable of generating a linear to angular energy asymmetry, ie. a path dependent interaction.

I read your thread with interest MrV.

And I hold firm to the belief that there is a primary system (the Prime Mover mech(s)) that acts pretty much as you say.

And that there is an abundance of options for the secondary OB mechanisms of which MT48 is but one (and glaringly not a 'normal' OB subsystem).

What I can't decide is if there is ONE Prime Mover mech or many Prime Mover mechs e.g. an 8 division wheel with 8 OB subsystems has 1 Prime Mover acting 8 times per revolution or 8 Prime Movers acting once per revolution ??

Keep up the great work. Hit a nut enough times and it will eventually crack.