Poss. Symmetry Break?

[MrVibrating]

"1 lb can cause the raising of more than 1 lb. Wagner says that, to date, no mechanism has been found fit to the task.. He's right, and so am I.. can anyone see why?"


We've been grinding away at the wrong end of the problem - what we've all been trying to do is engineer an energy gain; to design some sequence of actions that raises a weight on the cheap.

And what we've all been finding is that such a mechanical arrangement is intrinsically impossible. Nothing we do ever has or will produce a gain in energy. All else being equal, there is no mechanical possibility of 1 lb causing the raising of more than 1 lb.

As noted in the past, the only logical solution to Bessler's little poser here is that the means in question is, in fact, non-mechanical. It's not something designed or manufatured by mortal hands, nor can it ever be.

The 'mechanism' is no artifice, but nature Herself. 1 lb can cause the raising of more than 1 lb - energy can be created and destroyed - by an effect so trivial, a child can generate it, unknowingly.



In every other approach to this problem, we've sought to generate an energy asymmetry. How to actually harvest a gain that's already there is not something we usually come up against (since we never get that far). Here, though, the magic part is already done for us.. conservation of KE depends upon mass constancy, or, in rotating systems, the invariance of angular inertia.. an inherently variable quantity. So the issue here is not how to create the advantage - the form of the excess impetus is already writ large - all that remains is some means to harvest, or rectify it.

There's gold in them thar hills...

[MrVibrating]

What grabs me about MT 143 (the Roberval) is that it has counter-rotating MoI's with a separate axis to the system's main center of rotation... necessary requirements by my deductions.

I've been playing with such weighted armatures already in simulation - swinging inwards from various radial distances and directions, looking for interesting offsets between opposing centrifugal forces, but counter-rotating pairs isn't something i've tried yet.. seems a good place to look for an asymmetry..

[eccentrically1]

I just love the maths of this tho... P=MV, KE=1/2MV^2... so if P=100 kg/ms and M and V are both 10, KE = 500 J. But if the inertia is halved down to 5 kg, velocity must rise to 20 m/s to conserve P, and energy doubles to 1 kJ... the energy gained from the rise in velocity is not offset by the reduction in mass/ inertia, so although we have half the effective mass in flight, at double the speed it has twice the energy.

No offense intended, but doesn't the math for angular momentum, moment of inertia, and rotational kinetic energy contain terms for radius?

L = mrv

I = mrr

𝑲𝑬rot = 1/2 mrrvv

So mass and velocity are 10 initially and if the radius begins at 1, and is reduced to .5,

L = 10 x .5 x 20 = 100 (constant mass, new radius, velocity doubles to conserve momentum)

I = 10 x .5 x .5 = 2.5 (new moment of inertia, mass is easier to accelerate at new radius)

𝑲𝑬rot = 1/2 x 10 x .5 x .5 x 400 = 400 x 1.25 = 500 (KE conserved at new moment)

Unless I misunderstand, which has happened more times than I care to admit.

Okay, first let's run thru the linear scenario:

We begin with a 10 kg mass at 10 m/s, so momentum is 100 kg/m/s.

Then, because abracadabra, the mass halves down to 5 kg. CoM does its thang and V doubles.

Momentum is still 100 kg/m/s, but now divided between 5 kg and 20 m/s.

The initial energy was 10 kg at 10 m/s = 500 J.

The final energy is 5 kg at 20 m/s = 1 kJ.

So energy doubled because momentum was conserved.

Replacing mass with MoI just dispenses with the magic:

MoI = mass times radius squared.

So suppose we have 1 kg at 10 m radius, MoI = 100 kg/m/s.

If we halve the radius, MoI drops fourfold, down to 25 kg/m/s (because MoI doesn't simply sum but squares with radius).

So to halve the MoI by varying radius, we only need reduce it by just under 3 m; 1 kg @ 7.07106 m radius has 49.999 Kg/m/s of angular inertia.

Alternatively, it'd make the maths neater if we just halved our mass instead, since 0.5 kg at 10 m radius also has 50 kg/m/s of angular inertia. But rotation doesn't let us vary mass, anymoreso than linear accelerations. So all we can vary is radius...


Sooo, if radius drops by a factor of two, MoI is reduced, and velocity thus increased, both by a factor of 4 respectively (again, conserving net momentum).. then RKE has quadrupled.


It is most important that we recognise the special nature of this relationship; it is commonly understood that KE is "not conserved", but this is usually stated in the context of dissipative loss mechanisms, in which KE is simply transferred to ambient heat. Here, however, if the radially-moving masses are allowed to bounce in and out with perfect elasticity (ie. no frictional losses) then we observe the net system energy fluctuating, in this case, by 400%. This variable net energy is clearly not being radiated and re-adsorbed as it comes and goes (which in itself would challenge 2LoT); as a "loss", it is non-dissipative, and likewise, as a gain, it is ex-nihilo.

Precisely the kind of asymmetry we're looking for - non-dissipative loss in one direction, and non-2LoT violating in the gain direction. IOW a non-dissipative non-conservative system.

I have the growing hunch that MT 41 is all about MoI, rather than balance, and that the 'upper' amd 'lower' masses are more meaningfully regarded in terms of being inner and outer, ie. varying the MoI.

If this intrerpretation is valid, then it is interesting that the outer two masses are inducing cancelling forces via their interacting jacks and horizontal drive wheels, while the corresponding counter forces from the inner two are non-cancelling (jacks open)... IE. the 'counter-force' caused by extending our masses is a negative torque on the wheel. So, if these counterforces mutually self-cancel, but the ones from the inner weights accelerating the system, do not, then we have gain.


So with this in mind i spent some time today trying to find ways to cancel the negative torque caused by a radially-extending mass. No conclusive results yet but it's an angle. There's others yet to try. One way or another, Occam's razor suggests, to me at least, that Bessler must've found some way of freely or cheaply varying MoI...

I wouldn't agree Bessler found a way of freely or cheaply varying MoI.

If the radius drops, then there has been an external input of RKE to the mass. It explains the quadrupling magic; it hasn't been dispensed; the mass doesn't move to a smaller radius magically by itself. Its inertia is tangent to the curve, helpless to our quest. No one can comment any further than that on your generalized scenario, you've simply plugged numbers in to the equations for perfect conditions; you'd have to be a little more specific.

It would be the same for your linear scenario, because abracadabra, the other half of the mass went somewhere, or in a perfect collision, got left behind at rest.
An example would be dropping two balls, one on top of the other, and the lowest ball sticks to the ground and the top ball rebounds away at some velocity greater than impact velocity. A tennis ball and basketball sort of work, if you can get them to drop just right, but the lowest ball still has some motion unless you could use velcro or super glue to stop it. The velocity depends on the balls' masses and coefficients of restitution, also known as Newton's experimental law.

e = relative speed after collision/ relative speed before collision

https://en.wikipedia.org/wiki/Coefficie ... d_momentum

In a one-dimensional collision, the two key principles are: conservation of energy (conservation of kinetic energy if the collision is perfectly elastic) and conservation of (linear) momentum. A third equation can be derived [7] from these two, which is the restitution equation as stated above. When solving problems, any two of the three equations can be used. The advantage of using the restitution equation is that it sometimes provides a more convenient way to approach the problem.

Provided the speeds are measured accurately, obviously.



Stacked ball drop

https://www.youtube.com/watch?v=2UHS883_P60

As always, the motion of the ground is negligible.

[Fletcher]

What grabs me about MT 143 (the Roberval) is that it has counter-rotating MoI's with a separate axis to the system's main center of rotation... necessary requirements by my deductions.

I've been playing with such weighted armatures already in simulation - swinging inwards from various radial distances and directions, looking for interesting offsets between opposing centrifugal forces, but counter-rotating pairs isn't something i've tried yet.. seems a good place to look for an asymmetry..

In the Roberval the levers can be moved laterally at low energy cost i.e. to give the levers KE of movement some part of a system must lose GPE.

The levers can move in the same direction which shifts CoM (though it can't be used for torque to turn the system because it is anchored). Or they can be moved in opposite directions for example.

The problem being that as soon as they are mounted in a free to turn wheel they will seek their position of least GPE and then instead of having say one lever lift and the other lose height (where vertical CoM remains the same) they then both need to be raised etc, which requires Work.

So it seems very problematic mechanically wise to use this approach in a wheel.

[MrVibrating]

@eccentrically1

Leaving mass behind isn't an option - if the linear scenario were to work, the actual value of mass itself (the strength of the Higg's interaction) would need to drop, in order for CoM to cause a velocity rise.

And obvioushly, mass is constant and inertia is mediated at lightspeed (instantaneously), so 'abracadabra' is just a way to illustrate what would be on the table, if mass constancy wasn't.

But in a rotating system, 'mass' is replaced in our momentum term by angular inertia, MoI, which is variable, and without magic.

As already noted, i would expect that varying MoI usually costs energy equal to the RKE gained. Therefore the hypothesis here is simply that if this was the form of gain Bessler was harnessing, then there is some means of circumventing or economising that usual cost... More fundamentally, for a time-varying field to perform free work upon an invariant one, the energy cost of that variation has to be thermodynamically decoupled from the resulting net energy change... and we're not quite there yet, all we have is the implication (mounting, IMHO) that such an exploit is possible, and awaiting rediscovery.

Again, the "quadrupling magic" is simply nature herself - P=MV and KE=1/2MV^2, so the KE corresponding to a given P depends on its specific distribution of M and V. If that distribution changes in time, then so does the corresponding energy. This energy change "just is" - it's written into the terms of conservation. It does not have to be supplied by an external source, but simply comes into existence from nothing, as far as classical mechanics is concerned. The "inertia tangential to the curve" is CF, but the inertia of interest here is the angular inertia - and the corresponding balance of momentum and energy constrained to and following that curve.

In your example of coefficient of restitution, the 'gain' has a clear source - energy is being input from GPE or some other source of KE, and more input work is being done. But a thermodynamic gain requires a passively-time varying force (or at least partially-subsidised). As noted earlier, not really the same animal as a non-dissipative non-conservative system, which is what the CoM / RKE relationship boils down to.

As ever in such a situation, it can be just as informative to consider the loss scenario, as the gain - when a pirouetting skater extends her arms, slowing her rate of spin, she loses energy equal to half the square of the velocity change... so where has that energy gone? We can account for all the momentum - that's still conserved - but because KE and momentum have these different dimensions (the latter having an additional time derivative), and hence precisely because her net momentum is conserved, her net KE is not.

This energy has not been radiated as heat - it would still work in a vacuum, provided the skater had adequate life support, precluding mechanical osmosis... so then the skater would have to glow red hot to dissipate the energy by EM instead..

And so likewise, when she pulls her arms in again and re-accelerates, the energy gained has not been adsorbed from the environment. It is just an incontrovertible fact that the net energy of the system is not constant. Simply put, conservation of energy does not and cannot be applied to time-variant interactions, and read between the lines this is the real lesson of Noether's theorem..


I don't expect all of this is gonna make sense to everyone at first, and most likely it'll go nowhere and peter out like every other idea. Incredulity in the absence of understanding is the only sane response. But if you just consider that momentum (P) is conserved, and that P=MV, then M and V are covariant and if one changes then the other must compensate to conserve net P. But while this covariance is linear, the energy corresponding to a given distribution of M and V squares with V (KE = 1/2 MV^2).. so, like i say, the gain in energy when V rises to compensate a drop in M (inertia), just is. It hasn't come from any classical source, but has simply been conjured from a reshuffling of the conserved momentum. IOW the 'magic' is already there in nature... Sure, i'm plugging in nice round numbers, but only to keep things simple - use any figures you want. If P's conserved and moment of inertia's variable, then velocity rises because the given momentum is conserved through a smaller displacement, and less angular acceleration = less inertia and hence higher velocity. P sums linearly, but KE rises exponentially with velocity, so the gain is caused by nature's maths, not selective input values (try any values you wish).

And to be perfectly modest, the "maths" here are just the most fundamental axioms of classical mechanics - the terms and inter-dependence of momentum and energy.


Hence my point this evening, that we've been attempting alchemy when we should've been trying geology..

At issue here is the revival of the "vis viva" debate of antiquity - Newton, Leibniz and S'Gravesande, the latter two both personal aquaintances of Bessler, and witnesses to his wheels. So Bessler was most likely acutely focused on the vis viva debate, as much for fraternising with the very epicenter of its protagonists, as for its obvious implications to his ordained mission..

I've previously noted the possibility that his prediliction for the number five could be related to certain solutions to inertial asymmetries producing five-fold (or 50x, or 500x etc.) energy gains (discussed in other threads). I think there's a very good possibility that Bessler had not only already resolved the vis viva debate for himself, but gone that extra step and noticed the implicit asymmetry in their scaling dimensions.

But enough waffle for now.. the solution to the vis viva debate for rotating systems is clearly that CoE is subordinate to, and dependent upon CoM.

[ME]

In the Roberval the levers can be moved laterally at low energy cost i.e. to give the levers KE of movement some part of a system must lose GPE.

The levers can move in the same direction which shifts CoM (though it can't be used for torque to turn the system because it is anchored). Or they can be moved in opposite directions for example.

The problem being that as soon as they are mounted in a free to turn wheel they will seek their position of least GPE and then instead of having say one lever lift and the other lose height (where vertical CoM remains the same) they then both need to be raised etc, which requires Work.

So it seems very problematic mechanically wise to use this approach in a wheel.

For the Roberval it's CoM can be shifted horizontally because of it's connection to the ground (which creates this horizon). Removing that anchor, will remove the PE eventually - so that's indeed not helping.
For a rotating wheel this ground would be the rim and upwards would mean towards the axle.

If the geared weights of MT143 can be scissored relatively easy, then it should be as easy to put one weight to the rim and the other towards the axle and thus change the MoI *).
-According to WM2D this just results in a (familiar) wobble around a mean velocity.
*) I'm in doubt now: as MT143 clearly shows the Com will only shift horizontally, so for a wheel that's tangent and not radial....

But enough waffle for now.. the solution to the vis viva debate for rotating systems is clearly that CoE is subordinate to, and dependent upon CoM.

I simply conclude Energy is basically (on itself) a useless number when it's not converted to momentum or potential.

[MrVibrating]

What grabs me about MT 143 (the Roberval) is that it has counter-rotating MoI's with a separate axis to the system's main center of rotation... necessary requirements by my deductions.

I've been playing with such weighted armatures already in simulation - swinging inwards from various radial distances and directions, looking for interesting offsets between opposing centrifugal forces, but counter-rotating pairs isn't something i've tried yet.. seems a good place to look for an asymmetry..

In the Roberval the levers can be moved laterally at low energy cost i.e. to give the levers KE of movement some part of a system must lose GPE.

The levers can move in the same direction which shifts CoM (though it can't be used for torque to turn the system because it is anchored). Or they can be moved in opposite directions for example.

The problem being that as soon as they are mounted in a free to turn wheel they will seek their position of least GPE and then instead of having say one lever lift and the other lose height (where vertical CoM remains the same) they then both need to be raised etc, which requires Work.

So it seems very problematic mechanically wise to use this approach in a wheel.

At this stage gravity is an unnecessary complication - my sole concern is getting the masses back out again without incurring the full compliment of negative torque - IE. i want the skater to re-extend her arms, but without slowing her rate of spin, or at least without the full normal deceleration.

This is one reason paired inner / outer masses might be useful - as noted already, this changes the MoI 'hill' into a valley... normally, MoI = mass times radius squared, so for a given mass it is minimal close to the axis and maximal out at the rim. But a rotating system of opponently-paired inner / outer masses has minimum MoI when the masses are halfway between inner and outer extents...

In the skater analogy, she'd be extending and retracting alternate arms, and would reach peak speed and energy during the moments when both were at equal radius.

I haven't yet noticed any specific benefits to this but it changes the lay of the land enough to be worth bearing in mind..

So yep, activating the levers will also exert counter torques from N3 reactions, but, like gravity's influences, i'm trying to ignore these as incidental...



The rabbit hole my head's currently stuck in looks something like this - i'm breaking everything down into elements; MoI squares with radius because angular acceleration does, so if we want to extend our radius while maintaining angular velocity, we need to either add velocity.. or else reduce effective distance, or the net acceleration across an effective distance, or etc. etc.

From the external frame, the angular inertia of the net system modulates the angular inertias of rotational axes orbiting the common center, as thier net trajectories become epicycles, and so will integrate differently depending on their directions relative to that of the net system (CW vs CCW).... as such, the net angular accelerations from the external frame could be asymmetric - although the physical distance between the center and rim is invariant, the actual distances travelled per net angle of rotation by the counter-rotating orbiting masses in transiting between inner and outer positions, may be variable.


If that makes any sense. Possibly a reference-frame thing, it's still a bit foggy.. But as you've noted, if the prime mover's inertial then gravity's application is merely an interior design issue. So for now, the orientation of the levers WRT gravity is not a factor - all i'm thinking about is how to either retract MoI on the cheap, or extend it without incurring the usual drop in RPM.

I think re-embiggening MoI on the sly looks like the tastier option - we know we want to reduce MoI to boost the energy, so the challenge is simply keeping posession of it when MoI has to increase again - ie. getting the masses back 'out' there (whichever direction that may be), to the peak MoI position.

As you've already acknowledged, the trick then would simply be a matter of raising weights when MoI is low, and lowering them when it's high.

It would seem ideal if it transpires that the inertial and gravitating masses can be one and the same, but this would just be a convenience... my attention for now is focused on the cost / benefit constraints of the MoI / RKE relationship.

Linear radial translations seem to be of limited versatility in this respect, but angular ones - especially equal opposite ones in a rotating frame - open the door to a more diverse and complex range of inertial integrals.. hopefully, to be demonstrated shortly...

[MrVibrating]

I simply conclude Energy is basically (on itself) a useless number when it's not converted to momentum or potential.

"Converted"? It needs no conversion since it is already manifest as RKE!?

It's not some notional energy rise, as if only on paper... we're not simply swapping price tags, the energy corresponds to - and only has meaning in the context of - a real rise in velocity. Of real mass.

Again, go back to the pirouetting skater - when she extends both arms she decelerates, and sheds energy equal to half the square of her change in RPM... her initial energy and final energy are both 'real' and perfectly valid measures, but since they correspond to different dependent velocities the net system energy is not constant. It really is that simple.

If we input 10 J when MoI is high, and MoI then halves, we now have 20 J. Not a cashier's cheque, not bearer bonds or casino chips, but cold, hard RKE, 100% pure and unadulterated. The skater really does change angular velocity with MoI, and RKE really does follow half the square of velocity. Not just disembodied abstract figures of expression but literal, corporeal quantities.


Again, if P = I * V and is conserved, then V must rise when I falls, and since KE evolves as half the square of V, if MoI is time-variant then so is net RKE... it's just an unassailable consequence of the vis viva resolution; the nature of the relationship between CoM and CoE.

[ME]

Absolutely true.
For a single reference, and within one system.

But I do think it's indeed similar to currency, and has no problems when things are interchangeable; while potentially skewing things when comparing (possible) unrelated/relative things as being absolute.
It's just a hunch (a feeling, or <I don't know>, something indefensible, perhaps even invalid or simple paranoia :-) that the term "energy" is sometimes used to point towards things which are unconnected. Where a sudden raise in energy (esp. in "distance-squared"-situations) is an explanation of something that might work while what's implied could be unrelated, while (as you almost point out) it would be more realistic to convert things back to RPM's, F's, a's, V's, dx's, etc.. - things which can be measured and tested - A minor obstacle, and shouldn't be hard to do.

Ah well - don't mind me too much, I guess everyone has his own rabbit hole :-)

[KAS]

After reading all the evidence and witness statements, it is my humble opinion that the 2 pendulums were in fact energy storage devices. It is possible that the enemy of Bessler's wheels were centrifugal force. Especially if the weights fell inwards on the up cycle. A means of slowing the rotation may have been sought whilst preserving the gain energy in the up swing to be released at energy flat spots in the cycle. I think he got around this in his later wheels by incorporating a fly wheel which achieves the same effect. This to me is the only logical explanation.

[MrVibrating]

...for any of those in still in doubt about the veracity of the gain, its form here is exactly the same one Fletcher identified, which would arise if momentum could be tranferred between unequal masses (essentially, inertias).

The only difference here is that the two inertias the momentum is transferred between are one and the same body, just at different times.

Anyhoos, tried a quick sim this eve of an MT 143-inspired concept mentioned yesterday - using the counter-rotating paired weight levers.

First attempts allowed everything to move freely under gravitation, just to see what would happen... the upper and lower weights were geared together, some PE was given and i sat back to watch the chaos unfold..

Energy wasn't constant, with small gains above the starting PE, probably due to error but with a chaotic mechanism it's hard to pin down sequences of causation...

So, just for the sake of a clarity, i've added a stator wheel, against which the main wheel and armatures can be synchronised - so a given anglular increment of the wheel causes a controlled degree of rotation of the arms..

This allows me to configure an idealised illustration of the concept i mentioned yesterday... again, the stator wheel here is a temporary compromise, just to be able to examine the relative inertias and momentums being induced. If a clear gain trajectory does become apparent, presumably synchronisation can likewise be configured by tuning the harmonics (lever lengths / masses etc.)..

So in the following sim, gravity is disabled, and the system is given a 10 RPM initial velocity, then allowed to coast without losses. Upper and lower masses are geared 1:1, turning equal rates CW and CCW. They're synched to the main wheel via the little stator wheel on the left.

Obviously in MT 143 the weight levers collide every 180° - i'll try playing with that concept later, but for now the object of interest is simply the relative paths taken by the masses..




...note how the upper two masses (blue) trace a perfect circle, while the lower two (red) follow a mirrored epicycle.

This can be highlighted by tracing the trajectories:




So the inertia incurred by the two blue masses is minimal - they take the shortest available route through net space between inner (min MoI) and outer (max MoI) positions.

But the net distance traversed by the red masses is much greater - from the wheel's resting frame, both masses have travelled equal distance between inner and outer positions, yet from our external perspective, there's a massive inertial asymmetry - and since the blue mass trajectory is optimal (a constant rate of curvature / angular acceleration), and geared 1:1 to the red masses, the asymmetry must also be optimal for this particular arrangement.


Sooo, what was the point of all this again, i hear you ask?

Gaining RKE from closing MoI is a done deal, cash waiting, no questions asked. The only hitch is that we have to wind out our MoI in order to wring it in again.. and on the outbound stroke we lose all we gained on the inbound stroke..

So the question is, could there be a way of re-extending our MoI, without it clawing back all of our gain in the process...? We wanna get our MoI back out there, but without slowing down...

So the obvious implication of this latest observation is that, perhaps... we can take a long route in, and a short route out..!?

Again, both routes are equal from the internal (inertial) frame, but unequal from the external (non-inertial) frame, and it's the latter that traces the line integral of the net acceleration, and thus the work done.

Such an asymmtery may raise the possibility of unequal inbound vs outbound energies..!! Perhaps bouncing the masses off one another, per MT 143, facillitates this.. need to play more.. for now tho, the TL:DR is that i'm hoping to cause more net acceleration from decreasing MoI, than deceleration from increasing MoI.. so the wheel should gain momentum throughout a cycle..

[ME]

KAS, AFAIK those pendulums where never part of a real setup and only added to the drawings.
When a single mechanism has a positive behavior over a full rotation, then there should be no need for your proposed additions as it'll be smoothed out by multiple mechanisms.

A single mechanism could have a torque behavior like a Sine-wave (as an arbitrary choice), but with the positive being a bit larger when in overbalanced position. It still has some negative torque for the other halve of its cycle.
With two mechanisms such function will reach its minimum of 0 torque two times per rotation.
With three mechanisms the minimum torque will reach 0.866 times the amount of overbalance and doesn't have dead-points any more.
With this (note: a bit arbitrary) behavior it's also clear that an odd amount of mechanisms will result in a smoother outcome.

[ME]

Sooo, what was the point of all this again, i hear you ask?
Gaining RKE from closing MoI is a done deal...

The question is: how to get them back to the previous low RKE (or RPM) position while keeping the current RKE (or RPM).

[KAS]

ME, I am not sure where your assumption that the pendulums were added to the drawings came from but I agree that there may have been balanced points in the cycle. When a wheel becomes balanced, no torque is evident.
IIMV that at this point the pendulum(s) would lose energy and fall to giving back the energy stored. My own experiments with 4 pivotless mechanisms have indeed exposed 4 points of rest and 4 points of torgue as you would expect but the detrimental effect is CF. My system incorporates a strange phenomenon whereby the sub system's mass falls without losing height. I know, hard to believe but I assure you it does. The energy raises up to the mass through leverage causing it to fall against the force under the influence of gravity toward the axle.
I am sure Bessler's systems incorporated a similar moment, hence the pendulums.

[MrVibrating]

...an obvious interpretation of MT 143 at this stage would be that the purpose of the Roberval is to indicate that, although alternating between inner and outer positions does change the net system's balance, this overbalance itself is immaterial to the real matter of interest..

..and if it's not gravitation that's being affected.. that just leaves inertia.

Additionally, it would also indicate that rotations between inner and outer positions (esp. opposing CW / CCW rotations) are more interesting than linear radial translations between the same postions...