Blood From Stone

[Georg Künstler]

MrVibrating wrote:

It does this by varying the time spent gravitating on either side of the interaction.

It is exactly that, what is going on in the Bessler wheel.
The time spend on one side is longer than on the other side. This will create you the momentum.
Give gravity the time to act.
It is the time difference between up and down.
The movment up is faster than the 9,81... here on earth.


Very good insight, MrVibrating. Don`t leave this path, even nearly all of the experts will disagree, you are right.

Maybe this will help you, to go over the fence:

there are 3 types of gears:

the first two types are used in the mechanic, but the 3 is for the mechanicans to today worthless.

It is a turning wheel in a bigger wheel.
Both are turning in the same direction.
This will allow differential speed.
The inner mechanism is beeing used as a driver, the outer as the cyclical braker.
This construction I have presented already in the university of Bremen years ago.
They have messured it, input and output and it got more output than input.
But this was achieved only at a specific speed.
They declared, it can not be and it must be a measure error.
As all experts disagree with me, I stopped this development.

Now I see, I was right and the experts wrong.
You proof my physical model with your knowledge.
Again, great work of thinking.

[MrVibrating]

Thanks Georg - like i say, this trick does not look very impressive if one is only considering energy - i have previously used MoI reduction to boost a lift, but usually with a focus on energy cost / benefit, of which there is none.

However it does affect the rate of change of upwards vs downwards momentum exchanged with gravity.

Only inertial torque - from an MoI variation - can do this. If we tried it using conventional torque instead, we'd find that we must inevitably apply counter-torque back to the wheel, so that even if we achieve a faster lift than drop, the momentum in each direction remains equal and opposite.

But because inertial torques incur no instantaneous counter-torque, or thus, counter-momentum, we can boost the lift speed without applying a corresponding counter-momentum to the rest of the wheel, hence spending less time gravitating.. and since our lift & drop speeds are different but gravity's rate of change of momentum is constant, we've introduced a time-dependent momentum asymmetry.


Thinking a little more about it, i still suspect that this won't be enough to break energy symmetry - collisions will still be required to break things up further in that regard, but i'll have the energy balance metered eight ways from Sunday soon enough.. it seems for now though that we'll get less energy back out from CF F*d as the masses move back out at slower speed than they're moved in, so it's "3 steps forwards, 2 steps back" - that 1 step difference being our energy cost of momentum, and since CF=mV²/r, it's going to square with rising velocity, hence following ½mV².

But this is expected, and no disappointment - we know how to tackle it (inelastic collisions) - again, it's not the above interaction that's such a breakthrough, as comprehension of what's happening momentum-wise. This is a major battle won - the war's not over, but final victory's in clear sight.

As i've long said, even if you had a system that could freely vary MoI - adding free KE to the system like this still couldn't change its momentum - if you began with 1 p of momentum and 1 J of energy, you might end up with 1 kJ of energy, but you'd still only have 1 p of momentum.

You cannot get momentum from just anywhere, and certainly not from just energy..

But gain momentum, and all bets are off.. wrt energy, too.. We can now create a divergent reference frame.

[Georg Künstler]

Hi MrVibrating,
your system is a fix point system with a fix axle.

A swing beam on a axle will allow you that the mass can be transferred inside and outside.

If you turn the swing beam on an axle, and increase your speed, then CF will take over the control, the then mass is on the outer rim.

I used a second system, running with a different speed, differential speed.

if you make now a collision between this two system, while turning in the same direction, you will get a shift of the mass inwards.

Doing this collision shortly after the 6 o'clock position it is like a lift against gravity in a part of a second.

After 180 degrees, this action can be repeated. Gravity has time to act on
a longer lever arm.

[cloud camper]

Great effort to try and document your experiments Mr V but I'm seeing an endless procession of nice smooth mathematical curves - perfect mathematical symmetry!

I don't claim the answer but I believe we're going to need a step function maybe something like the flea jump mechanism to break free of mathematical symmetry.

Keep trying tho!

[MrVibrating]

The first runs of the current system were symmetrical; the current version is an asymmetric gravitational interaction, gaining momentum from gravity every cycle.

It's a gravity motor, bashically.... all of the angular momentum is provided exclusively by gravity. Just not the energy.. yet.

Unlike almost any other kind of motor, it has no stator, and no top speed - provided you can keep the MoI variation going fast enough anyway. It's more like a waterwheel than a conventional motor - it has a greater preponderance of angular momentum of one sign over the other, caused by gravity's constant acceleration in relation to the time-delayed fall vs fast-lane rise.. in short, the weight spends more time gravitating on one side compared to the other, and this is where the continual rise in angular momentum is coming from...

Yes it doesn't look like much, but it's a key condition for breaking PE to KE symmetry. A necessary condition, if not a sufficient one. For now, its efficiency needs establishing, then we can start looking at collisions and whatever. First things first: collect underpants, ie, momentum, from gravity (from gravity! (how nuts is that?)), and like, measure it and stuff innit.

[MrVibrating]

Consider this alpha state - it definitely has bugs (current work integrals appear way out) - but it just gives an idea of direction of progress:





Initial system energy:

KE = 0.5 J

GPE = 39.2266 J (9.80665 N * 1 kg * 4 m)

Total = 39.7266 J

___________________


Final system energy:

KE = 30.03949 J

GPE = 39.2266 J

Total = 69.76609 J

___________________

Increase = 38.53949 J


____________________

Initial System AM = 2 p

Final AM = 15.82641 p

increase - 13.82641 p

____________________


Actuator F*d = -16.71154 J

CF/CP (F*t)*t = -32.3921 J

CF/CP F*d = -5.83162 J


_____________________



Formulating and debugging the meters takes time - but you can get a sense of the range of variables that could be used to determine the input energy costs; these are going to have to cross-reference consistently to fill in the picture of exactly what's happening.

In the last set of sims for my "inertial torque vs counter-torque" tests, i found that WM's measure of forces acting on the actuators is unreliable unless they're set to apply 'acceleration', whereas i was initially only controlling their length, leading to erroneous work plots. Here, i'm using them this same way, so the "Radial F*d" graph is almost certainly junk. I did try syncing them using 'acceleration' instead but couldn't get it stable, so stuck with 'length'.

CF/CP force, as a function of mass * rad/s * radius, seems a safer metric - and the closest answer to the 38 J KE rise (CF force * time = power, times time = energy). CF force plotted against displacement however seems to be giving an inconsistent result.. clearly something wrong there..

Give it a few days, it'll get there...

Doing it one cycle at a time for now - attempted a 10s run initially but useless to try run before you can walk.. next cycle will pick up where this one leaves off... tho that could be a week away; one step at a time eh.

My money's still on ½mV² FWIW..

[MrVibrating]

Found a pair of seemingly-reliable metrics - got actuator power times time working, plus CF/CP force times radial velocity times time, for the CF/CP work integral. Here's the first cycle:




Stats:

Actuator power times time: 29.57742 J

CF/CP force times radial velocity times time: 24.64732 J

29.57742 - 24.64732 = 4.9301 J (radial KE is 5 J)
__

Initial KE: 0.5 J

Final KE 30.05253 J

Difference = 29.55253 J - equal to actuator work

__

time / stroke / Sys Momentum (kg-m²-rad/s)

0.00000 / TDC / 2
2.12960 / BDC / 23.91240
2.80560 / TDC / 15.82621

e/p efficiency:

29.55253 J / 15.82621 p = 1.86731 J per kg-m²-rad/s


Now that the accounting seems somewhat reliable, i'll try a longer run of many cycles, to see if that e/p yield changes. If not, then this stage of testing is pretty much done, and we can start adding in collisions.

Consider what will happen if this mechanism constituted 50% of the net system MoI, with the other 50% made up by a uniform disc, and both rotating together at equal speed. The mechanism accelerates ahead under gravity, as here, and having been further accelerated by an inertial torque it then collides with the solid disc, instantly doubling the MoI the gained momentum is distributed into, hence halving the velocity, and hence halving the CF force that this quantity of angular momentum pertained to when MoI was 50% lower.

Hence we can grab some momentum from a gravity/time delta, knock its velocity down with a collision whilst still preserving all of it, and thus lowering the CF force and thus energy cost of the next such cycle - and so stepping off the ½mV² escalator.

Alternatively, we could try use a further MoI variation, instead of a collision:

• suppose we have another pair of radially-sliding masses - a second variable MoI

• this secondary vMoI remains constant throughout each cycle, but increases between successive cycles, thus raising net system MoI, lowering velocity for the current given net angular momentum, and hence lowering the CF workload and thus energy costs on the primary vMoI

• so as with collisions, this now forms the step-wise sequence; gain momentum from the gravity/time delta, then extend the secondary MoI, and repeat

Doubling MoI - whether by collisions or further MoI variation - quarters current RKE either way, so we still get that 25% per-cycle efficiency accumulator, starting out under-unity, then hitting unity, then topping it.

Furthermore, limiting the speed like this, in spite of rising net system momentum, helps maintain the time/momentum deltas, and thus the e/p efficiency. Obviously, the faster the system gets, the less time overall is spent gravitating per cycle - in principle we can maintain the asymmetry ratio, but the absolute amounts of momentum in and out per cycle decreases with speed.

And that's the simple picture of it - in a nutshell, momentum is being pumped into and then back out of the system, by gravity, but we're skimming some off. The total amount pumped in and out per cycle decreases with RPM, because gravity's acceleration is constant, so less time gravitating per cycle means less momentum exchanged, and so a smaller cut per cycle.

All else being equal, system velocity rises with rising angular momentum, and hence so does CF force, which sets the energy cost of running our skimming operation. Hence, the cost of accumulating that momentum squares with rising velocity, per ½mV², and no banana.

But because the inertial component of angular momentum - MoI - is dynamically variable, and momentum is conserved, we can step down the velocity, and thus the impending energy cost of the next momentum-gain cycle, so accumulating reactionless angular momentum embezzled from gravity for a fraction of its ½mV² value..

I don't foresee any real challenges anyway. The overriding exploit - the 'zero day' - is the time-dependent momentum asymmetry, an effective N3 exception that allows accumulation of angular momentum exclusively from gravity. No stator required. Zero counter-momentum induced to the system. Perfect conditions to start decoupling PE from KE..

All that remains is to tame the CF as the momentum builds up, thus keeping input energies down - and we still have a third technique on reserve - off-axis (ie. orbiting) moments have a CF profile completely independent of the main system axis (as demo'd previously). All in all, as far as i can see there's no reason CF has to increase linearly with rising angular momentum..

Can't repeat this enough - whatever kind of motor one may employ, the best efficiency you could hope for would be 0.5 J for your first kg-m²-rad/s of momentum, and then 2 J for the second. But if you could somehow pay only 0.5 J again, you'll have bought 2 J worth of KE for just 1 J.

Even if they were 0.75 J each, that's still only 1.5 J of PE for 2 J of KE. ½:1 is obviously the optimum e/p ratio to aim for, nature's base rate - we're currently at 1.86:1 here, but again, whatever the ratio, even if it's speed-dependent, so long as it doesn't quite track ½mV², there's our energy gain margin. In practice i think OU in 5 cycles should be easily attainable..

[silent]

.

[cloud camper]

Mr V has designed a very novel mechanism here and does all the right stuff to analyze the reactions.

Unfortunately the actuators have to work against CF at the worst possible time (when rpm is greatest).

There is nothing anomalous occurring here as the work done by gravity on the weight is exactly equal to the work done by the actuators.

The novel aspect is using a forced MOI variation to lift the weight back to the initial condition.

Perhaps combining with impacts in some manner the momentum can be conserved so it is an interesting concept!

[ME]

Nothing anomalous.
At least in V's latest reply to me he "wasted a lot of effort" to repeat his claim that he showed that "momentum accumulated from gravity". But that's simply false.
As a pendulum-part the green-weight may prime the system with its tendency to convert GPE-->KE-->GPE but that's just a fixed amount and unable to accumulate the momentum per cycle as visualized . This conversion also gets limited by those actuator weights- which are balanced around the axle (so no effective GPE from their part).
But GPE-->KE conversions are, as usual in PMM research, the culprit to overcome by other means than hidden energy sources.

This mechanism under investigation is about pushing the pendulum (green weight) by exploiting Conservation of Angular Momentum (CoAM) when this pendulum goes in a certain direction (up), and then wonder why things accelerate....

--

To expand on that:
Forcing the actuator outwards when the Green goes down does not require much energy... It may take some effort to keep its traversal speed in check. In fact, at a certain rotational speed one could even extract energy from those actuators! -- or I think it would practically require energy to slow down the expansion.
The real action happens when the Green goes up. Of course this pendulum-part should drain velocity from the system.
By pulling the actuator-weights inwards then, by a reduced MoI and what is known as CoAM, we'll in fact speed this wheel up to overcome such drain. To make that happen it also needs to overcome the current centrifugal force while designed to have its speed synchronized by frequency... it needs more energy on multiple counts.
It may be the imbalance by the green weight (gravity) causing an imbalance in the energy requirement, but that is programmed/delivered/overcome by the actuators. It is basically going down automatically, going up for almost free.

Nevertheless, mrV will have a really nice job in figuring out the energy flow of these things, likely requiring some sweat and perhaps tears: But its seriously a very rewarding exercise!
I advice to attempt a specialized solver, or otherwise just extract all the values from the simulator and puzzle them together.

On a practical side-note:
I seriously doubt any affordable powered actuator is up to specs to perform the required action... Plus a feed-back controller is a requirement, perhaps an Arduino/Rasp.Pi would do.
When you'd choose a mechanical solution like a "trammel of Archimedes"/"bullshit grinder" then this whole setup would just become a fancy pendulum - behaving only slightly different from a normal pendulum.

[MrVibrating]

Right, instead of continuing one cycle at a time, i've just run a 10 second burst, with eight consecutive cycles, at the highest possible resolution using all the memory WM will address:


(static image, anim would take too long)

Actuators power * time = 236.9473 J:

• CF/CP workload = 197.3897 J

• Radial KE = 39.52161 J

• 197.3897 + 39.52161 = 236.91131 J


Input and output reference frames are still coupled as expected, because the centrifugal/centripetal workload is increasing linearly with rising angular momentum.

Now to negotiate the next step... decoupling the I/O energies!

The first, most obvious way to attempt this is by using inelastic collisions, which conserve momentum but not KE..


Currently we have the ability to cyclically accumulate momentum from gravity in a closed (ie. statorless) system of interacting masses. For our purposes, this is an effective violation of the first and second laws of motion.

But to capitalise on it, we need to apply it to generate a "divergent inertial frame" - so, accumulating angular momentum to create an accelerating reference frame, within which the energy cost of producing that momentum is less than its KE value in the external, stationary (ie. non-inertial) frame.

If we collide the current apparatus with some inert but equal MoI, such as a co-rotating solid disc, at the culmination of each momentum-gain cycle, we halve its current KE value, but also quarter the cost of raising the next metric unit of momentum, since the CF/CP workload squares with velocity.. thus we should be able to step off the ½mV² ladder, and climb onto an altogether different one..

So as before, the plan is to test a series of such cycles, to see how the I/O energy efficiency evolves..

Obviously, the expected pattern is going to start out lossy, but with an incremental improvement over successive cycles, as net velocity increases.. Remember, we can tolerate a rise in input energy costs with velocity - potentially, quite considerably so - provided it doesn't simply square with velocity.. so if it just doubles per cycle, fantastic. If it triples, we're still OU. Just so long as it doesn't quadruple, there should be a break-even velocity threshold within reach..



Attached sim only useful for data extraction (takes an age to run).

[MrVibrating]

Mr V has designed a very novel mechanism here and does all the right stuff to analyze the reactions.

Unfortunately the actuators have to work against CF at the worst possible time (when rpm is greatest).

There is nothing anomalous occurring here as the work done by gravity on the weight is exactly equal to the work done by the actuators.

Nothing anomolous!? Are you for real?

There was never any anticipation of a break in I/O energy symmetry!

How many times do i need repeat this? The purpose of the experiment is to establish cyclical gains in momentum, from gravity.

Moving masses in and out cannot add any momentum - on the contrary, it conserves it.

But the torques applied in that effort cause a differential between the time spent gravitating between rising and falling strokes of the gravitational interaction.

Hence, because the rate of exchange of momentum of a gravitating system is a time-dependent constant, it is trivially simple to engineer a time-dependent momentum asymmetry, accumulating momentum, without counter-momentum, in a closed-cycle system of interacting masses.

Thus achieving the prerequisite conditions for a subsequent I/O energy asymmetry; a divergent inertial frame.

As ever, if you just stubbornly refuse to grasp the relationship between CoE and CoM, you deny yourself the prospect of ever comprehending the implicit nature of mechanical OU, and thus whatever must've been going on inside Bessler's wheels..

The novel aspect is using a forced MOI variation to lift the weight back to the initial condition.

No! Talk about not seeing the wood for the trees... the novel aspect is what this achieves! A 'statorless' cyclical momentum gain! The basis for a frickin' divergent inertial frame! The rolling platform upon which we will attempt to build momentum at less energy cost than its resulting KE value in the external reference frame!

It's like i'm showing you the plans for a hydro-electric dam, and you're all like "but it's just a big wall, how's that make energy? All that concrete must cost a bit?"...

Perhaps combining with impacts in some manner the momentum can be conserved so it is an interesting concept!

Argh! Mate, it's already being conserved, and accumulated! The purpose of impacts is to break up the distribution of 'inertia' and 'velocity' comprising this conserved momentum, thus disrupting the usual ½mV² accumulator on the cost of its production!


The exploit is little to do with 'lifting weights' - that's almost incidental - it's all about accumulating cut-price momentum from gravity.

That singular source of momentum is the titular vis viva being 'wrung from stone' - scoff as you may, but sh!t just got real..

The intention is that finishing each of these momentum-gain cycles with a collision into a co-rotating equal MoI will 'bank' the gain whilst halving the velocity, so providing a significant advantage to each successive interaction's e/p yield, that will cause a divergence from the usual ½mV² accumulator on the cost of raising momentum.

Please try to refrain from misrepresenting my intentions?

[MrVibrating]

Nothing anomalous.

..ditto what i said to CC.

The momentum gain is precisely equal to the lift/drop time differential, relative to the gravitational constant.

This has been conclusively and empirically established for the last year, and is not news...

You cannot add angular momentum to a system via MoI variations alone, period. They can only vary the KE.

Only gravity can contribute momentum, because it is a uniform acceleration - an ambient rate of change of vertical momentum that is constant with respect to time - hence by lingering during the drop, whilst expediting the lift, we gainfully manipulate the weight's exposure time to that ambient rate of change of momentum.

Momentum's being pumped in and out by gravity, but in unequal measure.

For your reference, consider again what happens when the two interactions are rotated 90° relative to one another:



..the system no longer gains momentum each cycle, because the net lift and drop periods are equal, thus so is the net momentum in and out from gravity.

Again, the anomaly is a closed-cycle, cumulative rise in momentum, sans counter-momentum, from gravity, in an otherwise closed-system of interacting masses, with no stator.

You may find this trivial, but it's an effective violation of all three laws of mechanics (and would be all the moreso, by your analysis).

...and FYI, the energy accounting is not that hard - i figured it out in a day (ie. in a few hours after work). There's radial KE, and a radial inertial workload, and their sum is equal to the net input energy and net KE rise, precisely as expected.

Still, with your focus on the inertial workloads, consider what future consequences will transpire when we cap each successive cycle with a collision into an equal co-rotating MoI.. thus halving the velocity of the gained momentum, and so quartering the energy cost of the next cycle, while likewise doubling its time-spent gravitating..

Does the energy cost of raising momentum still follow ½mV² under those conditions? That's what the next round of experiments will investigate..

[Georg Künstler]

Hi MrVibrating,
don't waste your time to fight against others.
They don't know it better.

What is named an expert ?
An expert is someone until someone better is showing up, and then this person is the expert.

Expert knowledge normally has an expiration date.

So keep going on, you are an expert on this development field, the others will follow soon.

The momentum gain is precisely equal to the lift/drop time differential, relative to the gravitational constant.

This is the reason to build a wheel big, so that you can see with your eye this difference. In my wheel the up is faster than this 9,81...

[MrVibrating]

eta:

And as for build-practicality, it's completely irrelevant to the current experiments, which are designed to be elementary measurements, with no fat on them.

It is far too premature to speculate on what form a practical implementation may take, though needless to say, it need not be mathematically optimal to merely harness the effect.

On that count however, note that the current design is probably not fully optimised either - we could further increase the up vs down time differential by using more non-linear actuator timings; you'd think that the optimal performance would see the MoI changing much more quickly, through a much smaller angle of rotation, so that MoI is maximal for most of the drop time, and minimal for most of the lift time. Currently, the radial speed - and thus acceleration - is purely a linear function of the weight's vertical speed and accelerations.

But all we need for now is a simple test rig that can consistently reproduce the effect, with clean data.

If however you're already trying to race ahead and visualise a wheel design using this principle, you'll inevitably be dealing with a more complex system; here, half the mass is gravitating, the other half's radially-translating, so straight off the bat, we're not gonna get such a clean-cut distribution in the real world - some mass will inevitably be neither gravitating nor changing radius.

Furthermore, don't necessarily assume this interaction will be centered at the main system axis of rotation - it could be performed by orbiting mechanisms, which then transfer their momentum gains to the net system via the 'torque couple' effect (basically, angular momentum's conserved regardless of its radial location relative to the axis of rotation; step onto the outer edge of a kids' roundabout whilst carrying a heavy spinning gyroscope, and then brake it to a halt with your hands - the roundabout will spin up with precisely that amount of angular momentum).

But that's getting much too far ahead. Spring-loaded scissorjacks, some kind of crude-but-effective artificial horizon sensor / trigger system, how to actually capture the KE gains and reload the springs etc. etc. - all those no-doubt vital details - are engineering issues for later, not physics issues for now.. 'Step 1' is simply to collect underpants..