Blood From Stone

[TheVisitorV]

What? :)))))))))))))))))))))
no idea what you just said :))

[MrVibrating]

..on 2nd thoughts, Fletch, you might've been too prescient - so we've got 16 J of rotKE we've paid 8 J for; if we let the orbital MoI back out, whilst locked into the 'extended' (max MoI) position to 4 m radius, we simply harvest that 8 J gain as CF PE, right? Just load a spring or whatever - the rotKE will drop by 8 J whether that PE harnessed or wasted to heat.. so why waste it?

Thus we start with 8 J of rotKE, and end with the same, plus another 8 J of sprung PE.

We'd only destroy KE if we raised MoI whilst rotating, instead of reducing it!

Ie. once set in rotation, the orbital MoI converges to the net mass of the orbiting rotor, at the motor axis - so, down to '8' from '16' as in the first example - but whilst in that 'suspended' state, the actual absolute radius of the masses could be changed to anything - any radius, wider or smaller... and the orbital MoI will remain pegged at a value of '8' until the orbiting rotor stops again relative to the central rotor. The instant it stops, the orbital MoI returns to being a function of the actual mass radius, however high or low that is!

So we might think of what happens when we fire the motor as a kind of 'curtain fall', that sets an absolute MoI of, say, 8, no matter what may be happening to the axial radius of the constituent masses themselves whilst their actual positions are 'hidden' from the net system MoI calculation. We can change that axial radius to anything - it's utterly irrelevant to the orbital MoI so long as they're still rotating relative to it..

It's only counterintuitive till you start to see the warped logic..

But, harvesting rotKE via CF PE; definite angle, there..

[MrVibrating]

What? :)))))))))))))))))))))
no idea what you just said :))

"ninja'd" = a cross-post pre-empting another.

I posted the 90° drop in the post right before yours asking for it.. if that was the interaction you meant..

[TheVisitorV]

had a similar test in my software.

did not save the result that i will describe. Anyway, had 2 weights connected by scissors on one side,dropped them from 12, and when they reached about 6oclock had a spring preloaded that pulled the weights together, it kept going for several turns as the weights started to oscillate a bit due to the spring. weights going in and out.

anyway i'm trying as well to see if i can come up with something that does not require motors. i guess here is where gravity actually played it's part.

there are still many questions to be answered.

[TheVisitorV]

anyway, went ahead and redone the experiment...

https://youtu.be/No8Y4XRxcMA

[MrVibrating]

I'm guessing all options are open as to how to produce torque - taking it from CF force does seem viable, since that energy ultimately comes from rotKE, which is where our gain is.

The first thing that comes to mind tho is to take it directly from net system torque, via a stator - i can't see why a stator would destroy the effect since no net torque is actually required, according to the torque * angle plots - it's a flat trace, hence no counter-torque or counter-momentum is being applied either.

What i might try is a hanging stator - like a heavy pendulum that doesn't swing - hanging down from the central rotor axis (what Fletch refers to as an 'artificial horizon'); attach a 'transmission sprocket' to the bottom, driven by the main rotor via a chain / belt drive, with a second chain driving the orbiting rotors from that lower sprocket via a clutch or latch that can be switched on & off...

..a bit like this thing from page 27:




This way i can precisely control the relative orbital / axial angles as a function of the gear ratios, so they're always in-sync.

If that doesn't work, i may switch to trying to take the gain from CF force, or else perhaps using GPE's to drive the axial counter-rotations, tho that seems inherently messier..

At this stage, a neatly-revolving ' Bessler wheel' seems more of an aesthetic priority than a mechanical necessity - i can't see why a crudely-oscillating Rube Goldberg installation wouldn't suffice..


The stand-out feature of this particular scheme is that the maths and sims say we already have the gain in the bag - whereas, normally, any given attempt hinges upon trying to meet 'x & y' conditions that would access a prospective gain, here, we already have it.. As rotKE on the central axis - surely the ideal form of mechanical gain!?


There are still questions that need answering - for instance, does the axial counter-rotation have to be applied via torque against the main rotor (as by motors / drive shafts / rotary springs) - ie. does the effect depend upon that counter-torque being applied back to the central rotor, or else, could a 'reactionless' form of torque suffice, such as a counter-rotating OB weight attached to the axial rotors (which thus applies no counter-torque at the orbiting axis)? Using another MoI variation to generate inertial torque on the orbiting rotors wouldn't seem to work, since that torque would be in the wrong direction (in the same direction as the orbit, instead of against it).

Will just have to trawl thru the various options and see what works and what doesn't..

[MrVibrating]

So here, the orbiting motors have been replaced by an internal transmission system, using an external stator wheel affixed to the background:



I did try a hanging stator instead, but it has to be very heavy to prevent significant swing, and measurement accuracy takes precedent over aesthetics.

So back down to Earth with a thump we go. While ostensibly preserving the sequence of actions, we've killed the gain. The sim ends when it runs out of energy (not interested in the back-swing).


So, what's happening?

At the instant the rotor clutch unlocks and the transmission engages (90° in) - where the rotKE is supposed to suddenly double.. it halves instead!

One frame before 90°, rotKE is at 23.531 J..

..one frame after, it's slashed to 11.82 J!

So, very interesting result, and equally informative, with a bit of prodding..



Why's this happening then?

Is it something to do with trying to employ a stator, as 'forbidden' by Bessler - ie. we're now applying orbital counter-torque / counter-momentum to Earth, thus breaking the conditions for an effective N3 violation?

Or else, perhaps there is an issue with the motor torque * angle plots after all?

For a third, somewhat more convoluted, option; maybe the analysis framed on the previous page is closer to the mark - work is being done by the motors, but is being instantaneously reciprocated with a reactive inertial torque of equal sign and magnitude..?

Because for now, the results seem paradoxical... and there can be no paradoxes!

I've considered trying to use rotary springs instead, since a spring is definitely either loaded or unloaded with PE as a function of angle.. surely then, the 'gain' would be equal to the PE expended by the springs?

Yet in that case, why can't we see any torque on the motor T*a plots?

Intriguing riddle if nowt else, eh? :/


Note that the orbital CF profiles still cancel nicely tho.. :\

[MrVibrating]

..just musing aloud - if i replace the motors with rotary springs, how would i govern the axial rotor speed to keep it counterposed - and thus non-rotating - relative to the orbital rotation?

The precise torque profile required for that balancing act is presumably non-linear, at least if we're also applying a GPE..

..remove gravity again, and the velocities stabilise such that the spring simply has to unload at a constant 1 rad/s, which could be accomplished with a linear spring constant...

..but then there isn't really any such thing as a linear spring constant, at least, not across any significant range of displacement.. I could model one... but in effect it'd really just be another motor by another name, rather than a true 'spring' per Hooke's law..

Dunno.. first thing to do seems to be: get rid of the stator. Keep all torques and counter-torques internal to the system, to contain all momentum and counter-momentum. Whatever's going on in the motorised version, the momenta seem to be battling it out to produce a 'win' in one direction..

One more thing i should prolly do first is rebuild the MoI and momentum meters... because in that negative result above, is the net system momentum still constant throughout the interaction? Or does it get slashed along with the rotKE at the 90° mark?

And isn't the transmission clutch actually dissipating half the rotKE when it engages? Does a motor act as a 'generator' under that same condition, thus preserving that torque * angle?

First things first; let's get the momentum metered..

[MrVibrating]

..just to keep yuz on yer toes, here's the original result again, slightly more spruced:



If i may say, this is a very clean result, using functions about as elegant as i can formulate.. the outward simplicity of those nice round numbers belies the tedium of obtaining them!

But just look at the MoI / momentum situation - the former suddenly halves, the latter holds fast unwaveringly, hence rotKE duly doubles..

Hence my presumption that we must be losing half our momentum to Earth when attempting to redistribute it via the external stator..!?

So, something to chew over while i knock up a fresh MoI meter.. (and maybe first, some dinner)

[Fletcher]

FWIW .. I've encountered it before in sims.

Doing various experiments to see if different types of Momentum is in fact conserved along with E in the Outputs and Meters. What I seemed to find was that E was conserved as per CoE, but Momentum could disappear from the screen and the traces in front of my eyes.

I concluded that momentum wasn't really annihilated as the meters suggested but that the sim simply couldn't show the momentum gone to earth, tho that is what presumably happened to conserve momentum too. It seemed the logical explanation for the disappearing momentum from Outputs.

It lead me to presume that the sim has an order of priorities in how it calculates its Outputs and shows them.

To conserve E > to conserve momentum (as a second level visual conservation).

I was never able to get a programmer etc to confirm that wm2d basically conserved E above all else in ordinary sim situations.

You can of course override the sim software to some degree, by introducing fake force vectors etc but in the final washup they still come under the CoE umbrella because of the Work Energy Equivalence Principle of Physics. So the sim is consistent and keeps things nicely accounted for, imo.

[MrVibrating]

I think CoM's probably the only thing the sim needs to get right, since CoE simply emerges naturally from N3 & N1. KE's tertiary to the equality of momentum and counter-momentum - and we already know (examples earlier in this thread) that it will show OU when N3 and/or N1 are circumvented.. which is no surprise, since it basically cannot avoid doing so; its job is simply to report half the mass times the velocity squared, and since motion's relative, so's velocity and thus KE, to a given non-inertial (ie. zero-momentum) reference frame.

What i suspect's happening here is that the counter-torque from the motors is being neatly cancelled by a reactive inertial torque caused by the instantaneous halving of orbital MoI the split-second that axial torque's applied.. but let's get all our telemetry back up & running - hopefully it'll speak for itself..

[MrVibrating]

So with MoI back on the plot we can see that net momentum does indeed take a dive when we pop the clutch:



Whereas, going back to the prior, motorised, version, we don't see that at all:



..from observing the momentum trace alone, there isn't the slightest blip to betray the MoI suddenly halving.. The net momentum's conserved, only changing as a function of GPE in/out!

Given that the clutch can only dissipate KE, not momentum, the only possible source of leakage is its only other point of contact with the outside world besides the main axis.. that stator transmission wheel!

So, stators kill the gain... suggesting we are indeed in N3-busting territory.. even though the effective symmetry break is mass constancy rather than N3 per se.. the net result is a divergent reference frame; the KE gain only exists in relation to the outside world (which is all we need, ie. it's effectively 'real'), but its generation depends upon the isolated containment of all momentum and counter-momentum during the gain cycle.

Afterwards, once coasting at a fixed MoI in its new higher-energy state, the gain's immediately ready for use - it's twice whatever we began with, in the form of rotational KE, and obviously at that point we could then harvest it by torquing against a stator or whatever, like any normal motor.

But to cause the gain in the first place, it seems it's essential to apply these axial torques without recourse to an external stator..

So what to try now, but springs..?


We know that Bessler was seen to "push down on a spring" when reloading the Kassel wheel, implying they were radially-oriented. This is the ideal plane of alignment for harvesting rotKE internally (ie. without needing a stator), via output CF work..

But the clue we need at this point is that we might be able to use such radially-loaded springs to drive the axial rotation..

If that works, and all our momentum's securely under house arrest, then converting the resulting rotKE gain back into sprung PE should be a doddle..

So that's my weekend set.. provided we're not all jellified by microwave weapons overnight.. (silver alert! Tinfoil EVERYTHING!)

[MrVibrating]

I believe what we're actually seeing is an orbital variation on this:



..only here, the inertial torque is being generated by the MoI halving when we fire the motor, whereas there, it was generated by physically changing the radius against axial CF force... the cost of which equals the 'gain' (OU work by the motor).

And that's the other difference here - the CF profile's now orbital, not axial, and so the CF integrals now sum to zero!


As such, this thread appears to have culminated in a solution to the impasse the previous one concluded in - exploiting the reactionless nature of inertial torque to exact OU performance from an otherwise-ordinary motor - or more generally, any application of conventional torque & counter-torque, per N3.

We know Bessler wasn't using motors so we're looking at torque from PE.. weights, or springs..

[MrVibrating]

..for anyone not quite seeing it:

• In that older "Counter-Torque vs Inertial Torque" rig, the motor has only performed ½ J of work (torque * angle).

• Yet the rotational KE produced by that work rises by 1.5 J; 1.5 / 0.5 = 3x OU.

• This result arises because whilst the motor's applying torque to the rotor, it's also applying counter-torque back to a 'stator'...

..however that 'stator' is a co-rotating 'vMoI' - a pair of counter-balanced orbiting masses, pulled into the axial center at a controlled speed, such that the resulting inertial torque perfectly cancels the counter-torque from the motor spinning up the rotor.. get it?

So you have a closed system of mutually co-rotating masses, in which you can spin up a rotor, without de-spinning the world around you!

Spin up a heavy rotor in a spaceship, and the view outside's gonna change.. but cancel that counter-torque with an equal opposite inertial torque, and due to its reactionless nature, the scenery holds still while the net system KE must rise accordingly.

And so, the only thing thwarting OU in that rig was the CF cost of pulling those two orbiting masses inwards..

..whereas now, 30-odd pages of mostly-redundant work later, we've finally found a way of getting that CF cost down to zero!

That last motor-plus-GPE sim really is showing 3.57x OU.. it's being calculated independently in real-time by both me and WM in parallel, with perfect agreement..

Ain't the home straight no more, we're over the line.. swapping from motors to springs seems almost academic..

[silent]

.