Towards A Complete Maxwell's Daemon System

[d3x0r]

Some glossary sort of definitions
   angle - the idea of angles requires 3 points, 2 on the arc covered by the angle and a central origin point; the origin is the centroid of the circle containing the arc.  The angle and the arc-length for 1 time unit are the same thing.
   curvature - the idea of curving, requires 2 points, from, and the to, with a forward vector, which is curved a certain amount per time...   The origin that is being moved around can be computed, but it's not really part of the system, but a projection of the real value.

   spin - the idea of spin requires 0 points, but and moves all other points around it the same amount around that point, it's value is not a point, but a curvature.

   Angle in unit-circle/unit-sphere math which is measured in or radians, is exactly the same as the arc-length, and really only differ in the idea of them; the values and functions applied to those values are exactly the same.
   Curvature is usually measured by the arc of a curve covered in a certain amount of time...  the full arc can be covered multiple times per tick, wrapping many times, before coming to a final state where there is a principal angle plus the number of wraps it did..  A thing that has 0 curvature is a straight line.

There's a theory that one would be able to make a maxwells daemon that could move things from the cold side to the hot side, and not lose entropy, but convert that entropy into a sort of state memory which moves the entropy to the daemon instead of the space.  Being a software developer, I'm all about 'state memory' which can take the forms of many things, including implicit definition in the system itself; like there's certain things you know just from having the numbers in the first place, like in Physics there's 'inertial frames' and 'non-inertial frames' which really are 'real frames' and 'relatative frames between other frames'... the latter don't REALLY exist, except in math, and maybe for discrete systems.  Quantum physics sort of puts a box around the combination of frames that averages out any individual motion into the average of the system over time... Consider a few real frames like a box filled with a bunch of (ping pong?) balls, loosely packed, and in microgravity.  Moving the box is going to effectively make an 'invisible' acceleration force in the relative frame between the box and the balls such that the balls would 'observe' that they were falling towards the size of the wall... their impact with the box is going to adjust the boxes frame and adjust the relative frame to actually be moving faster than the box as a whole...

Anyway... if we had perfect knowledge of the current state of all spins in a system there wouldn't really be uncertainty... but the exact spin cannot be known by observation... but moreover, electrons spin, as they spin, they orbit, as they orbit, they are in bodies that are on a planet that spins, which rotates around a sun, which spins around a galaxy, which orbits other galaxies and have virtual origins they rotate around...  Even if the exact moment of spin is known for a thing, the composite of all other frames at the same time that a certain spin happens cannot be known....

Let's suppose there's a few 'states' of quantum matter... and having more than 1 or 2 states starts to give enough room to have memory just by things being in a state different from another state... so we can move a bunch of entropy from one state to another state, without losing or gaining overall energy.  What sort of states do we observe?  Photons?  Matter/condensed photons?  (black holes? pure spin 0 linear component?)  Photons seem to have a linear component and 0 spin; that is as they travel, when they arrive they are in exactly the same state as they left, unless there is an interference along their path (quantum cryptography is based on this idea).  They might have an spin component coincident with their velocity though, and spin helically in their same direction (but I expect this isn't the case; later, limit on spin vs velocity esp. 'causality' constraints).

Let's suppose also that the only 'frames' in the universe contain a velocity axis and a rotation axis, which are often nearly the same.  Let's also suppose that spin, or curvature is a primary component of things, not just a projection of the points a rigid body representing rotation, but as a truly intrinsic part of 'things', notably at the very smallest and very largest of scales.

---

I should take a moment to reflect on some origin story of this theory... I was working on a few projects this year.  the first, takes a cloud of points in an x/y/z grid, and computes from their values the shape of a surface.

https://github.com/d3x0r/MarchingTetrahedra  (notes and references in readme) https://d3x0r.github.io/MarchingTetrahedra

I then added better texturing based on the 'element type' of the voxel (water,sand,dirt,grass,granite,... ) https://d3x0r.github.io/IsoSurface-MultiTexture/ 

So, these make pretty pictures, one of the 'input data' choices is 'Sine Waves' which is (0=sin(x)+sin(y)+sin(z) ) where the 0 surface is drawn.  This can obviously go from -3 to 3 as a maximum value... if the '0' is moved to +/-1 the shapes start to look like steinmetz solids(wikpedia).  at +/- 2 they are more isolated forms and plot as little balls... But so far I can only say 'this is a pretty picture generator' the significance will be mentioned later.

---

So; I have this mesher that does a really good job of giving me the shape of things, now I just need a good system which I can move and collide these clouds of points... the intersection actually isn't that hard, there's a common technique called 'octree' in 3d graphics, which doesn't have to actually be a 'thing' but can just be a computation laid on an existing grid of points to easily resolve 'is there something here in this region or not'.  I did have to take a few leaps and experimental stabs that were outside of the 'official' standards; but having been an independant researcher for like ever, I have the freedom to make such choices...

But, since I first wrote my first 3D thing; there hasn't been a system where I can just add and subtract the angles, and get the relative orientation of one body to another; and this is my 25-30 year contribution...
https://github.com/d3x0r/stfrphysics#space-time-field-reactor-physics

Rotations that you can just add and subtract!  They're NOT non-linear... in fact the problem that has kept this from being done until now, is that they're maybe TOO linear.

It turns out, the sum of the angles around 3 axles of rotation is the sum of the rotation of the system.  There's a measuring system called 'taxicab geometry' which has a 'manhattan normal' which, because of the constraint of x/y summation restrictions, the sum of x+y is the distance, and not the sqrt(x*x+y*y).  The normal of rotations is also a manhattan normal; and makes it so 1 rotation (a complex number) works great, but when you start adding more dimensions, things start 'not working'.  This 'manhattan normal' is actually a 'new idea' to geometry... and none(!) of the rotation systems really account for this...

There are a bunch of references in the STFRPhysics readme; a few of those are log-quaternion systems, where they did come up with the (|x|+|y|+|z|)/2 is the effective value to convert log-quaternions into quaternions.  The basic definition of the log of a quaternion is actually somewhat 'wrong', and that lead to 'Reimannian Space' which itself is a hyperbolic space... which only leads more to 'maybe real space is curved too'   It's not, and rotation space doesn't have any invisible walls or asymptotic conversion points.

I made a lot of graphs along the way...

https://www.geogebra.org/u/d3x0r

https://www.geogebra.org/m/jnk82ncd (arccos(cos(a)+cos(b))

https://www.geogebra.org/m/pwjdwzrz  ( phase space of rotations around other rotations... is an interesting transform... meaningless to someone who doesn't know what part of the equation that represents... so 'oh look, pretty pictures, you can change with sliders)

https://www.geogebra.org/m/whnktbpt  (this is attempting to build an extended arcsin() function... arcsin ends up having infinite holes (see attached graph of 'Sine Waves' which the surface is also the 'arcsin(x,y,z)' well each plane slice through that is an arcsin curve... the problem with arcsin, is it goes up, and then backwards... never effectively leaving the origin, it just loops... cos on the other hand ended up being a solid surface...

So arccos(x) is like 'the shape of the surface of things'  and 'arccos(x) is the shape of 'no-thing', or the space outside of a surface'  That space is itself a continuous surface, but also has infinite holes in it.

So at this point, I can also realize that there are straight lines on 'Sine Waves' graph which are helical lines.... in pretty much every direction; and I begin to wonder to myself if spin is really the all there is, and that linear motion is emergent from just spins.   I still sort of think that the 'inertia' or velocity vector might just be a computed value that doesn't have a 'real' part to play... that any change to linear motion is actually a change in the helical spin paths of all the spinning things making up the body...  You really can't instantaneously move from forward to backward, you can either slow down and then reverse, or curve around a path (perhaps at a greater speed than you could stop and then reverse, batmobile grapple hook turn?)

I've sort of settled on, in the meantime, keeping the values separate, so in my new 3D phsyics system, I have time, spin and velocity as my primary dimensions.

So then, I'm getting around to initializing this frame-thing, and am looking at other ratios, such a elliptic cuves would have two scalars instead of 1 common scalar... so there's a ratio between spin and linear speed available.. it would appear that matter would be like 1/2 spin and 1/2 linear component, where the spin is still at 1x the speed of light... a photon (as previously mentioned) has only a linear component and 0 spin, the other thing would have 0 linear component and 1 spin; this would seem like a place black holes fit in.... Especially since they seem to 'warp space' such that although they are incredibly massive, they're really not... Black holes emit very high energy gamma particles - neutrons are high energy gamma particles... would seem black holes take energy from ambient energy that's otherwise lost (including the heat phonons in their area) and make matter, and is 'endothermic' to the universe; where suns on the other side are 'exothermic'. 

So that leads one to wonder, We (as humans) are really good at taking things and breaking them and getting energy out... we're not so good about the other side; that is taking some energy and reforumulating matter.

We know breaking things apart (fission) releases energy... so obviously the other process should absorb energy.

---

Is there any documented sciences which are endothermic?  Yes, Geet Reactor, electromagnetic devices from John Bedini, The SEG(which, this theory validates most), ...

---

I'm kind of wander all over at this point, and I do really want to get to the point.

1) process A reacts and generates heat, that difference of heat vs the ambient cold can be used to do useful work.
2) process B reacts and consumes heat, the process itself creates more 'power' by increased current at the same voltage.

process A destroys matter(generates heat), and in the process can do other useful work.
process B creates matter(consumes heat), and in the process can do other useful work.

It would seem to me, that there's a complete energy loop available that you can 'do work' on both sides of the conversion.

This is like part 4 of a series... this one though really bring in a lot of other things, and my require some of the other videos...
https://www.youtube.com/watch?v=9QLJQHqu_KY

---
Let's presume that the 'previous civilation' had  'clean' energy, and it was all endothermic, and non-explosive and 'in harmony with nature' but that sort of thing will lead to 'global-freezing' ... they probably blamed it on the orbit of the earth changing so the earth was further from the sun, not that they were using up all the heat producing matter (electrons even, not even nesscarily full protons).  Our current situtation is of course the opposite, we use only exothermic energy (well, except wind/solar, which also absorb entropy/heat from their systems) which is leading to global warming.

---
14Billion years... that's a really small number; feeling nearly as ridiculous to me as 'we were created 6000' years ago.
The current theory says that 'photons last FOREVER'... in an infinite universe (or at least one that has existed more than 14B years) we would be blinded by all of the photons hitting us from infinite points... but they have applied 'oh well, space stretches so they lose energy' ... that idea is a skew of the value between time and space, when really it could all be projected on a time axis, and say 'oh,photons only say linear and energetic for 14Billion years, then they scatter and become part of the background heat of the system.

Measuring the pressure of a casmir cavity probably is very similar to what our CMBR would look like 14B years from now 14B light years away.

But now I'm way WAY off the map, "WHAT? Space doesn't warp!?  Blashpemy, of course we are the most important thing in the universe, it was created 14Billion years ago!"

----

I can find evidence and references for everything I've stated... but I don't really know where to go from here either....


At this point, I'm 'in a class of my own' or rather 'I seem to be the only one involved in this class, and the teacher doesn't know jack shit' (the teacher of the class is also it's only student)

[d3x0r]

Suduko is an interesting problem space... you know; 3x3 grids of numbers 1-9 and then a grid of those cells...
the possible combinations is fairly large, but the possible combinations that are 'valid solutions' is  pretty small...
you can throw random things together for along time and not get anything, and in fact the hardest suduko is the one that's blank.


===


I realize I've just kind of jumped into the middle of my train of thought an threw out a few highlights, I think I got the basic idea across, but probably that was lost in the details... well that and I'm not entirely sure 'The Idea' ?  In this case... can we take a exothermic work producing system, and a endothermic work producing system, and have a complete system?  It appears the universe does... but it also predicates dismissing many preconceived notions, to replace them with different ideas


It doesn't change the math - like 1+1 isn't suddenly 3, projections and all the standard things still work, and produce the same numbers (maybe with scalars in different places, and/or different units because the decay of light is just change in time (dT) and not (dT/dS) change in time per change in space.


There is no coodinate system for rotations.  Spherical trig fails at the point that it only represents the rotation with 2 angles; this is enough to get you to any point on a globe (say the earth) and either mark that point, or assume that from that point and 'up' is you standing at that point.  To also include the direction you're facing there has to be another angle. So Spherical trig is really only a 2d projection of the 3d rotation needed, and there's nothing filling that space.


There's Hopf Fibrations, which are basically a way of analysing the function of what the rotations would look like, but it's not just a direct interpretation... the hopf fibration demonstrations that are also found on youtube are renders that took time to compute what that point in space should look like.... And this rotation system I've implemented just Does that; like I pick (x/y/z) rotatate it around (xR1/yR2/zR2) by an angle, and rotate that around (xR2,yR2,zR2) and just 'get the answer'.


There's a quantum uncertainty problem related with the intersection of 3 cones... which is another way of projecting rotations into 1 d projections... and considering the full scale of angle as the diameter of the cone... which is great an all, but didn't actually tell me 'how to do it' ... so now I do it, and not even thinking about it ended up with a simpler system than the standard matrix multiplications and quaternion multiplications that normally happen (and fail, a rotation matrix, only represents 1/2 of a rotation, from -1 to 1... it doesn't account for passing 1, and going to 1 in the other direction - it's assumed that it's a linear bounce... so in the long run it doesn't matter, and it's true, that's really all you need in order to apply the curvature to a point is that principle rotation, but if you continue to work with just principle rotations, you lose the full rotation value.


A series of 5 robot arm segments with motors all in the same direction... each motor can go from -90 to +90.  If all of them are at +60, the total rotation of the hand on the end of the arm is 300 degrees around a circle (almost reaching back to the base).  The rotation matrix simplifies +300 to -60 because that's the same direction.  So when you turn off the thing and turn it back on, the current position is -60, so it sends an update to the motors to go to -10 on each, so suddenly all motors rotate from +60 to -10... Quatnerions represent only half of the rotation space also, because normally we work with 'unit quaternions' which are quaternions that are '1'.  but there's also unit quaternions that are '-1' ... but those are disregarded....   From the angle-angle-angle space, for any rotation anywhere, I can get a valid rotation matrix (basis frame) or a quaternion (which then gets applied to points using a basis frame anyway) .   The instantaneous work to do the conversion is cheap, the information required is 4 coodinates for a quaternion, and 9 or 16 for a matrix.  Rotations take 3.  Linear motion takes 3. 


I could do 2D or 1D rotations by just ignoring one of the axles... (axles are axises, but things spin around them, the X axle and the X axis are the same direction in space).

Anyhow these are technical justifications that don't really matter... the point is , simpler, smaller, faster, need to compute less frequently, all better.  And Me, I'm one guy, who knows noone, with a simple equation for a rotation space (hey, since Reimann got a space named after him, suppose I get a space named after me?)  ... that everyone needs to know, to invest as much time in the real way things work as they have with the reimannian space, and we'll be good to go for a few years... From a math community standpoint; it's crazy that 'all math isn't just known' and that we actually had to discover math as we went... now maths can even be created

---

Extended arccos - that is another thing that I'd love for every computer math library and implementation to just support; that'd be nice...
I think I could even live with arcsin(x) = arccos(x) +pi/2 or -3pi/2   .. either way sin(arcsin(x)) = a good X...
but the way arcsin genuinely works is that it's the opposite slope of arccos, and is not just a parallel offset.


https://www.geogebra.org/m/na2qdbn4

https://www.geogebra.org/m/e98czhgc

(if two things behave the same, are they the same?) 
So I have these mobial wound coils, following patterns that this spin geometry has reinforced, and I find effect, and then I find equivalence, and I doubt I can identify unique effect....


Like I have this coil, (two coils) would approximately the same, I put a ferrite around the core of one, and get X inductance, I put it around the other I get X*2 inductance, for 3 seprated ring ferrites...
I wind a coil on A ferrite ring, (mind you this was previously never broken ring) with a loose enough geometry to change the windings the same way, and I get X inductance.
I wind a coil, around a ferite, and get 0.5mH of indutance (for about 12 turns)  I wind an air core coil with 24 turns (48 really, has to be in series bifilar), and get 0.5mH of inductance...
so really looks like the ferrite isn't worth the bother of winding around it....(at least that's where I went the first time through this... it really seemed that ferrite within the toroid of a mobial torus was basically ineffective, because it wasn't a complete path between the inside and outside
In fact, a cuff around the toroid - passing through the center, instead of around the center - is a better place to use ferrite...  so adding it after the fact was even more affirmation it didn't matter...
but, now looking back there IS some effect... like 0.5mH from a few turns on a life-saver(somewhat bigger) ferrite is pretty high inductance for like a foot of wire...  (these ferites worked really good for joule theives, (which can be wound to be bucking.... instead of bifilar) )


----

[d3x0r]

https://phys.org/news/2020-08-black-silicon-photodetector-efficiency-limit.html[/size]


1 photon creates 1.3 electrons... '130% efficient' though I doubt they measured the loss of heat to the panel... a really good solar panel shouldn't get hot.

[sm0ky2]

When doing tests of this nature it is best to build multiple prototypes


Carefully winding each one, with brand new wire.
bending the same piece of wire unwinding rewinding
you change the inductance of the wire by stretching it
coherency is decreased, and there are small changes in diameter
also the insulation is often weakened or compromised
which can result in shorting of a winding


also variances in ferrite construction, even brand of wire when different copper qualities are used.


you are correct in your observations, a difield coil cancels its own external field.
There is not enough external  flux to polarize the ceramic ferrite
Electrically it is more efficient, magnetically it is less efficient.
The energy trade off is equivalent


As an inductance used in a circuit this is the way to go


If the goal is inductance in the magnetic domain, you would want a unified field

[d3x0r]

the circuit I used was a mazilli/royer - the drive coil is split with a central power source, which was behind an inductance of its own; a good amount of inductance is good there, there is a point where too much and it doesn't really support the power levels you're feeding it.


THe tests were just signal->current coil-> (and then a split)  to the test coil, which goes to a cap to ground (limiting the amount of current and getting the most pulse)
      and on the other side of the split - to load - but that didn't really demonstrate anything since the flat solenoidal - just gathered into a ring but definitely don't cross through a magnet loop... which reminds me - that video just came up.


https://www.youtube.com/watch?v=nGQbA2jwkWI ( here he uses a plastic bag to show 'the surface of a field from a loop of wire'  which the mobial coils violate. )


https://www.youtube.com/watch?v=X4dXXnUMHbQ (just a review of classical magnetism)