Continuing on with previous research and countless of simulation iteration to try and find its breaking point but the results keep showing something that should not be possible. So I thought I would throw it in the community and see if it does have breaking points.
Well after all those battles with FEMM. And the secret was? A combination of an air gap and a coil, the latter will not be affected by the magnets due to the former. This synergy does very cool things. There is near zero voltage induced in the coil when the shown magnets move in and out of the region of the coil, why? Well Because of the air gap. The core near that part is not affected, most of the "flux" goes around the path of least resistance. We have no induced EMF HOWEVER when we apply a current to the coil ourselves this DOES affect the force/torque on the motor. In fact it has a quite large effect on the order of 5-10 Newtons which is quite significant torque wise. However the coil remains purely inductive when powered, its inductance is barely affected by the effect of the magnets as FEMM also shows.
How does it work? Well you can imagine the magnet assembly is rotating around a drum. When it comes close the coil energizes and causes a stronger pull force without affecting its inductance. And then when it flips polarity it gives the magnets a weaker pull back force AGAIN without affecting its inductance or current flow.
It also seems to love high currents as well. The more current is even more crazy force differentials.
The design is btw not limited to the shown toroidal design either. Attached is an example of what would typically be called a gapped C core arrangement.
As you can see whenever the coil over the airgap is energized the force is changed however its barely changes inductance. The force can be quite significant the higher the current. Shown is a 4N force differential from 10A of current however at 20A this force diff becomes 25N, almost 5x times greater!
But since I am working with 200 turns I wanted to keep the current at a realistic level that a coil with that many turns could handle continuously preferably.
The left coil has even more interesting implications especially if you use IT to drive the current in the air gapped coil ;). All simulations were ran multiple times on higher mesh refinement steps to exclude any simulation accuracy anomalies, the results however persist.
I hope anyone else understands what is going on here.
Quote from: broli on 2024.02.17, 09:54:21
The design is btw not limited to the shown toroidal design either. Attached is an example of what would typically be called a gapped C core arrangement.
As you can see whenever the coil over the airgap is energized the force is changed however its barely changes inductance. The force can be quite significant the higher the current. Shown is a 4N force differential from 10A of current however at 20A this force diff becomes 25N, almost 5x times greater!
But since I am working with 200 turns I wanted to keep the current at a realistic level that a coil with that many turns could handle continuously preferably.
The left coil has even more interesting implications especially if you use IT to drive the current in the air gapped coil ;). All simulations were ran multiple times on higher mesh refinement steps to exclude any simulation accuracy anomalies, the results however persist.
I hope anyone else understands what is going on here.
Broli,
Most interesting simulation results! I'm sorry but what kind of drive is "IT" you mention above?
Also, the current you are producing in the winding over the gap, is it generated by a constant voltage or constant current source?
Regards,
Pm
Quote from: partzman on 2024.02.17, 14:22:37
Broli,
Most interesting simulation results! I'm sorry but what kind of drive is "IT" you mention above?
Also, the current you are producing in the winding over the gap, is it generated by a constant voltage or constant current source?
Regards,
Pm
Hi partzman, thanks for the interest.
Sorry I meant the literal word "it" by referencing the left coil in the simulation. As for the current, currently its generated in a static way. Essentially I tell the coil how much current flows through it and I measure the forces. I do this for all kind of variations, positive, negative currents, position changes of the magnet assembly and variations in the core material (linear and non linear).
If we only used the "regular" coil on the left in the previous design, we would see a typical motor coil/core behavior. Where a back EMF is induced and impedes the applied the current. However this is not the case with the right coil. And what I was hinting at is that we can even take advantage of both attributes by hooking them up in series, you now have a coil inducing EMF and pushing current through itself AND the air gap coil. BUT this only seems to work if the winding ratio of the air gap coil is larger than the "normal" coil. I found a 1:10 or even 1:5 to be adequate. So for example the airgap coil would consist of 200 turns whereas the "regular" motor coil would consist of 20 turns.
But that is already getting into more controversial territory where the thing powers itself. Having a motor with little to no perceivable induced EMF (by only using the air gap coil) is a good start and quite a strange result to see in FEMM no matter how I change the design and problem definition. Whether reality has a different say I dont know but I have yet to come across this design in any OU community.
Hi Broli,
Since FEMM is based on the equations of physics, in particular Maxwell's, whose mathematical formalism guarantees the conservation of energy, any OU that appears is either due to bugs, calculation uncertainty or unrealistic parameterization. This is true whatever the simulation software (I've had OU as an artifact of ltspice or Working Model 2D).
That said, this may be a sign that the system has very few losses and is therefore prone to runaway at the slightest parameter that might be at the limit of the calculation uncertainties. It may therefore be worthwhile pursuing this in reality, as the high sensitivity of the set-up could then reveal a real anomaly, such as a magnetic flux conservativity default.
https://youtube.com/shorts/jeziPoJia6w?feature=share (https://youtube.com/shorts/jeziPoJia6w?feature=share)
Does it can be cause increased oscillation of my vibrator ? This also has coil,magnet,gap.
Quote from: F6FLT on 2024.02.18, 17:56:55
Hi Broli,
Since FEMM is based on the equations of physics, in particular Maxwell's, whose mathematical formalism guarantees the conservation of energy, any OU that appears is either due to bugs, calculation uncertainty or unrealistic parameterization. This is true whatever the simulation software (I've had OU as an artifact of ltspice or Working Model 2D).
That said, this may be a sign that the system has very few losses and is therefore prone to runaway at the slightest parameter that might be at the limit of the calculation uncertainties. It may therefore be worthwhile pursuing this in reality, as the high sensitivity of the set-up could then reveal a real anomaly, such as a magnetic flux conservativity default.
EDIT: Retracted original post.
Quote from: broli on 2024.02.17, 09:54:21
The design is btw not limited to the shown toroidal design either. Attached is an example of what would typically be called a gapped C core arrangement.
As you can see whenever the coil over the airgap is energized the force is changed however its barely changes inductance. The force can be quite significant the higher the current. Shown is a 4N force differential from 10A of current however at 20A this force diff becomes 25N, almost 5x times greater!
But since I am working with 200 turns I wanted to keep the current at a realistic level that a coil with that many turns could handle continuously preferably.
The left coil has even more interesting implications especially if you use IT to drive the current in the air gapped coil ;). All simulations were ran multiple times on higher mesh refinement steps to exclude any simulation accuracy anomalies, the results however persist.
I hope anyone else understands what is going on here.
Hi Broli,
Quite Interesting - Thanks for sharing your discoveries!
SL
Hi Broli,
You have in FEMM the ability to derive a force v. distance profile for your rotor movement in the x y plane hence obain the mechanical energy gained or lost over that movement. You also have the ability to derive the electrical input energy gained or lost during that movement using the flux linkage profile over the movement. The flux linkage change between steps is N times the integral of the voltage, hence when multiplied by the coil current it yields the energy transfer at each step. Creating the profiles is quite easy using Lua to create the stepping program and to output the data into a text file. As you have not quoted any actual COP results I assume you have not done this. If you need assistance in this I can help as I have done many Lua runs, and I must say that so far FEMM has not given me overunity.
Smudge
Quote from: Smudge on 2024.02.20, 20:09:10
Hi Broli,
You have in FEMM the ability to derive a force v. distance profile for your rotor movement in the x y plane hence obain the mechanical energy gained or lost over that movement. You also have the ability to derive the electrical input energy gained or lost during that movement using the flux linkage profile over the movement. The flux linkage change between steps is N times the integral of the voltage, hence when multiplied by the coil current it yields the energy transfer at each step. Creating the profiles is quite easy using Lua to create the stepping program and to output the data into a text file. As you have not quoted any actual COP results I assume you have not done this. If you need assistance in this I can help as I have done many Lua runs, and I must say that so far FEMM has not given me overunity.
Smudge
Dear Smudge,
I am aware of the LUA scripting in FEMM however honestly lately my time is going fully into my job and have little time and energy to do this kind of testing. So if you want to assist I would be very grateful for that! I attached the latest design I am playing with that is showing very asymmetrical forces. I completely got rid of the air gap as honestly it was not changing that much. Now its akin to a completely closed toroidal core fully wrapped with a coil.
I have done a very crude and linear analysis of what you asked and it shows a
COP of 4.5. And honestly I have given the inductive energy way more energy than it really contains. In fact if I compare the magnetic energy of the core before and after the movement I get a much smaller difference. And this would make sense as the core I chose has a non-linear permeability. At high currents the magnetic energy of the core no longer changes much however the force keeps getting significantly asymmetrical. I know the force calculation is also not ideal and linear but if I only consider the actual magnetic energy of the whole magnetic region I am seeing a
COP of 30-40.
As you said this needs a finer step by step analysis using scripting and perhaps at different currents to find the peak COP current as going too high has diminishing returns. Using flux linkage would also not be accurate and the total magnetic energy of the core region should be compared instead for a more accurate analysis.
I have attached and example of this of using 20A. And a small ball park calculation assuming even the inductance is linear with current (which it is not) and you can see there is a big difference in apparent energy. I also have attached a zip file with the FEMM file. I am eager to see the analysis, thank you.
I have a hunch that operating at near the saturation current is key.
For those interested. Attached is a setup which could be Lua scripted. I tried to show in a limited way how the impact of current, force and energy in the (non linear) core have a unique interplay.
The calculations should iterate over a current range and displacement range to find the highest energy deltas. However even at 100A you get a reduction of almost 500N of the original 740N while the energy of the core goes down much less ratio wise. Around 3x for the force and 1.25x times for the energy in the core. And this is not even measuring the actual Work done by the force.
I dont think the core matters much as long as its non-linear which is any ferromagnetic core is in the real world. Even core losses are not that relevant! Because you dont have to flip the domains 180 degrees. According to research core losses and remanence ar much smaller if the domains rotate with a field rather than having to flip 180 degrees aburptly. I tried to illustrate this below. The domains align due the force of the permanent magnet, however if you apply a current coil you will torque them away from this position. The overal field barely changes because it is ONLY these affected domains near the magnet that are changed enough to reduce the mechanical force on the magnet. Yet at a much smaller magnetic energy cost. The argument here is that the applied field due to the magnet is at right angle with the current field. And second due to the relatively small "effect" area where the magnet acts on the relative inductance or rather magnetic energy change will be very small. This effect area essentially depends on the size of the magnets used and is where the domains are being torqued away from the magnet which costs little magnetic energy.
I hope someone can help with extracting the step wise data of this. I attached a cleaned up version of the previous FEMM file to avoid confusion. It contains a single magnet, the previous double magnet was mainly to demonstrate that the "total flux" change can be 0 due to the movement of the magnets alone if you used one on each side. Do note that current in the opposite direction does not give the same symmetrical force result either! This means that we could drive it with an AC tank rather than a PWM that switches the coil on and off at the right moment.
Broli,
I have had a brief look at the FEMM file you put in your previous post. There are three things that affect the COP figure you derived, two of them you probably expect because they relate to the non-linearities and the third is an error of omission. I will demonstrate these in a graphical way but that will take time as my wife and I have just moved house and we are in a bit of chaos at the moment. I am 90 in April and my wife is 89, we both have health problems and we have moved into retirement apartments that have extra care facitities. Here is a summary of my findings.
1. Your magnetic energy derivations using 1/2Li2 assume the core is linear. Your flux data for 5A, 20A and 40A coil current indicate the significant non-linearity, saturation occurs at around 3A current. Thus your input energy being the difference of the two magnetic energies in in error by a factor close to 2.
2. You have omitted to take account of the flux change during magnet movement that induces voltage across the constant current generator hence demanding input energy, and that increases the error by another factor of 2, so we now have a factor of 4.
3. Your average of the forces at each end of the magnet movement yields an energy output that assumes force v. distance is linear, and we know that it is not. That introduces another error that overstates the output energy. The net result is your crude analysis giving a COP of 4.5 now reduces to a COP near unity. I am quite sure that a detailed attempt using say ten 1mm steps will result in that unity COP.
I will write this up with some charts showing the non-linearities that demonstrate how the energies are derived.
Smudge
Welcome broli.
I read your study with interest.
I really like the results! The results are similar for some machines. Of course this is just my opinion.
I am not familiar with the spiritual world of the software you have just shared with us. So I am doubtful, but I am still interested.
I mean. How can this be a practical guide?
I don't understand exactly if this is a motor concept or a static permanent magnet generator? Or can it be applied to both?
Well. If we consider the phenomenon in terms of mechanical force, then it can be supported. (Unfortunately, for other reasons, I cannot present it now. Sorry.)
In my opinion there are several independent experiments (but there is no excess energy, only energy utilization)
Or there is an implementation similar to your idea. But not in the same way.
It is completely different, but let's think about the operations.From the third minute.
https://www.youtube.com/watch?v=J61m6YY-2sY
A different idea.
In my opinion, your idea is similar here, based on the parameters given in the FEMM simulations. Of course, this is also complementary. Or the Bulgarian m.e.g. transformer.
Here we can't talk about forces only variables.I underestimate.Without air gap no result.In this state there is no OU yet.Only simple transformation.
I apologize if my opinion is irrelevant.
Sincerely, Atti.
....
Quote from: Smudge on 2024.02.26, 15:54:24
Broli,
I have had a brief look at the FEMM file you put in your previous post. There are three things that affect the COP figure you derived, two of them you probably expect because they relate to the non-linearities and the third is an error of omission. I will demonstrate these in a graphical way but that will take time as my wife and I have just moved house and we are in a bit of chaos at the moment. I am 90 in April and my wife is 89, we both have health problems and we have moved into retirement apartments that have extra care facitities. Here is a summary of my findings.
1. Your magnetic energy derivations using 1/2Li2 assume the core is linear. Your flux data for 5A, 20A and 40A coil current indicate the significant non-linearity, saturation occurs at around 3A current. Thus your input energy being the difference of the two magnetic energies in in error by a factor close to 2.
2. You have omitted to take account of the flux change during magnet movement that induces voltage across the constant current generator hence demanding input energy, and that increases the error by another factor of 2, so we now have a factor of 4.
3. Your average of the forces at each end of the magnet movement yields an energy output that assumes force v. distance is linear, and we know that it is not. That introduces another error that overstates the output energy. The net result is your crude analysis giving a COP of 4.5 now reduces to a COP near unity. I am quite sure that a detailed attempt using say ten 1mm steps will result in that unity COP.
I will write this up with some charts showing the non-linearities that demonstrate how the energies are derived.
Smudge
Hi smudge thanks for your feedback. I am amazed about your tenacity as an elderly gentleman. I would like to not let age come inbetween intellectual discord but I have great respect for your insights and ideas. I often see a lot overlap with ideas that keep my own mind busy.
As for your comments. I mostly agree. However the inductive energy would actually be overestimated if you considered it to be linear like I did. As for the force integral I assumed a linear degradation from the max to min force. Which is not ideal I know and needs to be scripted in tiny steps
I am not aware of handling changing voltages in FEMM. I kept it simple by assuming a constant current source over the duration of the magnets movement. However what I do want to highlight is when I instead select the entire core and calculate its energy In femm before and after the magnets moved, according to the literature this should be equivalent to calculating the changed inductive energy. In fact it would be much more accurate AND lower than considering the core to be linear like I did in the table posted earlier.
What I want to stress is the fact that not only does there "appear" to be more mechanical energy than magnetic, the actual reaction of the coil is such that the changing "field" it sees and its reaction to it actually help to reduce the force further rather than increase it as conventional motors do.
I just did the analysis for a single 1mm step instead of the previous 10mm step, attached is the result. Its interesting to note that the same 30x factor also arrises in this analysis. I tried to add numbering so the analysis is easier to follow. This time I used the magnetic energy calculator in FEMM and selected the enclosed core area. This gives a much better estimate, and is essentially a free value FEMM gives you, than using flux linkage as the inductance which is non linear above saturation currents.
There are four interesting aspects to this:
- The magnetic energy in the core is only reduced slightly. Aka the input energy.
- It seems to like higher currents due to saturation. There probably is a peak after which you get diminishing returns. (Must be parametrized with scripting)
- The potential induced "back emf" of the coil would act in in the APPLIED constant current direction further reducing the hold force on the magents as they move away from the core at a reduced force then they approached it. The coil current must act WITH the field of the magnet to reduce its pull force. How is no one talking about this effect alone?
- And then the ridiculous COP.
To me this seems like gently torquing the magnetic domains back and forth is much better than trying to fight their fields head on. The effect in the proposed design acts 90° against the field and thus will try to gently torque the domains not flip them which costs much more energy and work.
But exist other programs for similar modeling. Why don't try to do it by any other programs ?
Solarlab has been advertising here for a long time, for example ANSyS. O0
Quote from: broli on 2024.02.27, 12:48:09
As for your comments. I mostly agree. However the inductive energy would actually be overestimated if you considered it to be linear like I did.
I have to disagree with you there. The inductive energy 1/2Li
2 is also given by 1/2Phi*i where Phi is the flux linkage. In the plot of Phi v. current this energy is given by the area between the line and the Phi axis as shown below. It is seen as a triangle whose area is 1/2 base*height and that accounts for the 1/2 factor. Now look at the next image for a saturated core. Clearly the linear case is an overestimate as you say. But this is misleading when you take the enregy differences between the 0mm case and the 10mm case. The third image shows this difference for the linear case and the final image for the non-linear case. Now the linear version is an underestimate.
With regard to voltage, flux change is equal to the time-integral of the voltage hence fluxchange*constantcurrent yields energy, and this makes sense as the voltage determines the power taken from the current generator. This can also be presented as a rectangle on the Phi v current chart and has an energy value that is twice the incorrect value derived from the linear inductances. I have never used the FEMM magnetic energy facility so I must read up on that.
Smudge
Broli,
Having looked into the FEMM magnetic energy integral it is obvious to me that it does not necessarilty tell you the electrical input energy. You can have magnetic energy in the core put there from the nearby magnet, you can then alter that energy by applying current through a coil. That current can reduce the core energy but you actually supply energy to do it. In the limit for certain geometries you can drive the core energy to zero. So where has the input energy from the coil gone? The answer must be somewhere else and that is outside the core where the field is not considered in your treatment. As I see it using the FEMM energy facility is a huge mistake, the only sensible method is consideration of the field change seen by the coil yielding voltage*time that then gives energy from the constant current source.
Smudge
Quote from: Smudge on 2024.02.27, 20:46:38
Broli,
Having looked into the FEMM magnetic energy integral it is obvious to me that it does not necessarilty tell you the electrical input energy. You can have magnetic energy in the core put there from the nearby magnet, you can then alter that energy by applying current through a coil. That current can reduce the core energy but you actually supply energy to do it. In the limit for certain geometries you can drive the core energy to zero. So where has the input energy from the coil gone? The answer must be somewhere else and that is outside the core where the field is not considered in your treatment. As I see it using the FEMM energy facility is a huge mistake, the only sensible method is consideration of the field change seen by the coil yielding voltage*time that then gives energy from the constant current source.
Smudge
There shouldn't be anything wrong with using the magnetic energy and is often used in FEMM applications to measure total energy and to converge the simulation on. So its an important factor in EM simulators. Care should be taken when doing so over free space. But ours is bound by the core which should make it more accurate too.
This is what I am showing in the previous post. The top illustration shows the total energy in the core due to the magnets only. This energy due to the magnet alone is significantly lower in the core at 0 amps compared to 50A. Meaning most of the energy is due to the coil current which indeed has a cost. This cost can be easily calculated by measuring the core energy before and after moving the magnet at a constant current and should in theory equate to the inductive energy change of the coil. This eliminates the time dependent nature of changing voltage/flux linkage. But even taking THAT into account we are over unity.
I encourage anyone to jump in and validate. My own time is currently limited and cannot learn something new like Ansys but I welcome anyone else and if needed I can share the dfx file. I believe Ansys does have a transient solver. But the static solver like femm should do to compare results.
Quote from: broli on 2024.02.27, 20:51:36
There shouldn't be anything wrong with using the magnetic energy and is often used in FEMM applications to measure total energy and to converge the simulation on. So its an important factor in EM simulators. Care should be taken when doing so over free space. But ours is bound by the core which should make it more accurate too.
I think you have missed the point I was making. FEMM correctly calculates the BH product as an energy density that when integrated over the volume of the material yields the magnetic energy in the material. You start with the energy put there by the magnet. Additional energy from the coil can either increase or decrease that energy. You then move the magnet to a new position and the magnetic energy has changed. Removing the coil current now gives another value of magnetic energy. You are of the opinion that you can use those values to obtain the difference in the initial coil supplied energy and that which is regained when the current is swicthed off. I think that is a false assumption because the coil supplied energy influences magnetic energy in regions other then the core. IMO you have to evaluate the change of magnetic energy not only within the core but also within the air space outside the core, within the copper and within the magnet. The magnet energy will always return a negative answer and that introduces another intellectual challenge in deciding whether a reduction in negative energy there yields an answer that really is a positive supply of energy. Using just the core energy will not tell you the electrical energy input.
QuoteThis is what I am showing in the previous post. The top illustration shows the total energy in the core due to the magnets only. This energy due to the magnet alone is significantly lower in the core at 0 amps compared to 50A. Meaning most of the energy is due to the coil current which indeed has a cost. This cost can be easily calculated by measuring the core energy before and after moving the magnet at a constant current and should in theory equate to the inductive energy change of the coil. This eliminates the time dependent nature of changing voltage/flux linkage. But even taking THAT into account we are over unity.
I disagree, I think that method will give wrong answers leading to wrong COP values.
Smudge
Quote from: Smudge on 2024.02.28, 11:07:41
I think you have missed the point I was making. FEMM correctly calculates the BH product as an energy density that when integrated over the volume of the material yields the magnetic energy in the material. You start with the energy put there by the magnet. Additional energy from the coil can either increase or decrease that energy. You then move the magnet to a new position and the magnetic energy has changed. Removing the coil current now gives another value of magnetic energy. You are of the opinion that you can use those values to obtain the difference in the initial coil supplied energy and that which is regained when the current is swicthed off. I think that is a false assumption because the coil supplied energy influences magnetic energy in regions other then the core. IMO you have to evaluate the change of magnetic energy not only within the core but also within the air space outside the core, within the copper and within the magnet. The magnet energy will always return a negative answer and that introduces another intellectual challenge in deciding whether a reduction in negative energy there yields an answer that really is a positive supply of energy. Using just the core energy will not tell you the electrical energy input.
I disagree, I think that method will give wrong answers leading to wrong COP values.
Smudge
As I said previously I am open to anyone doing their own calculations or analysis hence the reason I shared the files. I am eager to see what your proposed method will show.
OK, here is a quick look at your single magnet movement of 1mm for comparison with your data. The image below shows flux linkage v. current at the two magnet positions. I initially did 10A steps but then added the 5A points so as to use Simpson's rule for the integrations. I get input electrical energy when the coil is energized as 0.7383 joules and returned energy when the current is switched off as 0.7146 joules, a difference of 0.0237 joules. To this must be added the energy taken from the 50A current becaause of the flux change applying voltage to the current source during the movement, that calculates to 0.12065 joules. Thus total electrical input is 0.1443 joules, to be compared with the mechanical output of 0.1455 joules. The COP is 1.008.
Smudge
Edit. added the chart
Quote from: Smudge on 2024.02.29, 11:25:59
Thus total electrical input is 0.1443 joules, to be compared with the mechanical output of 0.1455 joules. The COP is 1.008.
Smudge
It works out,that perpetuum mobile does not exist ? Nobody never will be able building it. :-[ :'(
Quote from: Smudge on 2024.02.29, 11:25:59
OK, here is a quick look at your single magnet movement of 1mm for comparison with your data. The image below shows flux linkage v. current at the two magnet positions. I initially did 10A steps but then added the 5A points so as to use Simpson's rule for the integrations. I get input electrical energy when the coil is energized as 0.7383 joules and returned energy when the current is switched off as 0.7146 joules, a difference of 0.0237 joules. To this must be added the energy taken from the 50A current becaause of the flux change applying voltage to the current source during the movement, that calculates to 0.12065 joules. Thus total electrical input is 0.1443 joules, to be compared with the mechanical output of 0.1455 joules. The COP is 1.008.
Smudge
Edit. added the chart
Hi smudge and thank you! I really appreciate constructive criticism like this rather than "You are wrong."
What I find strange is how using magnetic energy over the elements gives such a different results. Here is a FEMM example where they talk about the different way to calculate inductance and how they are pretty much similar. One utilizes the magnetic energy while the other would then be your current method:
https://www.femm.info/wiki/InductanceExample
And as they state the former is indeed more accurate but the latter should be plenty accurate when most of the energy is confined and not in open air.
Now I have two remarks for you.
- 1) I find it very peculiar that your calculated energy values are in the ball park of 2x of mine? In the attached sim images I changed the magnets to be air to eliminate their contribution. I then Energized the coil to 50A and calculated the magnetic energy in the region both with and without the surrounding air. As you see the difference does not account for the 2x. I find it strange that FEMM itself demonstrates that both of these values can yield the same results when used in the correct situation yet ours are very different.
- 2) Then what is also important in your analysis is to consider how this summation calculation of energy behaves depending on mesh refinement. Something I often do as a sanity check to validate the values on super fine meshes. where I do most of the quick checks with medium fine meshes and then the final with a superfine refinement that takes much longer to process but gives more accurate results. Does your energy delta change much with finer meshing? If so then this should be considered as an stastical error and accounted for especially when dealing with summations where errors tend to add up.
To expand on the latter point, I believe for flux linkage, FEMM is using a clever contouring technique to calculate the area encased by the coil terminals. You even see such contours when you work with very refined meshes and calculate forces in such regions. It draws little red contours around the magnets for instance and uses them to calculates the force. The more "smooth" these lines (aka more triangles) the more accurate the calculated values. Thus in your analysis this "enclosing area" should be more refined to increase the accuracy of this value. Whereas right now the core region of the coil does not have that fine of a refinement as you see in the attached image.
Also as an another example I compared the flux linkage difference between the current mesh and a more refined mesh at 50A current (again with no magnets). And as you see attached the difference cannot be underestimated. And following the rule of error propagation in statistics, and by assuming that the error value (sigma) is the same over every sample point. This simplifies to sigma*squareroot(n) where n is the amount of sample points (current values) you took. This means the error value grows by the square root of the amount of sample points. Whereas using the magnetic energy you dont need to worry about such propagation as you dont need multiple summations to approximate the energy. You just get it in one shot and thus focus all the processing power on the final run. Thus to consider this cumulative error I suggest you do one run with the current mesh and then one with a super fine one and take the difference. This difference can be your "fixed" error and used as the error for any subsequent run with a more coarse mesh to improve simulation times. From this you can determine the total error of the final result by multiply it to the squareroot of the amount of samples taken.
Of course this is a simplistic approach. A more thorough one would be to do the entire analysis with a fine mesh and more sample points.
Can you perform this suggestion at perhaps 100 amps with 10 sample points and consider the errror value?
EDIT: Forgot to ask, did you move the magnets 1mm away or towards the core/coil in your analysis? As in the previously shared file the magnet was already moved away 1mm and thus the analysis should be done by comparing the current magnet's location vs 1mm TOWARDS the core.[/list]
Quote from: Smudge on 2024.02.29, 11:25:59
To this must be added the energy taken from the 50A current becaause of the flux change applying voltage to the current source during the movement, that calculates to 0.12065 joules.
Can you elaborate how you have come to this value please?
Quote from: broli on 2024.02.29, 22:43:01
Hi smudge and thank you! I really appreciate constructive criticism like this rather than "You are wrong."
And it is via these dialogues that we all learn. Your replies here help me understand where you are coming from and I will do my best to clarify my perception of the problem.
QuoteWhat I find strange is how using magnetic energy over the elements gives such a different results. Here is a FEMM example where they talk about the different way to calculate inductance and how they are pretty much similar. One utilizes the magnetic energy while the other would then be your current method:
https://www.femm.info/wiki/InductanceExample
And as they state the former is indeed more accurate but the latter should be plenty accurate when most of the energy is confined and not in open air.
Note that they clearly state the integration should encompass the whole area of the problem, whereas you were only doing the core.
QuoteNow I have two remarks for you.
- 1) I find it very peculiar that your calculated energy values are in the ball park of 2x of mine? In the attached sim images I changed the magnets to be air to eliminate their contribution. I then Energized the coil to 50A and calculated the magnetic energy in the region both with and without the surrounding air. As you see the difference does not account for the 2x. I find it strange that FEMM itself demonstrates that both of these values can yield the same results when used in the correct situation yet ours are very different.
When you eliminated the magnets the field is then almost entirely contained in the core, but there is some in the air as you can see some field contours. So you would expect your two values to be slightly different as you found. Your two values are to be expected, and to check this I modified your sim so that the coil is wound around the whole core, not just the RH limb. Then the field in the air disappears and the two values are virtually the same. With the magnets present the field in the air has contribution both from the magnets and from the coil current, so the air contribution is even higher.
Quote- 2) Then what is also important in your analysis is to consider how this summation calculation of energy behaves depending on mesh refinement. Something I often do as a sanity check to validate the values on super fine meshes. where I do most of the quick checks with medium fine meshes and then the final with a superfine refinement that takes much longer to process but gives more accurate results. Does your energy delta change much with finer meshing?
If so then this should be considered as an stastical error and accounted for especially when dealing with summations where errors tend to add up.
[/list]
I understand what you are implying but your findings with finer mesh showing that your mesh size is adequate will also apply in my case since the fields are the same.
QuoteTo expand on the latter I believe for flux linkage, FEMM is using a clever contouring technique to calculate the area encased by the coil terminals. You even see such contours when you work with very refined meshes and calculate forces in such regions. It draws little red contours around the magnets for instance and uses them to calculates the force. The more "smooth" these lines (aka more triangles) the more accurate the calculated values. Thus in your analysis this "enclosing area" should be more refined to increase the accuracy of this value. Whereas right now the core region of the coil does not have that fine of a refinement as you see in the attached image.
That "contouring technique" is the Maxwell stress tensor mask and only applies to the force and torque measurements. It is not used in the magnetic energy integration, that is done in the area you select where the BH product is evaluated for a finite number or area elements within the material boundary.
QuoteAlso as an another example I compared the flux linkage difference between the current mesh and a more refined mesh at 50A current (again with no magnets). And as you see attached the difference cannot be underestimated. And following the rule of error propagation in statistics, and by assuming that the error value (sigma) is the same over every sample point. This simplifies to sigma*squareroot(n) where n is the amount of sample points (current values) you took. This means the error value grows by the square root of the amount of sample points. Whereas using the magnetic energy you dont need to worry about such propagation as you dont need multiple summations to approximate the energy. You just get it in one shot and thus focus all the processing power on the final run. Whereas with flux linkage you need to consider the error of each samplepoint which depends on the the mesh refinement. Thus to consider this cumulative error I suggest you do one run with the current mesh and then one with a super fine one and take the difference. This difference can be your "fixed" error and used as the error for any subsequent run with a more coarse mesh to improve simulation times. From this you can determine the total error of the final result by multiply it to the squareroot of the amount of samples taken.
I see flux linkage presented to three significant figures as 0.0595, 0.0595 and 0.0595 for the three meshes you used. That tells me even the coarsest mesh is OK to use and your argument on the statistics does not apply.
QuoteOf course this is a simplistic approach. A more thorough one would be to do the entire analysis with a fine mesh and more sample points.
Can you perform this suggestion at perhaps 100 amps with 10 sample points and consider the errror value?
I can do the 100A but I don't see a significant error value in my method. Using Simpson's rule takes care of the small number of current data points.
QuoteEDIT: Forgot to ask, did you move the magnets 1mm away or towards the core/coil in your analysis? As in the previously shared file the magnet was already moved away 1mm and thus the analysis should be done by comparing the current magnet location vs 1mm TOWARDS the core.
I used the two magnet positions that you used as shown by the agreement of my force values with yours.
From your next post
QuoteCan you elaborate how you have come to this value please?
It is simple. The flux change that takes place during the 1mm magnet movement with the coil current at 50A was 0.03987 at the outer position minus 0.037457 at the inner position yielding a difference of 0.002413 that when multiplied by 50A gives 0.12065 Joules energy.
Smudge
Forgot to add to the previous post. On the subject of inductive energy it is acceptable to use the FEMM flux/current value (henries) as the inductance for linear materials. Clearly in this case we can't because the material is not linear. But is is acceptable to plot flux v. current and to use the area to the left of the curve as the input energy. The FEMM magnetic energy integration also can be used if the fields inside the core and outside the core do not contain fields from another source such as a permanant magnet. With PM fields there the the magnetic energy approach falls down because some of the field regions can have reduced energy when the coil is energized. It is quite possible for the magnetic energy approach to yield a total loss of magnetic energy when the coil is energized yet the current source supplies energy (example a ring core of material that has coercivity with PM energy in the core that is neutralized by current in a coil wound over the whole core). In that case the only method for determining the input energy from the coil is the one I use and I am surprised that this is not made clear in FEMM tutorials.
Smudge
Quote from: Smudge on 2024.03.01, 19:17:10
Forgot to add to the previous post. On the subject of inductive energy it is acceptable to use the FEMM flux/current value (henries) as the inductance for linear materials. Clearly in this case we can't because the material is not linear. But is is acceptable to plot flux v. current and to use the area to the left of the curve as the input energy. The FEMM magnetic energy integration also can be used if the fields inside the core and outside the core do not contain fields from another source such as a permanant magnet. With PM fields there the the magnetic energy approach falls down because some of the field regions can have reduced energy when the coil is energized. It is quite possible for the magnetic energy approach to yield a total loss of magnetic energy when the coil is energized yet the current source supplies energy (example a ring core of material that has coercivity with PM energy in the core that is neutralized by current in a coil wound over the whole core). In that case the only method for determining the input energy from the coil is the one I use and I am surprised that this is not made clear in FEMM tutorials.
Smudge
I understand what you are trying to say, when both a magnet and coil apply an opposite field then the coil would still hold energy but the core would appear to have 0 energy as the fields cancel. My argument is that this doesn't matter. As it should be enough to only look at the delta energy change of the system. Because assuming a super conductor where current keeps flowing, when you move the magnet out this current would indeed drop due to the change of circumstances, mainly the influence of the magnet on the domains changed. This change causes an energy delta in the system to occur. So even if your initial condition was such that the whole magnetic energy was 0 in your region. The delta will tell you how much it changed from its INITIAL condition. You seem to have added this complete Initial energy to your calculation and is confusing.
This initial energy can indeed be 0 or any value in the core region really. It is the overal magnetic energy change of the core (due to the movement of the magnet) that results in "flux change" and thus a reaction from the coil. This reaction and energy change should match exactly with the provided energy from the coil when considering a constant current source. The initial energy of the core with or without the magnet should be irrelevant as you pay for that once at the very start, and from there on only the delta matters as the system oscillates between these two states. I believe this is why I was confused about that added final value to the difference. You are overcompensation for the total input energy by also considering the "initial" energy to get the coil to 50A. The advantage of the Magnetic energy method is that you dont have to consider this. Because 1) you get most of it back when you power the coil down and 2) You only need to know the total magnetic energy change to know what changed for the coil energy wise. And thus its this delta that determines your total energy cost.
I also redid the analysis. I used a fully enclosed coil now as you did and a much finer mesh as you can see below. This time I also selected the copper coil region itself in the magnetic energy calculation. This would actually be an overestimation as this region also contains the magnet's field and doubles the previous values I had. But even with this better refinement, enclosed coil setup AND doubling of the magnetic energy values. The COP is still at 3x.
I believe the real debate here is, did I omit the initial energy of the coil or did you over account for it. This overcompensation would be akin to me adding the
total magnetic energy of the core when the magnets have moved to the final difference value of the magnetic energy. It wouldn't make sense as this energy can be recouped when powering down the coil. Only the delta is lost for good. So on that merrit my delta is even double that of yours and still get a COP of 3.
Since this calculation table also shows how the energy behaves without a magnet it is clearly evident that the magnetic energy or "flux" (a term I dont like to use) is reducing as the magnet moves away. As a consequence the coil will try to compensate by INCREASING the current. What other motor coil reacts in such a way to aid the magnets motion rather than oppose it?
I thought this was really interesting, https://hackaday.io/project/11865-3d-magnetic-field-scanner
Why guess with a simulator when we could gather real data and facts with a magnetic field scanner?.
AC
Quote from: Allcanadian on 2024.03.02, 16:14:30
I thought this was really interesting, https://hackaday.io/project/11865-3d-magnetic-field-scanner
Why guess with a simulator when we could gather real data and facts with a magnetic field scanner?.
AC
It will have trouble telling you what the field levels are inside materials, which is what we are dealing with here.
Smudge
Quote from: broli on 2024.03.01, 21:38:15
I understand what you are trying to say, when both a magnet and coil apply an opposite field then the coil would still hold energy but the core would appear to have 0 energy as the fields cancel.
It is not a case of the coil "holding" energy, in the case of a coil around the magnet cancelling the field in the magnet there is no energy for the coil to hold. It is a case of the coil current source supplying energy, and for a constant current source it can only supply energy if it is seeing a voltage.
QuoteMy argument is that this doesn't matter. As it should be enough to only look at the delta energy change of the system.
But if the two energies are in error then the delta will also be in error.
QuoteBecause assuming a super conductor where current keeps flowing, when you move the magnet out this current would indeed drop due to the change of circumstances
That is not correct. The current is constant, the change of circumstances cannot change the current, but it does present a voltage to the constant current source.
Quote, mainly the influence of the magnet on the domains changed. This change causes an energy delta in the system to occur. So even if your initial condition was such that the whole magnetic energy was 0 in your region. The delta will tell you how much it changed from its INITIAL condition. You seem to have added this complete Initial energy to your calculation and is confusing.
I do not add energy!! I correctly determine the energy delivered by the current source on current switch on and the energy recovered back to the source on current switch off. There is no magnet movement while the current rises or falls. During the rise and the fall the flux change is delivering voltage to the generator. I also take account of the flux change applied to the energized coil during magnet movement that also applies voltage to the current source.
QuoteThis initial energy can indeed be 0 or any value in the core region really. It is the overal magnetic energy change of the core (due to the movement of the magnet) that results in "flux change" and thus a reaction from the coil. This reaction and energy change should match exactly with the provided energy from the coil when considering a constant current source. The initial energy of the core with or without the magnet should be irrelevant as you pay for that once at the very start, and from there on only the delta matters as the system oscillates between these two states. I believe this is why I was confused about that added final value to the difference. You are overcompensation for the total input energy by also considering the "initial" energy to get the coil to 50A. The advantage of the Magnetic energy method is that you dont have to consider this. Because 1) you get most of it back when you power the coil down and 2) You only need to know the total magnetic energy change to know what changed for the coil energy wise. And thus its this delta that determines your total energy cost.
I also redid the analysis. I used a fully enclosed coil now as you did and a much finer mesh as you can see below. This time I also selected the copper coil region itself in the magnetic energy calculation. This would actually be an overestimation as this region also contains the magnet's field and doubles the previous values I had. But even with this better refinement, enclosed coil setup AND doubling of the magnetic energy values. The COP is still at 3x.
Using an incorrect method.
QuoteI believe the real debate here is, did I omit the initial energy of the coil or did you over account for it.
There is no initial "energy in the coil" in my calculations as I only consider energy delivered by the current source. I think the real debate is what method is correct, yours or mine.
QuoteThis overcompensation would be akin to me adding the total magnetic energy of the core when the magnets have moved to the final difference value of the magnetic energy. It wouldn't make sense as this energy can be recouped when powering down the coil. Only the delta is lost for good. So on that merrit my delta is even double that of yours and still get a COP of 3.
Since this calculation table also shows how the energy behaves without a magnet it is clearly evident that the magnetic energy or "flux" (a term I dont like to use) is reducing as the magnet moves away. As a consequence the coil will try to compensate by INCREASING the current. What other motor coil reacts in such a way to aid the magnets motion rather than oppose it?
The current can't INCREASE, it is a constant current generator that has infinite internal impedance.
Smudge
I have to agree and concede on the part of where during the movement the ELECTRICAL energy should be considered as well to maintain the current during the flux change as you point out. So yes that was an error of omission on my part and have corrected the mistake by adding this to the total value now. And this does bring the system closer to unity BUT I have noticed something interesting between the relation of the magnetic energy of the core and that of the magnet and the flux linkage which I dont see in your graph.
I will post this when it is finished.
I went the extra mile and tried to structure the data more clearly and used a visual aid we humans are good at picking up, colors!
So anyway the initial underunity result, after the correction due to the omission of the electrical energy, becomes an overunity effect beyond certain current flow. You can even see this under/over unity inflection point.
Something I dont also see in your graph is the sign of the flux linkage. Negative areas would reduce the total area of an integral no?
I ran all these with quite a high mesh refinement so it took quite a bit to gather manually. I attached the excel file too if you want to parse and validate the data or compare to the Flux linkage integral method.
So what makes this overunity effect? Well its evidently clear now from this data. At high enough currents, saturation kicks in and the flux linkage no longer changes that much. Evident from the change going down beyond 50A. Now pushing the current beyond saturation costs a lot less energy BUT what apparently does keep decreasing in a linear fashion is the mechanical force. This inflection point causes an the mechanical energy to dominate the electric energy input required to maintain the coil at a constant current.
Can you post the Excell file zipped then I can use your data in my calculations for comparison with yours.
Quote from: Smudge on 2024.03.03, 18:57:13
Can you post the Excell file zipped then I can use your data in my calculations for comparison with yours.
My bad, has been attached now including the FEMM file.
If a solenoid with a current in it is allowed to expand (as if it was a rubber solenoid), it does mechanical work, and at the same time the energy of the magnetic field in it increases. But this comes at the cost of increasing current. something like this...
Broli,
I have added my calculations to yor Excel file. Column O is the flux linkage data adjusted so that it shows the cumulative flux change seen by the coil for each current step. Column P is the co-energy that is the area between the BH curve and the current axis. It uses simple trapezoidal integration between steps as the current points are not suitable for Simpson's Rule. Column Q is the BH energy between the BH curve and the flux axis, obtained by subtracting the co-energy from the BH rectangle. This is input electrical energy for current rise and electrical recovered energy for current fall. Column R is the input-recovered energy difference. Coumn S adds the input energy due to flux change during magnet movement (adds your column K). Column T is your mechanical energy column J divided by my total electrical input.
Smudge
Hi Smudge thank you for your post. I will be building a lua script to increment in 1A steps to make the calculations even more accurate. I have done some reading on the concept of coenergy and you are right I now see how the flux linkage method (aka co-energy) is the more accurate way to determine the input/output electrical energy. So I will be using that method as well going forward.
However I have hinted at something multiple times now. But why are we substracting the energy we found using the flux linkage integral method?? At 1mm shift the latent coenergy of the coil INCREASED which is evident if you compare the flux linkage values in the table between 1mm and 0mm. Electrically during the maintenance of the 100A value sure that cost us some energy (which I erroneously left out initially). But the coenergy of the coil after we drain the coil is HIGHER so that is a WIN not a loss! I often catch myself too thinking in losses when it comes to OU but often miss a win when its right in front of me.
The co-energy is not the electrical input energy, it is the integral of the current with respect to the flux-linkage. Although it has dimension of energy it can't be used for the energy audit without using it to determine the other BH energy the other side of the curve. An increase in co-energy becomes a decrease in the Important real electrical energy.
Quote from: Smudge on 2024.03.06, 19:12:02
The co-energy is not the electrical input energy, it is the integral of the current with respect to the flux-linkage. Although it has dimension of energy it can't be used for the energy audit without using it to determine the other BH energy the other side of the curve. An increase in co-energy becomes a decrease in the Important real electrical energy.
The other side would be the method I initially used by determining the energy in the space and core. This is the BxH magnetic energy or also just called the energy or "other side" as you call it. However its either you use one or the other not both. Because coenergy is the actual effect you see when you power up the coil and down again to extract its energy once more. So for our electrical input side all that matters is this flux change as we power up and down the coil AND the intermittent flux change as we move the PM. This would give us the complete effect we experience when interacting with the coil from the electrical side. The rest is a black box to us. Now in this black box we have a different system interacting mechanically with its own energy inbalance. What matters is what we input in the coil, maintain a steady state for a bit, and then get back out of it. This is only the coenergy difference + intermittent electrical energy.
In nonlinear systems, energy and coenergy are not the same. The difference between coenergy and energy is related to the work done by the system. This is especially relevant in systems where the properties change with operation, such as in the presence of magnetic saturation or variable reluctance. The energy field would be a great hint at what this means but it is not part of the energy balance we only care about. The electrical side and the mechanical side, the inbetween magnetic energy just happens to happen. We can study it or meanwhile also use the extra energy to warm our cold butts.
Quote from: broli on 2024.03.06, 19:50:24
The other side would be the method I initially used by determining the energy in the space and core. This is the BxH magnetic energy or also just called the energy or "other side" as you call it. However its either you use one or the other not both. Because coenergy is the actual effect you see when you power up the coil and down again to extract its energy once more. So for our electrical input side all that matters is this flux change as we power up and down the coil AND the intermittent flux change as we move the PM. This would give us the complete effect we experience when interacting with the coil from the electrical side. The rest is a black box to us. Now in this black box we have a different system interacting mechanically with its own energy inbalance. What matters is what we input in the coil, maintain a steady state for a bit, and then get back out of it. This is only the coenergy difference + intermittent electrical energy.
You have got things the wrong way round. The co-energy is the area between the BH curve and the H axis. (Actually that is volume density and must be multiplied by the core volume to get energy.) Using flux linkage against current gives you energy directly where the co-energy is the area between the curve and the current axis and that does
not give you the input energy from the coil. It is the integral of Phi.di and that cannot be changed to i*V*t. The area between the curve and the flux axis is the one we want, that is the integral of i.dPhi that can be changed to the integral of i*(dPhi/dt).dt that becomes the integral of i*V.dt that is Power*time. It is known as the magnetic energy, not co-energy. To use your terminology, we should use that energy difference + intermittent electrical energy. When there is some PM field present we can't use the FEMM magnetic energy value since that does not relate to the Phi v. i curve because there are B*H areas where B comes from the PM, not from the coil. We have to use the intermediate current value to generate the Phi v. i curve and do the area integration outside FEMM.
Again you are correct, to be frank this is also the first time I heard of this idea which makes much more sense to use now that I understand it, thanks.
However this made it now even MORE clear to me why FEMM shows this OU effect. EVERYWHERE, core saturation is seen as a very bad thing in other words you are forced to stay on the linear part of the BH Curve as soon as you go non-linear you have gone too far and have "losses". However these losses become gains in this design, the higher you drive the core into saturation the lower the losses! The only limit is the current through a copper coil. At high enough currents even the differential flux change at the intermediate phase becomes so low. Meaning the difference in magnetic energy becomes lower and lower as you go higher with the current and remarkably the force even flips at a certain point.
So why does FEMM show such atypical behavior? Because this has been part of electromechanics all along! We have been mostly using the "linear" part and didn't dare to touch the "non-linear" part however modern day PMs are very strong ex. NdFeB and this would allow for significant force differentials. So then the magnet is the energy source? Yep.
Still working on parsing the data of the lua runs, I am getting a bit over obsessed again and affecting my personal life because of this. I attached the data and some rough data structuring, would be nice if you did the parsing Smudge. Thanks
Have been watching an interesting web series on this subject:
https://www.youtube.com/watch?v=85EX0LpDrVo&list=PLHq09ObRyxDBD3AJXHW4LI8egLSd0Eb1F
I am still running all kind of tests and am curious to see what would be the most "power dense" design. Ironically the more coercive the core material the better, because you are not fighting the already alligned domains only the ones aligned by the permanent magnet. Your only true enemy is ohmic losses and eddy currents.
Attached is the data of two runs one where the core has an air gap and one where it doesn't. Using the flux linkage method now one can see clearly how the COP remains at around 1 because of the air gap provides highly linear behavior. However in the second sheet contains the data of the setup with no airgap, now when the system reaches its non-linear phase the COP goes beyond 1.
EDIT: A very interesting recent paper on the behavior of magnetic hysteresis. It shows the behavior of "branching" quite well too. I added a BH curve that represents the BH hysterisis losses and shows that a high core coercivity actually reduces the core losses in this setup where the green area represents the energy loss.
https://www.mdpi.com/1996-1073/16/9/3908
Broli.
I have calculated the COP using your FEMM data and I can confirm that it does indeed show COP>1 above 70A current. The attached image compares my COP to yours and the difference is merely a shift sideways by 1A and that is just our different ways of doing the fit to the FEMM data, shifting the results one line in the spreadsheet. The deviation from near unity at the lower currents can be explained by the crudity of the integration over the non-linear region of flux v current. I need now to find out why (a) the results are just below unity over most of the region below 70A and (b) why the results are above unity at over 70A. I know that the atomic current circulations (electron spins and orbits) responsible for the magnetization in both ferromagnetic hard and soft materials can deliver energy (and sink energy) as though they are actual currents in coils. (F6 will disagree but that doesn't matter.) In the case of PMs the effective coil current (just Google equivalent surface current) is constant over the full cycle so there is no net energy gain. But in some parts of your scheme the effective coil currents are not constant (the Fe part joning the two NdFeB magnets and of course the ring core) and it is here that I expect to find the source of the excess energy (and the sink for the lost energy below 70A)
Smudge
Quote from: Smudge on 2024.03.11, 16:14:25
Broli.
I have calculated the COP using your FEMM data and I can confirm that it does indeed show COP>1 above 70A current. The attached image compares my COP to yours and the difference is merely a shift sideways by 1A and that is just our different ways of doing the fit to the FEMM data, shifting the results one line in the spreadsheet. The deviation from near unity at the lower currents can be explained by the crudity of the integration over the non-linear region of flux v current. I need now to find out why (a) the results are just below unity over most of the region below 70A and (b) why the results are above unity at over 70A. I know that the atomic current circulations (electron spins and orbits) responsible for the magnetization in both ferromagnetic hard and soft materials can deliver energy (and sink energy) as though they are actual currents in coils. (F6 will disagree but that doesn't matter.) In the case of PMs the effective coil current (just Google equivalent surface current) is constant over the full cycle so there is no net energy gain. But in some parts of your scheme the effective coil currents are not constant (the Fe part joning the two NdFeB magnets and of course the ring core) and it is here that I expect to find the source of the excess energy (and the sink for the lost energy below 70A)
Smudge
Hi Smudge, thanks for the extra validation. I have decided to let it rest for a bit as the data was starting to play tricks on my mind. Depending on how I integrated both sides (power up vs power down) I got everything from under, over and unity which is quite strange to say the least for such a small change in calculating the area under a curve. It weirded me out for a bit and sometimes its best to approach something when your mind has been cleared.
I wanted to make a post but I am also a visual person and like to use pictures references throughout text. So I deciced to write it down in a word file which I attached.
Will attach the referenced experiment later but it cant be more simpler than a toroidal coil and a magnet sim.
EDIT: Updated the document with new illustrations and examples.
I believe the source of energy is Earth magnetosphere
Quote from: forest on 2024.03.19, 05:47:24
I believe the source of energy is Earth magnetosphere
But Earth magnetosphere is actually too slight. :o
Broli,
KJ Magnetics manufacture NdFeB circular arc segments that can be put together to form a ring where the field is virtually fully confined within the ring, see FEMM result in the image below. I have played with these and you need great care in building up the ring because of the huge attractive forces involved. So you can create your ring magnet experiment.
Smudge
Quote from: Smudge on 2024.03.19, 11:52:50
Broli,
KJ Magnetics manufacture NdFeB circular arc segments that can be put together to form a ring where the field is virtually fully confined within the ring, see FEMM result in the image below. I have played with these and you need great care in building up the ring because of the huge attractive forces involved. So you can create your ring magnet experiment.
Smudge
Thanks for the suggestion Smudge, I was thinking about doing something similar with very flat cylindric magnets and orientating them in a circular path but this seems less of a pain in the ass.
I am still not too sure how the coercive strength of different PM materials could affect the potential experimental result. I would assume you dont want something that is too hard of a PM for it to not reorientate much of any domains and not too soft so the domains can still "spring back" due to their neighboring aligned domains.
Here is another illustration I did and an analogy on the previous one in the long document I attached previously. To kind of also illustrate what happens inside of the core. All the forces and torques make more sense if you view it this way rather than the concept of imaginary fields. In fact I believe we could even physically SEE this gyromagnetic effect in action by using thin disc/square magnets on a wheel and attaching to bearings so they can swivel as seen in the attached illustration.
The "rotor" magnet would be in a perpetually torqued state due to the continuous angular momentum change it causes on the flat magnets (domains in a core). This continuous and perpetual imbalance causing a net angular momentum change over every angle causes a unidirectional torque. If the rotor magnet were a coil it would see a change in field and react to this by reducing its "current" which would be an energy loss as we know. But this does not happen due to the permanent spin of the electron, and THAT is where the energy comes from.
Would be interesting to see if such setup would impart enough angular momentum as the gyromagnetic effect is rather small. However recently we have seen that this is not always the case and that we can see this effect even on the macroscopic scale:
https://www.youtube.com/watch?v=oijGMLErqck
I believe the "relaxing" phase might also be important because you need a continuous change occurring.
PS: The force vectors shown is NOT what causes the proposed effect. These are merely the electromagnetic forces at play, these are conservative and dont cause energy changes, however the gyromagnetic effect is there too and THAT combined with the nature of spin linked to the crystal lattice mechanical momentum is what I propose. The magnet array just helps to align the misaligned magnets back and reset the whole system essentially for free which in a purely mechanical gyroscopic system would be almost impossible to emulate without major losses. From keeping your gyroscopes spinning with motors, to perhaps a hydraulic system for the spring back effect to using a conservative force like electrostatics to the direction flipping? Electromagnetism and Gyromagnetic ratio essentially eliminate all of this complexicity for us and we end up with little system losses, in order to maximize the true energy source by torquing down the spin of the very electron.
Recently I have been going back to the work done in this thread. More importantly the OU claim by FEMM data and also pointed out by smudge here:
https://www.overunityresearch.com/index.php?topic=4594.msg110807#msg110807
However after digging deeper I want to clear the record. As this anomaly was caused by using an incorrect current sign. You could say at the time I was too dumb to understand that this was crucial to the overall balance. As this lead to a portion of the electrical work contribution to be negative (energy gain) rather than positive (energy lost). After correcting that, the COP drops to essentially one across the whole range as can be seen in the attached images. However this gave rise to another OU artifact at low currents this time. However after running simulations at smaller current and distance increments this artifact was also flushed out. So this essentially means there is no OU in the system.
The reason why I went back to this is because first of all I wanted to be really sure about the results and second I believe there still might be way to produce excess energy by leveraging core saturation. However I believe there is only a very specific window where a few conditions need to line up properly otherwise it might end up being so small that it essentially falls below the noise in the data. I believe if you combine a specific BH curve with a specific strength of the magnet and geometry of the coil you can maximize this excess energy significant enough to show up in the data. At least the FEMM validations so far are encouraging as they show FEMM can validate these mechanical vs energy exchanges quite accurately and zoom in on anomalies if needed. However we need excess energy the range of at least 10% to be confident of something interesting going on.
I made a simple idealized BH curve to further study this and it seems you can achieve excess energy across a full cycle. Of course I might have slipped up somewhere as it felt like walking a mine field but if anyone can point out a mistake happy to look into it.
Attached is the energy book keeping for one full cycle using the different areas of the BH curve. It was not easy figuring this out because during operation the BH curve essentially moves around.
The BH curve can be found here if you want to validate it it yourself
https://cad.onshape.com/documents/ace0a70136def5c66634755d/w/96e2b40b2e89426c173a6b7d/e/b62559c0fb5200307010e3a4?renderMode=0&uiState=6a595e4f2b20a37e38cc3f53
Essentially the core and a coil with a very high current can double the mechanical work. Because the magnet essentially feels double the current that is flowing in the coil because of the core if the coil has a high current equaling the cores surface currents due to its aligned electrons. But you only pay an electrical penalty at the source for the coil not the core. Next step is to validate this in FEMM.
Your summation has some additional terms that shouldn't be there. There are seven different areas of the BH curve that you add or subtract, but you sum 12 values!!! When you restrict yourself to the seven values you get a net deliverance of energy from the current source of 14 units of energy that FEMM will equate to the work done by the moving magnet being attracted then pushed away.
Smudge
Quote from: Smudge on 2026.07.18, 07:43:22
Your summation has some additional terms that shouldn't be there. There are seven different areas of the BH curve that you add or subtract, but you sum 12 values!!! When you restrict yourself to the seven values you get a net deliverance of energy from the current source of 14 units of energy that FEMM will equate to the work done by the moving magnet being attracted then pushed away.
Smudge
Hi Smudge thanks for jumping in.
And I think you meant 11 not 12 if so then yes you are right because 4 of those 11 values I consider to be mechanical work. Why is that? Well the logic is as follows. When we power up the core with the coil, besides its negligibly small saturation current, we go so far up with the current to essentially double the saturated core flux by using it as pure air coil which means we need to drive the coil current to VERY high levels to about match the fictional surface current of the already saturated core. This can be 10000's of amperes for iron for instance. Now our magnet that is going to be attracted to this will essentially feel the coils very high current and an additionally matching current due to all the core's aligned electrons contributing to its fictional surface current. And that is the real key here. Without the core the mechanical work would indeed have a 1:1 ratio with the electrical source that is maintaining the constant current of the coil. By adding the core however the electrical source still needs to provide the same amount of energy to maintain the coil current but mechanically we about doubled the work because the magnet sees 2 nearly equal strong currents now versus the singular high current if there were no core. Again electrically the flux change is the same for the coil with or without the core as the core is pushed way beyond saturation so the only flux change it sees is due to the magnet approaching which is the same regardless of the presence of the core. And
that is how we gain an additional 1 unit of work during the mechanical cycle.
FEMM calculates forces based on field strength and since both the core and coil contribute to this field that would be as if the magnet sees not one but two coils so I am confident that the force will also increase by nearly double as much leading to double the mechanical work than if only the coil were there.
Now I know this is just an idealized example and I believe you wont get an exact equivalent extra mechanical work from a saturated core but I would bet it would be pretty darn close if we tried to match the coil and core's "surface" current, saturation levels and the strength of the magnet.
Quote from: broli on 2026.07.18, 10:13:22
Hi Smudge thanks for jumping in.
And I think you meant 11 not 12
Yes 11 not 12, my mistake.
Quoteif so then yes you are right because 4 of those 11 values I consider to be mechanical work.
But you have double accounted on the basis that the core creates a force value on the magnet and the coil also creates the same force value, so double the force and double the energy. IMO that is wrong an I would bet my life savings that FEMMM will show it.
Smudge
Quote from: Smudge on 2026.07.19, 07:29:09
Yes 11 not 12, my mistake.But you have double accounted on the basis that the core creates a force value on the magnet and the coil also creates the same force value, so double the force and double the energy. IMO that is wrong an I would bet my life savings that FEMMM will show it.
Smudge
Well exactly double would a bit too hopeful indeed. But here is a quick simulation run using axis symmetric for a simple coil+core and magnet setup:
core+coil:0A
z-component: -43.657 N
coil:5000A
z-component: -79.0977 N N
core+coil:5000A
z-component: -152.257 N
coil:10000A
z-component: -158.845 N
core+coil:10000A
z-component: -261.043 N
coil:50000A
z-component: -803.702 N
core+coil 50000A
z-component: -1113.47 N
coil:100000A
z-component: -1626.98 N
core+coil 100000A
z-component: -2085.26 N
What this shows is that there is a clear trend in that the core is providing quite a signifcant boost. However if the current goes too high the coil will dominate mostly and you get diminishing returns. So somewhere in that broad current range there is a point where you get the most bang for your buck so to speak. But I would say that 5000A is already quite a good candidate for being nearly double. I have no use for your life savings though :)
I can see where you find more bang per buck with the coil+core compared to only the coil. But you have not provided the input energy (bucks) for your more bangs which FEMM + some calculations will give you.
Smudge
Quote from: Smudge on 2026.07.19, 09:39:14
I can see where you find more bang per buck with the coil+core compared to only the coil. But you have not provided the input energy (bucks) for your more bangs which FEMM + some calculations will give you.
Smudge
Yes that will require a more complex position sweep analysis which is being worked on.
It's funny how things sometimes go. I dropped the mechanical aspect of this problem and essentially arrived at the most core representation of this idea. Instead of using a magnet to modulate the permeability of the core. We could reduce the problem to two problems that interact with each other. Essentially a toroidal current combined with a circular coaxial current. When you combine both without a core they would operate completely independent of each other as they would never couple to each other. However when you introduce a non linear medium between them they can interact. Now the two worlds affect each other around the saturation point of the core.
The good news is that this heavily simplifies the problem and gets rids of all kind of issue like dealing with the very strong demagnetization field which led to using extremely high currents. Because the field is completely bound now the currents can stay at a moderate strength to reach the saturation point. The bad news is that this no longer a purely 2d problem which FEMM can solve. However on the plus side it's more of a 2.5d problem. Meaning a 3d solver only needs to solve a tiny slice of this substantially reducing computation times. In fact you can reduce the problem to a slice of the azimuthal section and even a half slice of the cross section of it.
The theory goes as follow:
• Energize the toroidal coil to saturation of the core then keep the current constant
• Increase the current of the coaxial loop
• Because this reduces the flux of the toroidal coil it will induce an EMF
• Now deenergize both simultaneously back down
So where is the excess energy? Well the induced EMF is energy gained first of all. And the last part where both currents are reduced simultaneously means that the coaxial loop returns all energy it was given because from its perspective its flux did not change to warrant a changed energy state. That only leaves us with the initial energy given and final energy obtained from the toroidal coil. The argument is that this dropped energy + energy gained while the current was kept constant is more than the initial energy it was given. Essentially the area for one drops as the base of the triangle drops while at the same time we gain energy not as a triangular area but a rectangular in our B-H curve or in other words you gain twice the energy drop from maintaining the constant current.
Anyway who has Ansys Maxwell laying around to give this a shot?
AI assisted me with writing a script to better visualize the field inside of this configuration. It shows the behaviour of the field when you switch between the common toroidal current and the coaxial current. At their half mark their sum produces these very cool helical fields around the toroid. It's almso very similar or almost identical to the fields generated in a tokamak reactor as essentially the same coil setup is applied. But of course that is also where the similarities end as the idea here is to use these fields and couple them using a non-linear core rather than generate fusion.
https://www.youtube.com/watch?v=wu6n5A8OQNc
(https://cdn.imgchest.com/files/b7546743d1e3.png)
Quote from: broli on 2026.07.21, 15:56:18
The bad news is that this no longer a purely 2d problem which FEMM can solve. However on the plus side it's more of a 2.5d problem.
FEMM has this axisymmetric mode that allows quasi-3D simulations.
Quote from: Verpies on 2026.07.24, 09:51:19
FEMM has this axisymmetric mode that allows quasi-3D simulations.
Sadly it only allows you to simulate the coaxial currents which would be modeled as current going in and out of the screen in FEMM. However this configuration also requires you to define currents that would essentially be circular on the screen. See attached, FEMM's axis symmetry can not model the green circular current which would make up the toroidal winding. This is why I refer to this as a 2.5D problem as a 3d solver would only need to model a single infinitesimal slice of it.
I perhaps need to revise my previous statement. After doing some theoretical work which is essentially nothing more than simple Pythagorean trigonometric identity. The maximum "COP" of this system is projected to be exactly square root of 2 or 1.4. This happens when both fields twist the domains 45 degrees.
The previous description of the process led to exactly unity. However there is one path which does not.
1) You energize one coil
2) While maintaining a constant current on it you bring the other coil to the same current
3) Then you deenergize both coils simultaneously back down
What that does is rotate the saturate domains by 45 degrees and reduce the flux linkage of the constant current coil.
The math says the energy you get and need to provide for that stage are exactly equal. However what is not equal is the difference between the input energy of step 1 and and the output of 2 coils in step 3. The math says that is exactly sqrt(2) more energy which has to do with the sin and cos of 45 degrees.
Quote from: broli on 2026.07.24, 15:58:50
2) While maintaining a constant current on it you bring the other coil to the same current
Won't this rotate the ferrite's magnetization vector because of H vector addition of the contributions from the two windings ?
Quote from: Verpies on 2026.07.24, 17:25:33
Won't this rotate the ferrite's magnetization vector because of H vector addition of the contributions from the two windings ?
Yes that is the point. There is a very simple trigonometric relation of angle and flux change that competes with the linear part of the energy extraction.
I am currently extracting data from the previous FEMM sim I shared. However this might be the worst case scenario for this configuration because the setup nearly behaves like an air core regardless of high of a relative permeability you chose for the core. Unlike a core that has its field fully contained and saturates at low currents this needs very high currents to reach that point potentially drowning out the energy surplus.
Attached you see this. The blue curve is the current v. flux graph when you energize a single coil. The saturation happens around 100KA which is quite high. This does not depend on the core chosen but due to its own high demagnetization field because the field is not contained.
However what is very encouraging in the preliminary data I have gathered so far is that there appears to be a COP bump around the point where we should expect one aka the saturation point. At around 100kA there is a bump in the COP!
Skeptics might say that a COP of 1.00000189 is barely worth noting and they might be right. However I have three arguments against that. First this data says something interesting is happening around 100KA for some reason the total energy ratio is not stable around the saturation point. Is that a mere coincidence when the theory in ideal conditions says it should be around that same point? Second arguably the most accurate value FEMM produces is the flux linkage. Unlike force calculation that heavily relies on a stress tensor mask and thus the underlying mesh discretization, flux linkage is an incredible stable value that converges very quickly on even coarse meshes. So any fluctuation in energy from using this value is highly noteworthy for a solid state system like this. And the third and final point, the maximum energy surplus is a fixed constant amount that peaks only around the saturation point. If you stay below saturation you will never see it if you go too far beyond it you might end up losing it in the noise of those high energy levels. This might explain why this open core is essentially the worst case scenario as the window to find the current that gives the highest possible COP is very small.
So the logical next step is to hunt for the exact current value that gives us the highest COP peak. The current data uses very coarse current steps but it already hints at us to look around 100kA which is exactly where we expect to find any surplus. This search can take many days but hopefully it can provide an exact current value at which the most magic happens to motivate further investigation.
I just realized something. FEMM gives me the freedom to just surround the coil + core with the same non-linear material to contain the field as much as possible.
The preliminary data shows this reduces the saturation current by A lot! So much so I need to rerun the whole thing with smaller current steps at a lower maximum range. But what is very exciting is that this also increases the amount of flux drop by a lot. The theory predicts this should be around a max of 30% or 1/3 because 1-sin(45)= 0.3 in best case scenario. The current graph attached shows its in the ball park of this figure now opposed to previous weak flux drop. Very curious what the full energy analysis will show now.
Quote from: broli on 2026.07.21, 15:56:18
...the coaxial loop returns all energy it was given because from its perspective its flux did not change to warrant a changed energy state.
Please elaborate about this lack of change of flux.
Quote from: broli on 2026.07.24, 18:31:00
Skeptics might say that a COP of 1.00000189 is barely worth noting ...
Only if this is a numerical precision artifact of the simulation.
Also, please settle your directional terminology for me - IMO: In the illustration below:
1) the current flows in the "toroidal" direction when flowing in the green winding
2) the current flows in the "coaxial" direction when flowing in the one-turn thin yellow ring.
3) the current flows in the "? ? ?" direction when flowing in the yellow shell.
Confirm, deny, correct, amend....
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4594.0;attach=56772)
I wish you had made the ring and shell different colors for clarity. What CAD was this drawn in, anyway ?
Quote from: Verpies on 2026.07.24, 20:56:43
Please elaborate about this lack of change of flux.
Only if this is a numerical precision artifact of the simulation.
Also, please settle your directional terminology for me - IMO: In the illustration below:
1) the current flows in the "toroidal" direction when flowing in the green winding
2) the current flows in the "coaxial" direction when flowing in the one-turn thin yellow ring.
3) the current flows in the "? ? ?" direction when flowing in the yellow shell.
Confirm, deny, correct, amend....
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4594.0;attach=56772)
I wish you had made the ring and shell different colors for clarity. What CAD was this drawn in, anyway ?
"Please elaborate about this lack of change of flux."
You can ignore the procedure given in that post. After doing the math it led to perfect unity during the entire process. However the current process does not. Again repeating to make it clear what is happening:
1)Bring core to saturation with one coil
2)Now keep that current constant while you bring the other coil up to the same max current. Both coils oppose each other 90 degrees, that is why in air they would not interact at all but with a core that is non linear they do couple. This causes the flux to drop in the first coil because of the non linear core. Essentially you are rotating the saturated domains by 45 degrees. Say the flux had a value of 1Wb at saturation when the domains rotate 45 degrees it can drop to as much as sin(45) or to about 0.7Wb. Its simple trigonometric identity.
3)Now both coils essentially have the same current and flux going through them. Finally bring both coils down simultaneously and recapture their energy.
In ideal cases the theory says step two is complete unity and all that remains is the balance between the input of step 1 for one coil and the output you get back from the sum of 2 coils at a slightly lower flux aka that 0.7Wb. So say input 1A*1Wb/2=0.5J vs 2*(1A*0.7Wb/2)=0.7J or a max cop of 0.7/0.5=1.41 or sqrt(2) to be exact.
QuoteAlso, please settle your directional terminology for me - IMO: In the illustration below:
1) the current flows in the "toroidal" direction when flowing in the green winding
2) the current flows in the "coaxial" direction when flowing in the one-turn thin yellow ring.
3) the current flows in the "? ? ?" direction when flowing in the yellow shell.
1)yes
2) see attached
3) see attached
essentially the inner wire and outer sheet are one and the same coil and act just like a coaxial cable where current flows in opposite direction with respect to each other. One of them might not be necessary but having both of them present compliments the toroidal field nicely as the coaxial coil would generate a circular field that would be completely contained in the core just like the toroidal coils fully contained field. That would be the most ideal condition. That is why you need to use trickery in FEMM to truly simulate this. A setup with said currents would be the perfect configuration for this. Because the toroidal coil and coaxial coil generate a field which is 90 degrees from each other. You saturate the core with one coil and then use the other coil to rotate the saturated domains 45 degrees and finally capture the energy of both. The data I am sharing is going over that cycle using the current v. flux graph to calculate energy input/output of each step.
Just finished an 8 hour data collection run. However it is quite evident it overshoot the area of interest by a lot. Attached you see the flux v. current graph. The core reaches saturation around 50A this time so I must lower the max current sweep by a lot.
The energy data is coarse and unreliable around these lower currents and needs a more refined run but I shared it for completion. Time to start another 8 hour run.
Quote from: broli on 2026.07.24, 22:04:14
So say input 1A*1Wb/2=0.5J vs 2*(1A*0.7Wb/2)=0.7J
Why 1A ?
Quote from: broli on 2026.07.24, 15:58:50
2) While maintaining a constant current on it you bring the other coil to the same current
and how do you propose to do that?
You can assume an ideal coil when answering this question because non-ideal coils complicate analysis and don't bring anything beneficial to the table...
Sadly the new run highlighted a fundamental issue with this solid state design. The COP is essentially one across the tested range. What I did wrong in the math was only consider a drop in saturation field. However the simulation highlights that both saturation field and current drop which in hindsight makes sense really. After including the correct trigonometric value for that then the math also shows perfect unity.
I guess it would have been too good to be true to get rid of the mechanical aspect of this.
Quote from: Verpies on 2026.07.26, 04:01:31
Why 1A ?
and how do you propose to do that?
You can assume an ideal coil when answering this question because non-ideal coils complicate analysis and don't bring anything beneficial to the table...
It's just an example value to keep the math simple. And yes in FEMM these synchronizations are easy to pull off that that is why I am doing the validation in it. If FEMM confirms extra energy then that would be significant as it would imply its fully compatible with EM theory.
Quote from: broli on 2026.07.26, 20:23:30
It's just an example value to keep the math simple.
You misunderstood. I was
not asking about the particular numeric value of the current. I was asking why the current was the same.
Let me rewrite your mathematical statement in a proper symbolic manner so there are no misunderstandings:
Quote from: broli (rewritten symbolically by Verpies)
So say input i1*Φ1/2=W1 vs 2*(i2*0.7*Φ2/2)=W2
Why i
2 = i
1 ?
Quote from: broli on 2026.07.24, 15:58:50
2) While maintaining a constant current on it you bring the other coil to the same current
and how do you propose to maintain that current in your 1-2-3 plan ?
Quote from: broli on 2026.07.24, 15:58:501) You energize one coil
2) While maintaining a constant current on it you bring the other coil to the same current
3) Then you deenergize both coils simultaneously back down
You can assume an ideal coil when answering this question because non-ideal coils only complicate the analysis and don't bring anything beneficial to the table...
Quote from: Verpies on 2026.07.26, 22:00:14
You misunderstood. I was not asking about the particular numeric value of the current. I was asking why the current was the same.
Let me rewrite your mathematical statement in a proper symbolic manner so there are no misunderstandings:
Quote from: broli (rewritten symbolically by Verpies)
So say input i1*Φ1/2=W1 vs 2*(i2*0.7*Φ2/2)=W2
Why i1 = i2 ?
In the coming days I will make a better illustration of the different steps and their math. Granted I have made some mathematical mistakes so far but all of it should be ironed out now. The solid state run was an important aspect to this as it cleared out the mistakes and finally made clear why there IS a fundamental difference between it and the mechanical design using a moving permanent magnet. I finally have a good explanation of why and where the extra energy comes from. I believe Smudge has written a paper on this; a permanent magnet is essentially a constant current source that can be exploited.
Quote from: broli on 2026.07.26, 22:10:41
The solid state run was an important aspect to this as it cleared out the mistakes and finally made clear why there IS a fundamental difference between it and the mechanical design using a moving permanent magnet.
So you are going to revert to the mechanical design.
When you do, keep the answer to the following question in your mind:
Q: What happens to the current flowing in an ideal coil when a p. magnet is mechanically removed from it ?
(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55543)
A permanent magnet removed
from an ideal shorted coil.(https://www.overunityresearch.com/index.php?action=dlattach;topic=4525.0;attach=55545)
Note that the speed of the magnet's motion DOES NOT have any influence on the final current flowing in such coil.
I'm always seeing the 'ideal' examples... Neat to think on a bit, but it is beyond the real world.
The question would be, can you even remove the magnet from the shorted 'ideal' coil?
mags
Quote from: Magluvin on 2026.07.27, 16:26:13
I'm always seeing the 'ideal' examples... Neat to think on a bit,
That's for the reason stated below:
Quote from: Verpies on 2026.07.26, 22:00:14
You can assume an ideal coil when answering this question because non-ideal coils only complicate the analysis and don't bring anything beneficial to the table...
Quote from: Magluvin on 2026.07.27, 16:26:13
, but it is beyond the real world.
Superconductive coils really exist even if they exceed your budget.
Poor-man's alternative to S.C. is a short time scale relative to the L/R time constant.
Quote from: Magluvin on 2026.07.27, 16:26:13
The question would be, can you even remove the magnet from the shorted 'ideal' coil?
Yes, but the closer the coil hugs the magnet, the more force is required to remove that magnet. The mechanical energy needed for removal to infinity stays the same, though.
Quote from: broli on 2026.07.26, 22:10:41
... a permanent magnet is essentially a constant current source that can be exploited.
A permanent magnet can probably be exploited, but only as an accessory, not in its true nature.
The laws of nature are such that translations or rotations that are invariant over time require no energy to be sustained.
Whether it be electron spin or the current in a superconducting coil, as no energy is required for the motion, we cannot extract this hypothetical energy that would sustain the motion – it does not exist. Energy is only required to set something in motion, or to bring it to a halt. We can only recover the energy from the cessation of motion.
That said, slowing down or stopping electron spin is a grand undertaking; we would surely recover energy, but from where – really, at no cost? And I suspect that some version of Murphy's Law will prevent us from doing so.
Attached the concept. I hope some critical minds can find the flaw with it. Key things I want to stress is the anisotropic behavior of the core and the distance between each core as to avoid them mechanically interacting with each other. I perhaps didn't do a good job illustrating the latter but was too tired to put even more work into it.
But the two main issues is the mechanical energy and the condition that the cores are far enough apart as to not interact with each other but a bigger distance changes the area of each coil too. But if they are literally meshed into each other you are essentially going in exchange interaction theory.
EDIT: After some pondering I don't even think the last remark should be an issue. All you need to do is change the relative permeability to 3 of the cores and all the numbers of this toy model should remain the same essentially. Attached an 3d example model for this.
Quote from: broli on 2026.07.29, 17:44:38
2. While rotating and maintaining constant current:
1. electrical input from source: 2*1A*2Wb= -4J
2. Mechanical work gained: +4J
Please justify this.
Quote from: F6FLT on 2026.07.28, 15:44:27
A permanent magnet can probably be exploited, but only as an accessory, not in its true nature.
The laws of nature are such that translations or rotations that are invariant over time require no energy to be sustained.
Whether it be electron spin or the current in a superconducting coil, as no energy is required for the motion, we cannot extract this hypothetical energy that would sustain the motion – it does not exist. Energy is only required to set something in motion, or to bring it to a halt. We can only recover the energy from the cessation of motion.
That said, slowing down or stopping electron spin is a grand undertaking; we would surely recover energy, but from where – really, at no cost? And I suspect that some version of Murphy's Law will prevent us from doing so.
The Coler/Unruh Stromerzeuger discussed in another bench broke Murphy's Law as verified by some Professors some 100 years ago, but fearful for their reputations thay forbade publication of their findings. The unusual feature was current flowing within the ferromagnetic cores, and that brings us into the relatively new science of Spintronics that was unknown to those Professors. It is now known that electrons have, in addition to their electric charge and their mass, two other linked features, angular-momentum and magnetic dipole-moment, now under the name Spin. A rotating mass that occupies volume will have angular-momentum, and a rotating charge that occupies volume will have a magnetic dipole-moment, so using Spin to describe these two features is not unreasonable. The Law of conservation of angular-momentum applies to any system where momenta are mixed and that leads to the conservation of energy. I ask the question, is there a similar Law of conservation of magnetic-dipole moment that would apply to any system?
Smudge
Quote from: F6FLT on 2026.07.28, 15:44:27
A permanent magnet can probably be exploited, but only as an accessory, not in its true nature.
The laws of nature are such that translations or rotations that are invariant over time require no energy to be sustained.
Whether it be electron spin or the current in a superconducting coil, as no energy is required for the motion, we cannot extract this hypothetical energy that would sustain the motion – it does not exist. Energy is only required to set something in motion, or to bring it to a halt. We can only recover the energy from the cessation of motion.
That said, slowing down or stopping electron spin is a grand undertaking; we would surely recover energy, but from where – really, at no cost? And I suspect that some version of Murphy's Law will prevent us from doing so.
I agree for the most part but make an important distinction. If we read the literature and prior art of many of these FE inventors we find a divergence in their perspective from others. We find they are less concerned with science and more concerned with the result. They build weird stuff to see what happens.
For example, when someone says "permanent magnet" I suspect we both see a picture of a classic PM in our mind. The PM is rectangular with a red north pole on one end and a blue south pole on the other end just like our textbooks. This bias was programmed into us at an early age in school which is why we can all relate to the subject. As well we almost always take the purists view due to our bias and tend to reject anything different. I'm no different and when someone say's PM I see the same thing. However it took me about 40 years to understand I am biased and try to work around it.
For example, I heard a guy named Howard Johnson was doing a lot of stuff with magnets and magnet motors about 30 years ago but he seemed like a bit of a quack. On the surface this seems to be the case until we look deeper. So I did the research and found Johnson was doing something most of us cannot even imagine. Johnson was making his own permanent magnets which are nothing like the ones you and I presume to be normal.
Here is the AI version of how to make a PM,
"Permanent magnets are made through a process called Powder Metallurgy, which involves melting raw materials, compacting them into a fine powder, and then aligning the particles to create a magnetic field. After compaction, the magnets are heated in a vacuum to densify and then finished to achieve the desired shape and strength."
However what Johnson was doing was very different and some areas of the magnetic powder(s) were compressed to a higher density than others at a lower density. As well he used different materials in different areas which may have been diamagnetic. As such our purists view of what a magnet should look like and act is not accurate. Do you know what happens if we were to dope any area of a permanent magnet with a variable transition of material(magnetic/diamagnetic) and density?. I don't, I have no freaking idea because the complexity of the possible number of variations compound on one another. So the field could act normally as expected until it meets some threshold then it could be opposed, could turn, could expand, it could do any number of unexpected things.
It could be as simple as, if we always make our magnets the same way we get the same results. If we make them unlike anything else the results may differ...
And digging we keep on doing. The previous designs all lead to unity but it opened new insights and better understanding. Here is the latest design in trying to exploit the electron spin.
Say you had a coil which you could stretch out to make longer. What would happen? Well to keep things simple lets consider a coil made out of two layers so we can translate one layer of it to essentially double its length. Additionally lets say you hold the current constant during translation.
(https://cdn.imgchest.com/files/bb0114eabd31.png)
When you do this you will experience a force against said translation which requires you to do mechanical work to essentially stretch out the coil. Magnetically this would roughly lead to halving of the flux in the coil since we didn't change the current or # of windings only its length. Also since we keep the current constant during translation this would lead to pumping electrical energy back into the source. However the full round trip energy bookkeeping will conclude that the electrical energy gained will be exactly compensated by the mechanical work required to lengthen the coil. So perfect unity.
However what if we added a ferromagnetic material to this system. Consider this core material to be made out of long individual strips rather than one solid core. Again we magnetize the coil which will now also magnetize the core. And again we pull out the coil and this time we also pull out half of the core strips with it. What will happen now? You might naively state that this is exactly the same situation as before since all we did was change the relative permeability of the system so everything just multiples by said factor and the energy book keeping will ultimately lead to the same balance.
(https://cdn.imgchest.com/files/176791162771.png)
However there is one major flaw with that reasoning. Yes the coil-on-coil interaction remains the same. But the cores don't mechanically act like coils. In fact they do opposite thing. Their forces actually point in the direction of motion and thus assist the translation!
(https://cdn.imgchest.com/files/73d16479483f.png)
So now we have this uncomfortable situation where during the separation the flux is dropping so we are pumping energy back to the source just like we did before with the coreless coil. However mechanically we have a different story as we are now gaining mechanical energy because the core pieces are essentially pushing each other out rather than wanting to remain stuck to their neighbors.
If you are not convinced by that here is a quick FEMM simulation showing an array of magnets that are all aligned in the same direction but half of them have been translated. As you can see their force is pointing in the positive x direction.
(https://cdn.imgchest.com/files/e831651f72e6.png)
This of course does not represent the described system but at least hints at an interesting behavior. I also came to this conclusion from a completely different design but I wanted to lay the groundwork and use the simplest possible design to explain it.