No you can't lift yourself by the bootstraps, no matter how hard you pull. The argument that it is easier to confine electrons than ions is flawed, because if you use magnets to confine electrons and then use the electrons to confine ions, then you may as well use the magnets to confine the ions in the first place.
For once I agree with Richard, it sound a lot like perpetual motion
This said, I still believe we are going to crack the Lawson criteria, and burn deuterium, and my feelings at the moment is that magnets is the hard way to do it.
Derek I used to restore old Lincoln Continentals as a hobby in the late 70's and the early 80's. One of my many hobbies. I once had 10 restored Lincons registered to me, but once gas hit 75 cents per gallon they had to go with the 430's and quad carter guzzling 7-9 mpg. 75 cents per gallon was outrageous then. When I started on old cars, gas was 30 cents a gallon.
BUT, back to fusing ...er...or not fusing.
The electron loses are miniscule compared to the losses associated with the simple failure to fuse in the first place.
The difference between getting 300,000 fusions per second using 500 watts and getting the same number of fusion at only or 250 watts or even only 100 watts means zip in the grand scheme of things.
50% or even an 80% reduction in energy for a given output improvement is nothing when only a 99.999999999% improvement would make one really have a reason to smile. It's those order of magnitude improvements which pile on the trailing nines that turn up the smilometer
Now if you could get a billion fusions per second on 1,000 watts input, then that would be a nice improvement but still far, far distant of from any power device. (still 10,000X distant from simple breakeven, which even then will leave you with no real power to send down the wire).
In the ion-ion fusion biz we need a 10 or more order of magnitude improvement to be even remotely power ready. Reducing electron loses to zero is abysmally easy, given enough extra input energy and a much more complex infrasturcture.
The classic, simple fusor remains one of the best neutron sources and least expensive actual fusion systems in the under 500 watt total input power requirement range. It may forever remain one of the most interesting and cost effective of all the well polished fusion turds squeezed out since 1950.
The fusor still intrigues me and I tend to believe nothing will work to a grid level in the next century, unless the lucky donkey effect kicks in!
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
My particular vice is MK2 Jaguars but I only have four ranging from 42 to 46 years old ... gas price in England is currently running at around US$7.56 per US gal but I don't drive them much. My XJ8 uses about the same amount of fuel so I guess it doesn't matter anyway ...
Yes, I take your point about the inconvenient reluctance to fuse. Just have to drive the little perishers harder perhaps?
If one of the papers I read recently (God knows which one, I've read so many!) is correct, the average lifetime of an electron in the centre zone in a 'normal' fusor is less than one circuit: upping this to a few tens would be a start but, as you point out, a very small one.
You have made an interesting logical construction, but one that illustrates some misunderstanding of the physics behind Bussard's fusor.
>The argument that it is easier to confine electrons than ions is flawed, because if you use magnets to confine electrons and then use the electrons to confine ions, then you may as well use the magnets to confine the ions in the first place.<
It's a fact that electrons of some energy E are easier to spatially confine with a magnet than ions of energy E--to make the electrons stay within particular physical bounds, you need much, much weaker fields.
To imagine a fusor built with magnets that "confine the ions in the first place," as per the above, is to imagine a very different physical animal than Bussard's fusor, in important ways not at all analogous. There are magnetic ion-confinement fusion schemes out there, e.g. the Tokamak. Obviously, they're quite different from the fusor and they have some real issues particular to themselves.
In the Bussard fusor, electrons are being confined by electric and magnetic fields in order to (A) avoid the losses they cause when they slam into the anode, and (B) trap a sufficient number of them to form a virtual cathode that will attract a sufficient number of ions with kinetic energy to fuse. The magnetic field's effect on ions is not important to the idea. It's not perpetual motion, it's not absurd, it's not some flavor of "new" physics. It's understandable with the conceptual framework of classical electrodynamics.
Finally, a word about the Lawson criterion. This criterion applies when you're interested in the conditions for thermonuclear ignition--when you have established a dense enough thermalized collection of reactant and product particles to sustain fusion. To the best of my knowledge, this concept has not been considered practically relevant to the IEC devices we're talking about. These do not involve large, dense thermal plasmas that might potentially ignite. IEC people are more concerned about what it takes to do fusion efficiently enough to recover more energy than you put in--a much lower bar for success than the Lawson Criterion, and yet still oh so far beyond us now.
Just on your previous post, electron losses from the grid in a Hirsch fusor or a polywell is always going to be a problem, they naturally want to go to ground.
By immersing the whole guts of the machine in a vacuum chamber with a conductive gas, just makes it inevitable.
I believe that I have solved the electron loss problem with my S.T.A.R. device, where the cathode is surrounded by dielectric oil. If my inital four experiments are correct, and I believe that they are, S.T.A.R. is producing more neutrons with less input power.
Not everyone is convinced yet, but my next experiment will confirm it one way or another.
If I am wrong, and experiment #5 does not produce any neutrons, Richard will say "I told you so!", and if I am right, and S.T.A.R. produces lots of neutrons, Richard will call me "A lucky donkey.."
First - what Carl says about it being easier to magnetically confine an electron versus an ion of equivalent energy, is correct.
Forces generated by magnetic fields are proportional to the cross product of Velocity and Field, and scaled and directed by the magnitude and polarity of Charge, as in
F = Q (V x B). a vector equation.
It is clear from this, that only the velocity, charge and magnetic field determine the force. The effect of the particle's mass enters indirectly, since for equivalent energy U = 1/2 (mV^2 ) the heavier particle has lower velocity. Thus the deuteron will move much more slowly than the electron, being about 7200 times more massive, ( about 85 times slower, in fact).
Now when an electron cloud is contained by a magnetic field, so as to create a potential "well" as in the polywell scheme. things become a bit more complicated, when you begin to add positive ions.
The first thing that happens when positive deuterons are added to the negative well, is that the potential of the well decreases in direct proportion to the amount of positive charge entering. This can be countered by adding more electrons.
So that for some sort of steady state situation, one needs both an electron current and an ion current entering the well, just to maintain a constant well potential.
Since the deuteron has positive charge, it turns in opposite directions to the electron, at a radius of curvature some 85 times larger than that of the electron. Since the radius of curvature of either electron or ion increases as the energy (eV) increases, raising the ion energy and electron energy, requires a matching increase in magetic field to again maintain status quo.
Given the large disparity in mass, however, it NOT at all likely that ions will in fact circle within the magnet structure with a field intensity intended to confine electrons. A magnetic field sufficient to circle the ions, will keep the electron localized so far away that they will have zero effect on the ions. Circling ions have little or no likelihood of fusing.
But the forces between deuterons and electrons are not repulsive byt attractive. So that electrons and deuterons SHOULD seek each other and become nuetralized. Some have opined, that the ions are moving too fast to be neutralized by the electrons, a point that hardly makes sense, since the electrons which are moving about 90 times faster, are in fact contained by the magnetic fields of the polywell.
It does seem likely that the ions should pass right through the electron cloud. What the probability of "collision" will be depends on a pseudo mean free path calculation involving particles that are attracted to each other.
I haven't attempted this yet, but it seems plausible to me that if the electron cloud is dense enough to create a potential well that attracts the ions, it will be dense enough to ensure almost 100% probability of neutralization of the incoming ion.
Whether the neutral atoms will collide with one another and fuse more so than in other configurations, is the big question here.
Neutralization doesn't seem to be a problem to me.
As far as I know, in a fusor the fuel usually is ionised by avalanche breakdown (EDIT: breakdown probably was the wrong word).
This means that a fast electron hitting an atom is more likely to knock out further electrons, than one hitting an ion is to be captured.
In a plasma dense enough for anything near net power fusion I doubt an atom could stay neutral for any significant time.
It appears plausible to me that a "whiffle ball" of fast moving electrons could capture a cloud of hot ions in its center without excessive losses by recombination. The problem with the other loss mechanism, that of the electrons themselves from the magnetic trap, is what Bussard claimed to have solved by design.
This does look like a form of bootstrapping. The magnets shape electron beams which in turn capture ions. The electrons must not collide with the magnets, which apparently precludes the use of simple permanent magnets, e.g. NIB, where the field lines move through the magnet instead of around it. The "whiffle balls" use electromagnets.
Recombination should not be a problem as long as the electrons are fast and the ions relatively slow. In order to recombine, both particles need to have the same speed (or momentum in a center-of-mass system).
Deuterons are 3666x heavier than electrons, a proton has 1835 electron masses.
While it is fairly easy to bend electron beams with small magnets, much stronger fields would be needed to trap deuterons. Bussard’s “Polywell”, if it works, would make the extremely heavy and costly magnet structures of conventional fusion machines unnecessary.
If someone has an idea how permanent magnets could be used, that would make the device much more accessible, on a small experimental scale.
I suppose permanent magnets could be used in something polywell like.
The magnets would have to be placed on the cusp axis behind the electron sources and had to be at ground potential or e-gun potential, so they'd be electrostatically shielded against the electrons.
In place of the magnetic grid there would need to be a normal wiregrid at a positive voltage.
However if the magnets were to far away from the grid, it wouldn't be shielded properly and would only allow a small plasma density. On the other hand if they were to close, the potential on the cusp axis would drop to low, allowing more ions to leak out of the well along the cusp axis.
So the question is whether there is a right spot at all.
Buildling that still wouldn't do it though. Like in a real polywell, there would need to be a way to insert a small, controlled amount of fuel on the inside of the wiregrid. If the fuel is ionised outside it, it goes to the chamber wall and produces a lose current. Also too much fuel would release more electrons than the magnetic field could hold.
A possiblity might be a second vacuum chamber for the fuel from where a pipe would lead to the grid.
Sorry about the slip up on Deuteron to electron mass ratios. Been doing a lot with He ions, lately and absentmindedly used those ratios. So the velocity ratios and radii of curvature for the same energies and magnetic fields are about 60 to one instead of 85 to one.
If someone is attempting to do something specific and succeeds at it They are not lucky donkeys. Lucky donkeys are those who, like Becquerel and Roentgen, were looking for something specific, but discovered something far off the beaten path. They both failed at their original quest and ideas, but unlocked totally new physics, (stumbled or blundered into it). That is the lucky donkey scenario.
The guy who breaks fusion might be looking for an ion gas piston design or some such dream.
Finally, the polywell should not trap or build up ions in the center of the polywell at all. The deuterons will ideally whiz through the electron knot (negative well) and recirculate. There should be zero ions in the center in the form of trapped ions. The verbage is important, I think. The ploywell accelerates ions towards the negative well where they either fuse or zip through to recirculate just as in a grided fusor. I got the distinct impression from some of the foregoing that some thought we were trapping ions somehow within the polywell.
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Indeed my understanding of the polywell is, that ions are to be trapped inside of the magnetic grid.
Do I understand you right, that the ions would pass through the magnetic grid?
If that indeed is how a polywell is supposed to work, than I totally agree with you, it couldn't possibly be much of an improvement over a normal fusor. Ions passing to the outside of the grid would immediately be lost to the chamber wall.
However since the polywell seems to make so much less sense that way, than in the way I understand it at the moment, I don't think this is how it is supposed to work.
As I see it, the fuel is entered between the (positive charged) MA-grid and the electron cloud / 'wiffle ball', thus at a potential lower than the MA-grid potential.
Idealy the fuel is ionised by the fast electrons when it comes in contact with the 'wiffle ball', so the ions have nearly no kinetic energy at a potential which should be clearly below the MA-grid potential.
This leaves two ways for the ions to leave the well (as long as the electron trapping holds and the electrons don't cool to far down).
One way is upscattering, the other is the potential at the point where the axis crosses the grid being pulled down, either by electrons accumulating there or by something at low potential close to it (like an electrostatically shielded permanent magnet).
The whole point of my discussion above the correction post, was just that. Magnetic fields sufficient to confine electrons to make the well, are far too small to also confine the ions. Field strong enough to confine both, would leave them widely separated, so there would never be the density to fuse.
Attempts to use magnetic fields to aid in the confinement of ions have been made a number of times. A rather impressive machine was build by Wilson Greatbatch, otherwise inventor of the cardiac pacemaker. The Greatbatch machine could work with permanent magnets, mounted outside of the vacuum chamber.
Of course, none of these experiments come anywhere near the point of possible energetic breakeven. The machines from Greatbatch, Nebel's POPS and Bussard's Polywell are all studies of technical possibilities, and maybe all are fatally flawed. Considering that they cost very little, at least relative to the politically more ambitious projects, it would be it very unwise not to follow up on these ideas. Be it for the sake of learning what can't be done.
It's quite possible that a successful solution may be found in the future, by someone with an open mind and working in a field completely different from fusion. Probably not by potluck alone, since it takes an high amount of perseverance to reach such a goal.
The missing component is not new physics, it's intuition.
The magnetic field only serves to confine the electrons.
It isn't supposed to confine the ions, the electrical field produced by the electrons and the grid will to do that.
If the fuel is ionised inside of the grid at an electrical potential below grid potential, the ions simply won't have the kinetic energy to leave.
(Also the electrons knocked out of the fuel won't have the energy to produce a real potential well until they get heated by the electron beams from the e-guns.)
I hope all who are talking about apples are not in conflict through verbage with those who seem to be talking oranges.
What is the magnetic grid? Are these the coils? if so, they are not a grid at all they are the outside shell effectively of the machine. Within the coils in the potential well, a much smaller knoted region of electrons (theoretically) that are to attract and not trap ions. All the deuterons (ions) are genrated outside the potential well either by high field emission, ion guns or filaments.
The ions infall to hopefully fuse. Those that don't will zip thorugh as in a normal grided fusor and supposedly go for another pass. They are trapped by the electrostatic field within the fusor itself and not the potential well.
I don't see how deuterons trapped in the potential well would ever fuse, unless we are talking a target machine that you could as easily make with a linear accelerator and deuterium target.
The only thing I see different in the polywell device and the common fusor is zero grid loss and supposedly zero electron loss and I seriously doubt the latter. Otherwise, this is just a farnsworth fusor with half the losses removed (in theory), but with no effective gain in fusion.
Any deuterons trapped in the polywell would not be energetic! Would not fuse and would ultimately neutralize.
I saw the various iterations of this thing and the grids were external on the spherical chamber. Once they left a spherical chamber and had a giant tank, the outer shell would still have to be a grid of sorts just inside the coils.
The coils will certainly not trap the ions as Dave has already noted.
Someone needs to define where the ions are created and how and where. Where is the ion trapping electrostatic field established? Between what and what. The assumption is the negative limit is the potential well or electron knot of high density. where is the positive shell or origin of the ions to establish a fixed positive point, which would be at some zero or ground reference to the electron knot. This knot mimes the inner grid in our machines..
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Yes the magnetic grid refers to the coils, I prefer the term grid because I'm not sure whether some other geometries that might be used, could still be called coils.
Anyway one can look at them as the shell of the actual machine.
With the potential well I was refering to most/all of the space inside the coils, basically everywhere were the potential gets smaller when moving towards the center. So when you were speaking of recirculating ions and I of ions trapped in the well we meant the same thing.
Pardon my ignorance, but what are the other 50% losses besides grid and electron? I was under the impression in a polywell electron loss would dominate, unless ion density was low compared to neutral density.
Did I overlook a loss channel?
OMG, I meant to provoke a mild discussion ... Oh well, all discussion is good!
I don't wish to enter discussion with Richard on what constitutes 'the machine' ... some might look at the core and others at the overall effect.
As I understand it the position is as follows:
Farnsworth notes that it might be a bright idea if ions (of whatever fuel) were encouraged to the centre of an electrostatic device to collide with each other (probability = small). The big problem with a simple two electrode device (albeit spherical) is that the ions have other ideas and like to whack into the cathode instead of each other. Recirculation of ions within the machine is thus severely limited.
Elmore, Tuck & Watson observe that it is possible to create a virtual cathode with increased electron density in the centre of a reversed polarity two electrode machine (with an outer cathode and an inner anode) thus creating a three electrode machine (with a virtual cathode). This means that ions contained within the anode cannot strike an actual cathode but only (possibly) an electron which is part of the virtual cathode. Unfortunately the electrons which create the virtual cathode have a nasty tendency to impact on the anode grid which attracted them initially (which is quite understandable given the relative charges/potentials).
Bussard notes that electrons are very susceptible to magnetic fields and will tend to shear away from them. Bussard suggests 'magnetic shielding' of the anode in the ELT machine configuration to prevent electron losses to high potentials. For some reason Bussard suggests that the magnetic shielding can also be used to confine the electrons in the virtual cathode region (apologies Carl W.)
The original question remains: is it a requirement that the electrons are magnetically confined or is it sufficient that they are discouraged from impacting with the anode grid?
To the original question, confinement seems to be a requirement to me.
First there is always some unshielded surface, if the electrons are confined to the center there are less electron losses.
Second for meaningful power incredible ion and electron densitys are required. If the electrons aren't confined they will pull the potential on the coil faces far enough down to let the ions go outside. Once there the ions will be attracted to the vacuum chamber wall and the electron guns and since their density is incredible so will be the loss.
We have to clearly define an anode in this device. A place where the OTHER power lead goes to. The cathode is being called the electron knot in the center which replaces the grid wire structure in the simple fusor. Fine. No problem here.
Where is the anode? What is the anode structure? To what do we connect the other power lead. In the simple fusor it is classically the SS shell or body of the device.
In a giant cylindrical tank, one would assume the anode would be a wire mesh grid at least! The magnetic coils would be just outside this large grid. If a sphere, the coils would be just outside the metal sphere's body.
However, if we are to inject electrons via a gun how would they get to the cathode potential well as this gun would have to reside out near the anode???
This is a mechanical thing. How do you impliment this in a real world? The norm would be as I noted above a regular fusor with and anode shell, external coils and normal high field ionization where ions and electrons are created all over the device but are mag focused or herded into the center magnetically to form the well.
Ion guns could be used to supply directed ions out near the anode perfectly with a mild bias.
We have to consider the ion current. For any ion current there is an equivalent electron current. No way to get around that one. Do we just squirt in x amount of ions and x amount of electrons and just let the thing run? Not even! What are the new realistic losses. Where do they come from and where do they go? I leave this as an exercise to the student. (Didn't you just hate that in college - Yeah, it forced you to think for once)
I worry that many in the theoretical discussion may not pack th' gear to have physically implimented an ion-electron circuit which is demanded of real world devices.
I have no theoretical issues with the elimination of electron losses within the polywell. It is more with the physical implimentation that I am concerned with, based on some of these discussions. Some seem to think there are ions present and there are electrons present, as if by magic.
Such free, charged particles are all part of an electrical circuit with a source and sink. Beyond the realization that there is a source and sink and that there is a perpetual circuit and current, there is a question as to how to implement the mechanical embodiment so that theory can be made reality.
The source and sink are just a must have mindset and part of any theoretical embodiment.
The physical embodiment of the wheelwork are where the fly is located in all such theoretical ointments.
I do believe that good ole' Carl Willis, Dave Cooper and myself have already voiced many of these concerns earlier in other polywell related posts.
This latter item is what stops perpetual motion machines and apparently what has always kicked thermal machines in the groin.
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Perhaps I've missed something in the construction of Bussard's poywell device. The pictures show a polyhedral assembly of solenoidal coils that actually are more like thick rings, than solenoids. They have been called "toroids", but the windings are all oriented around an axis perpendicular to the plane of coils... which makes the field somewhat solenoidal as opposed to toroidal (or around the perpendicular axis).
The array of these coils (how many, I can't recall) is claimed to produce a closed magentic surface than will trap electrons, once they enter through one of the larger openings. It is part of the inner assembly, we are used to calling the "inner grid".
This is claimed through arm waving and some simulations to create a potential well of many kV.
One has to sit down and walk through the trajectories of some incoming electrons to see how this form of magnetic containment functions. It does NOT create a microscopic point in the center of high electron density and consequent high potential. That sort of concept requires stationary electrons. The concept of this device requires an electron trajectory reversal, so that the change in velocity creates a dynamic electron concentration. This would be the place, that would have a coulombic attraction to an inbound positive ion.
Depending on the exact electron flow within the polywell zone, there will be some sort of effective "concentration" of negative charge - hence the "potential well". Since the electrons will have position dependent velocities, dynamic charge densities will vary all around the interior. Only if a majority of the electron orbits pass through the center will the center be the high negative potential zone.
If you look carefully at Bussard's data and simulations you will realize this "well" is rather broad. Some simple calculations will indicate the magnitude of charge it must contain to have the potential depth needed.
Bussard does some circulating current estimations to indicate how much charge be in motion. I think he was talking about 100's of amperes..
As such the potential well structure resembles a simple electrodeless capacitor (or is it an inductor?) , of probably some fraction of a uF (or Hy) . The equivalent inductor idea runs into difficulties since the electron paths are not constrained to any one trajectory. The net current in some direction could well be close to zero... so.. an equivalent capacitor is the most likely component representation.
This is what is claimed to be the attractor or cathode for the positive deuterium ions.
To repeat my point above...a magnetic field sufficient to create the cathode condition is vastly insufficient to also contain the ions. A field strong enough to also produce closed ion orbits, would effectively isolate the two charges, creating two counter orbiting clouds of opposite polarity. Interesting, but not particularly useful for fusion.
What some do not seem to realize, is that in the simple Farnsworth IEC fusor, the inner grid emits electrons first by secondary emission from ion bombardment, and only after it has reached temperatures above about 2500K - very white heat-will it become a significant thermionic emitter . The original ions are created far out toward the fusor shell by the usual methods of gas ionization, at these pressures.
The ions thus created, travel inward. Many hit the grids, and cause additional... upwards of 2 or more electrons for every ions impact - while the rest pass on through the grid and out the other side.
Somewhere on this route, an occasional deuteron collides with an deuterium/or deuteron and produces fusion products. Since this does not happen very often, the yield as a percentage of total ions flown is very low
If the Bussard configuration were 100% effective in creating a lossless electron well, the only current it would draw, once the well was established, would the current of ion neutralization and fusion.
Predicting how big this current would be depends on your view on ion-electron probabilities inside the inner grid. Ignoring the magnet power, (since they could be replaced by permanent magnet assemblies), the device then resembles a simple capacitor with a bit of leakage, from a strictly electrical standpoint. The neutron & proton flux would account for some fraction of this presumably small leakage current.
Clearly, the polywell concept presumes a number of processes, which while feasible, may not perform the assumed functions in the device.
As Dave points out, in the end, 'herein lay the rub'
Don't look for a smallish polywell machine at all and don't look to any successful prototype producing much more fusion per watt spent than a simple fusor. The size of even a medium sized successful prototype obviates this.
Fusion's issues are strictly 'failure to fuse' due to rotten probablistic collisions in a low density environment. Increase the density, coupled with increased fusion, and you are back to all the issues with a thermal device!
The way I look at fusion is that it is promised, but non-extant potential energy on the other side of the reaction. It is not making use of real, extant potential energy, but instead, the process of fusion might be termed, the promise of extant potential for potential energy.
It is like robbing Peter to pay Paul. Sort of a balloon credit scheme.
Not at all like fission or burning wood or other such true potential energy releases where the potential energy is real and extant just waiting for a trigger to release it.
The fusion we see done in stars is pretty much reliant on the potential energy exchanges between colombic and gravitational forces within matter and not due to the good offices of any external forces. We, on the other hand, are trying to make fusion happen using only external forces in all out schemes. Sort of backwards, if you will. We can certainly do this, but not to a net power level yet noted.
Stellar fusion is an exchange process we can't initiate or mechanize on the small.
I have always felt that Jupiter was fusing at some subliminal level to pour out more energy than it receives.
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
The coils are the anode, cathodes are the electron guns, the shell of the device and (in a way) the electron cloud formed inside of the coils by the electrons.
Apart from the 'virtual cathode' in the center these are sinks respectively sources.
A big high voltage DC power supply is hooked up between the coils and the electron guns. The device shell is grounded and the electron guns are either grounded too or ideally biased to a potential slightly above ground.
Electrons starting from the e-guns should very likely go straight into the magnetic trap.
This is because the e-gun are placed on the axis, near them the electrical field points directly towards the center of the nearest coil. When the electrons get closer to the coil where the electrical field pulls stronger away from the axis they already have a substantial velocity towards the center and are in a strong magnetic field also pointing to the center. They should have a very good chance of not hitting the coils but entering the trap.
However electrons escaping the trapping are not likely to come out so nicely focused. If the device is good engineered most of them should return into the trap anyway, the coils are shielded by the magnetic fields and they won't have the energy to reach the shell unless they were upscattered. Still some will hit the coils and cause a big loss current for the power supply to struggle with.
Now to the fuel. In the simplest case it is just injected behind the coils and ionised by the fast electrons confined there.
The electrons knocked away from the fuel get confined by the magnetic field and heated by the other electrons via collision.
Eventually they manage to leave the trapping and hit the coils, they then need to be replaced by electrons from the e-guns. They also could recombine with ions so they wouldn't directly produce a loss current but still carry energy away, energy which was introduced to the system by the electron guns.
The ions could leave the confinement via upscattering and would thus (together with the electrons with which they were entered into the machine and which went to the coils) cause a loss current from the device wall and the e-guns to the coils.
Ideally the ions will stay confined until they fuse, the positive charged fusion products would have enough energy to leave the confinement. They would either hit the device shell or the electron guns, also producing a loss current or the coils in which case they cause no current.
In summary:
There is a big electron current from the electron guns into the magnetic trap heating the confined electrons inside it.
This electron current then goes from the inside of the magnetic trap to the coils.
An ion current proportional to the number of ions escaping or fusing goes from where the fuel is ionised to wherever the ions end up (mostly the shell). An aquivalent electron current goes from where the fuel is ionised to the coils.
Now as a side note, I don't think the electrons would form a small knot in the center. Since we want to have as many ions as possible there would also be as many electrons as possible, which means adding fuel until the magnetic trap is about to blow out. At that point electrons would be all over the space inside of the coils.
Of course the physical implementation is the problem, even if everything goes according to theory the engineering might not be feasible or feasible but totally uneconomical.
However I don't see the same fundamental limits as in the common fusor, where part of the losses are (it seems) proportional to the ion density.
I know a lot of loss channels weren't mentioned in this thread, but I think the most important ones in respect to scaling the machine up were considered and I can't see why a polywell shouldn't be able to get a better gain than twice that of a common fusor.
