Building a WB-2 Polywell

Discuss the technical details of an "open source" community-driven design of a polywell reactor.

Moderators: tonybarry, MSimon

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

Sparking aka arcing is in no way beneficial and is bad in a number of ways.

Firstly, it can damage the equipment - the chamber and the magrid. And if your magrid coils fry, have fun winding a new one.

Secondly, it's a huge energy loss. Forget net power if theres an arc. The whole hunk becomes as useful as a rock.

Thirdly, it destabilizes confinement. Any fusion would cease immediately. sparking is not harmful to you physically, provided the chamber is well grounded, but as far as the machine goes its a show-stopper.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

happyjack27 wrote:If the charge of the whole system doesn't match that of the chamber, that will create a voltage gradient that will induce a current. To maintain that voltage gradient, you have to constantly pump in energy equal to the current. If you don't do this, then the system becomes net neutral relative to the chamber, like I said.

I don't see the benefit of having a voltage gradient on the chamber, unless you're using it as an electron source (not a bad idea) but that would mean a net positive system.
Talking about the charges in a Polywell is complex. Like you said , the confined plasma needs to be net negative, this creats the potential well and electrostatically confines the slightly fewer positive ions within the magrid. This net negative charge will flow to ground, and do so very rapidly for the electrons. they are faster than the ions at the same energyand feeds into discussions about Debye Length, etc. The key of course is the cusp confinement and then the Wiffleball confinement and then finally recirculation that slows this electron current to acceptable limits. The "ground" for these electrons is actually biased positive surface of the magrid. This positive charged electron "ground" is what accelerates the electrons. You can actually have the magrid at system ground - Earth Ground, the dynamics are the same, though I don't know of any tests in this mode, just that Bussard said either approach is equivalent. When the vacuum vessel case, standoffs, ion gun cases, e- gun potential is considered it becomes more complex, now add in direct conversion grids. Insulation is also critical , in this case magnetic insulation.
From a funtional viewpoint the magrid is charged positive, and with inneficiencies, this positive charge has the greatest magnitude. The case is Earth grounded. The contained plasma inside the magrid where the action takes place is negative. The plasma outside the magrids is mostly electrons and neutrals, so this area is net negative. The electrons want to go to the positive charged magrid, but have a hard time due to the magnetic insulation, their eventually likely to hit the case due to scattering , etc. I don't know the relative currents that reach the magrid or recirculate back into the Wiffleball, or the case. Obvously when arcing begins, this trend reverses and most of the electron current will flow to the case. This current can becom so great that the voltage drops- potential well dissipates very quickly.

There is no simple answer, it depends on which part of the machine you are referring to and to the density of the charge carriers.

A key is to remember that the environment is different inside and outside of the magrid. The electrons and ions inside the magrid see each other and interact locally and through space charge effects, but they do not see the positive charge on the magrid due to Gauss Law effects. The contained charged particles only see the magrid once their momentum carries them outside the magrid. Now the electrons are attracted towards the pos magrid, and the pos. ions are repelled towards the Earth ground case (which they do not see untill they hit, again due to Gauss's Law. This 'simple' interaction id further complicated when other electrodes, supports, etc are introduced.

Sometimes I like to just use a wire analogy. The current flows through the wire based on it's resistance and the applied voltage. The magnetic confinement of electrons equates to a high resistance in the wire. Little current is lost, and the potential can be easily maintained. The arcing comes from a close by parallel wire. When the insulation breaks down (dependent on the gas density) the current will short to this low resistance parallel wire and pass to ground at high current due to the low resistance in this second wire. The system shorts out. Just like a wire, the positive charge carriers are trapped (by the potential well). So they do not contribute much to the current so long as the potential well is maintained and upscattering is not to bad. This upscattering is an important issue as taken to extreams it can allow the pos. ions to escape almost as fast as the electrons and this is bad. I don't know the numbers, but my impression is that while the Wiffleball is maintained, the ion current is only ~0.1 to 10% of the electron current. This is why electron losses is the dominate loss mechanism.

Dan Tibbets
To error is human... and I'm very human.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

Lelephant wrote:First of all, general questions (some are meant to be for D Tibbets but I can't remember which):

1. Wouldn't physically running a wire into the vacuum chamber to power the coils compromise the vacuum of a belljar vacuum chamber?
2. How to keep my magnets cool and how to measure their heat?
3. How can I calculate emission rates of x-rays, neutrons, etc?
4. How do I detect aneutronic fusion?
5. Is aneutronic possible with permanent magnets and WB-1 or WB-2 configuration?
5. Is D2 fusion possible with permanent magnets and WB-1 configuration?
6. Do I bias the cage of permanent magnets? Why? To what degree?
7. How do I measure the potential well and electron to ion ratio?
8. What proportions should I maintain?
9. Permanent magnets vs. coils?

Mattman:
mattman wrote: IDK. You may need some electron emitters to get a small electron cloud going first to attract D2. Electron emission could be done by heating up a wire next to a voltage drop. This is called thermionic emission. You could then probably just release the D2 gas straight in. More electrons would come right off the D2 itself, when the D2 becomes an ion and loses its electrons. If you get a spark, you have too much electrons and D2 in the chamber. 
1. What sort of voltage drop would I use and how would I regulate it to emit the correct quantity of electrons?
2. Would I shut it off just by turning off the power going toward the voltage drop?
3. How is the D2 ionizing in your explanation?
4. Why would a spark mean I have too many electrons as well as too much D2?
mattman wrote:Teflon seems to be the material of choice for the "first timers", Both Joe Khachan and Mark Suppes built small (about the size of a coffee cup) devices from Teflon discs with wire spun around them. Joe screwed L brackets into his discs to hold them together and probably attached the rings to an aluminum bar. His device pushed apart with ~0.2 Newtons of force - barely any repulsive force. You can read this: http://thepolywellblog.blogspot.com/201 ... sults.html for analysis. Teflon gets brown and can be a pain to degas because of gas pockets stuck in there that take forever to leave. 
5. Why would he attach them to an aluminum bar?
6. Why is repulsive force important and how can I measure the repulsive force of my rings?
mattman wrote: Yes - sparking. Watch out for sparking. Mainly from the outside cage to the rings. You will need neutron detectors and the rumor is they need to be shielded (allot) or you will get false counts. Especially when you are just starting out, the machine you build won’t give off much neutron signal compared to the neutron background noise.
7. How can I calculate if sparking will occur and prevent it?
8. Why is sparking bad?
9. Could sparking be beneficial to fusion?

D Tibbets:
D Tibbets wrote: Providing power is simple. A wire connection. That said, the magnets will need a case that seals the magnet wire inside. The connecting wire entrance also needs to be sealed. I can speak from experience that hot melt glue is bad. A slow set epoxy may work, at least for prototyping. The magnet needs to be sealed due to outgassing concerns. You cannot have just a role of wire held together with wire ( note though that some have done just that) Short run times and low concern about outgassing contamination, allows this. 
1. How do I know I'm preventing outgassing and how can I measure outgassing?
2. How can I make sure the wire entrance is sealed and that the case is sealed?
3. Are there any pre-made, purchasable cases (where can I find cases)?
4. Why can't I have the coils exposed at such low operating levels if outgassing is a minor concern?
D Tibbets wrote:
Actually, net neutral is the last thing you want. It is nit picking, but one of the essential properties is that there is not net neutrality. There must be an excess of electrons. Granted this bias is only ~ 1 part per million, but still essential. With gas puffing ion source this bias is naturally maintained, mostly due to less than 100% ionization efficiencecy. Ionization is almost totally due to electron collisions, and the ionizatios adds corresponding stripped off electrons to the mix, thus does not change the ion/ electron ratio. You need to flood the magrid with enough neutral gas molecules that the ionization process is saturated- in order to get the maximum internal density/ Wiffleball. Not too much though, as this shortens the time to the beginning of arcing. This might benefit from precision but it is not absolutely essential, there is a considerable amount of wiggle room as demonstrated by the simple apparatus used in WB6. 

With ion guns, the system may need active and precise control of the ion flux. 
5. Could I detect fusion in a small model like the WB-2 by simply releasing D2 into the chamber?
6. How will it ionize by itself?
7. How can I make sure/calculate whether or not (or how much of it) will ionize?
8. Why are ion guns preferred if D2 will ionize anyway?
9. How can I actively and precisely control ion flux and compare it to other measurements in the system?
10. What proportions should I maintain?

Happyjack27:
happyjack27 wrote: Make sure it's off and any capacitors are fully drained. Rubber gloves are unfashoinable, but so is baldness. 
1. Do I need capacitors? Why?
happyjack27 wrote: Its not a safety hazard, but you do have to balance the pressure with the electromagnetic force. That is key to the whole thing. That's "b=1". Meaning you have to measure the gas loss through the vacuum pump very precisely and inject enough to keep the net pressure in the chamber at a constant proportion to the field strength. 
2. How do I do/measure all this?
3. What proportion should I maintain?
So many questions.

You should really go to a vacuum forum- "The Bell Jar " search sould find one. Then go to the "Fusor.Net" (or is that Org?). site and work through the frequently asked questions.
You can pass a wire through or under a Bell Jar but it is difficult. Usually they pass the wires, ports, etc. through the base plate.

The WB2 was samal0, and I don't know if EMC2 managed to detect neutrons. WB3 was a scalled up version, and then WB4 did detect D-D fusions by counting neutrons. The ~ 1000 fold gain with WB6 was due to round coils and spacing. In your vision (?) a small, weaker magnetic field would perform better than WB2-4, but where the threshold for detecting fusion through neutron counting becomes possible. It depends of the sensitivity and setup of the counter. A typical electronic neutron may detect ~ 1/1000 neutrons that hit it, and the surface area may be only ~ 1 percent of the total surface area of the shell, so neutron fluxes approaching ~ 100,000 neutrons per second may be required. Perhaps very good/ expencive counters could do better. Also, thes fluxes are the output per second. If you run the machine for only 0.1 secnds, the total counts will be 10 times less.

As for detecting fusion, neutron counting is relatively easy. The neutrons leave the chamber and can be detected by several means. Depending on the fuel, charged particals like protons, alphas, He3 nuclei and tritium nuclei may be produced. Thes can be detected with a simple Geiger counter, but the particles are not very penitrating so the sensor head has to be within the vacuum chamber and this causes all sort of problems. It has been done by small operations, but it is a step up in complexity and problem solving compared to neutron counting.

The P=B11reaction may produce a gamma ray in a branch reaction about 1/10,000 of the time. This could easily be detected with an external Geiger counter. The problem is that the reaction rates for P-B11 fusion in current experimental machines (such as WB8 or FF! (DPF)) will probably be several orders of magnitude less than corresponding D-D fusion. Since the expected gamma ray flux would be 1/10000 less than this, the S/N ratio would be perhaps ~ 1 million times lower than D-D fusion.
And, forget trying to detect excess heat. You might get a few Watts of fusion heating, with ~ 1 million or more Watts of input heating.

Now a WB20 * may be a different story. With a 50 cm radius, and a B field of 2 Tesla, the scaling would predict a fusion output ~ 4.5 million times higher. Now the fusion thermal output may be very many thousands of Watts compared to an input of ~ 10-20 million Watts. Now you are cooking. You are also getting very nervous about radiation shielding.

* WB20 is an imaginary machine that reflects the scaling laws.


Dan Tibbets
To error is human... and I'm very human.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Post by mattman »

1. What sort of voltage drop would I use and how would I regulate it to emit the correct quantity of electrons?

I do not know. You can buy electron guns - http://www.kimballphysics.com/ or http://www.ebay.com/. Thermionic emission works by heating up material. The amount needed to heat is measured by a metals work function. Metal has lots of extra electrons so it is good for this. Bussard put the rings at 12.5 kilovolts positive from the electron emitter. If he used this emission he probably had a thin wire he was passing current through it. This heated up the wire. Electrons got kicked off the wire. The electrons then “fell” down the 12.5 kilovolts towards the rings. The electrons were then “caught” by the ring fields. You can measure the amount of electrons in the center using a Langmuir probe – that is what the U of Sydney did. They just stuck the probe in the center. You can probably buy a probe off e-bay - some may come with connections to multimeters.

If it were me, I would get the probe, get a wire, and try to generate an e-beam between a high and low voltage source in a bell jar. This is kind of like measuring the voltage of the e-beam in an old CRT television. It’s a good first test system. It is also well understood, meaning it’s a good place to learn. You can then play games with the current in the wire and the voltage and measure the number of electrons using your probe.

2. Would I shut it off just by turning off the power going toward the voltage drop?

I don’t know. Turning off the voltage means you are stop the flow of fresh electrons towards the rings. But, the magnetic fields are still on. This means whatever electrons were already there - are still flying around. They ride the fields like cars on a highway, back and forth into the center. We know from U of Sydney’s December paper that – for low numbers of electrons - they won’t stay in there forever. The paper estimates about 1 microsecond for them 1400 of them to “evaporate”. What is exciting is when we get lots of electrons, new behaviors may emerge – the Whiffleball, virtual anode, ect… but we do not have any work on this yet, so who knows...

3. How is the D2 ionizing in your explanation?

It is Bussards explanation. It is from his paper: “The Advent of Clean Nuclear Fusion: Superperformance Space Power and Propulsion”, page 11 & 12. It takes ~14-16 eV of energy to ionize D2. That’s nothing. An electron coming from Bussards’ wire into the rings could have up to 12,500 eV. When the electron hits the D2, it gets hot; hotter than 16 eV. The D2 vibrates and kicks off an electron. The D2 is now an ion. Remember the ion is ~1,800 more massive than an electron. We need an ion collision of ~10,000 eV to start getting fusion – so we are in the ballpark.

4. Why would a spark mean I have too many electrons as well as too much D2?

Since the electrons comes from your wire, and the D2 cloud - then sparking means you probably have both mechanisms running too high. Bussard listed arching as a major problem at the end of his Google presentation. Here is good reference from NASA which predicts sparking for different metals, voltages, ect: http://misspiggy.gsfc.nasa.gov/tva/mark ... spcing.pdf Predicting sparking was pretty hard to research, most of the work was done in the 20’s, 30’s and 40’s.

5. Why would he attach them to an aluminum bar?

Good Question. IDK for sure. It was hopefully aluminum. Aluminum has a magnetic permeability close to vacuum; so from the magnetic fields point of view, the bar is barely there. Aluminum also conducts electricity – but if the bar was held at a uniform voltage, the electrons should have no reason to leak through the bar. The L-joints and screws have similar properties. In U of Syndeys’ first work they were merely trying to experimentally prove the fields could hold electrons, so they could do things you would not do in the final machine.

6. Why is repulsive force important and how can I measure the repulsive force of my rings?

It is not critical. It is the force two magnets feel when facing each other. I just wanted to double check that the rings would not fly apart. You can estimate it using http://en.wikipedia.org/wiki/Magnet#The ... c_Currents

7. How can I calculate if sparking will occur and prevent it?

IDK. Sparking is hard to nail down. I got the books: Gaseous Conductors and The Electric Arch both from the late 50’s. This was a nebulous topic to research. The NASA paper provided a solid first pass equation to predict sparking – and it predicted it would be a problem.
8. Why is sparking bad?
Other people have answered this.

9. Could sparking be beneficial to fusion?

IDK.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Post by mattman »

Question:

How can the D2 be an ion flying into device center at 12,500 eV, if it needs to be inside the device to become ionized?

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

Mattman, it appears that your research is surpassing my hobbist level, but to answer some of your questions further,:

Power supplies whether home built or commercial can grow to great complexity. There are ways to control the current and the voltage. The costs grow accordingly. Charging up capacitors and then feeding the current is the simplest high capacity system. Bu the only current control is the resistance of the system. The voltage can be maintained until the capacitors drain to a certain level. This is why the current in WB6 climed to ~ 45 amps while the Wiffleball condition was met, but once the magrid external density climbed past a few Microns of pressure and gas discharge/ arcing commenced the current went up to a few thousand amps. A regulated power supply may have limited the currant gain, but I don't know if the potential well voltage could be maintained, even if the voltage to the magrid was controlled. The potential well is created by trapping high energy electrons efficiently. If the trapping fails through a low resistance pathway to ground, he potential well drops despite maintaining the potential on the magrid electrode. It is complicated and confusing for me. My understanding is that the potential well is created through the improving confinement of the electrons and their injection energy. Both conditions need to be maintained for the potential well to persist. How the arcing effects the electrode potential is straight forward for an electrical engineer (not me), and how arcing effects electron confinement directly is also a mystery.

The WB 6 electron guns were thermionic emitters- hot auto light filaments . During Wiffleball operation the current from the filaments was ~ 45 W. This current was limited by the resistance represented by the electron confinement. The "back pressure" represented by the confined electrons limited the current . This implies that most of the filament (E-gun) current was entering the Magrid confined Wiffleball and then leaking out after ~ 100,000 passes. When arcing started, there was now a second path to ground (vacuum vessel wall or Faraday cage surface). This second pathway (based on Pashin breakdown ) was very low resistance compared to the Wiffleball mediated path to ground.

You might be able to maintain the accelerating voltage with a very strong power supply, but the energy flow would be too high for any possibility of Q's over one. I suspect that secondary effects of the flood of mobile charge carriers (electrons and ions) modifies the effectiveness of cusp confinement. So there may be two effects of arcing. The massive increased load on the power supply and it's ability to maintain voltage; and modification of electron confinement (for the worse).

Turning off the voltage obvously turns off the machine. Depending on how fast the confined charge drains off determines the lag time and a few milliseconds mentioned is reasonable and consistent with the electron confinement time. Turning off the electron guns and/ or (?) the ion guns would do the same thing, though perhaps not quite as fast as the e-gun wires have to cool off. Alternately, shorting out the resistance represented by the Wiffleball confinement may shut down the system quicker- turning off the magnets or flooding the system with inert gas. By playing with the voltage, magnetic field strength (and thus the moment to moment Beta condition) you may be able to not only shut down the system very quickly (no more than a few milliseconds), but also throttle the system over a broad range.

It takes ~ a dozen eV to ionize hydrogen. This is the base line. It takes more to ionize quickly and efficiently. Space charge is very inefficient in ionization. Microwaves and needle electrodes can help, but by far most of the ionizations occur by collisions between electrons. Just like the fusion cross section curves, there are ionization cross section curves. For hydrogen this peaks at ~ 100 eV. This means that so long as you have electrons above this energy (accelerated electrons from the e-guns) the secondary electrons- those knocked off of the hydrogen molecules/ atoms will have an energy averaging around 100 eV. These secondary electrons at this energy then can efficiently knock of further bound electrons. There is a cascade that will continue until almost all of the available gas molecules are consumed, or the voltage drops. Due to head scratching interactions of confinement times and e-Gun current, this process can be supported and the secondary electrons can even be further heated to near the temperature/ energy of the injected electrons. So it is a three step process. The introduced high energy electrons, produce many secondary electrons through collisions and these secondary electrons have an initial energy of ~ 100 eV due to the ionization cross section curve. This allows for cascading further ionizations due to the favorable energy of the multiplying secondary electrons. A single 10,000 eV electron may ionize - release ~ 100 electrons at an average energy of ~ 100 eV. This mixture of low energy electrons are then heated by additional high energy electrons. Thus with good confinement the the high energy e- gun provided electrons (accelerated by the magrid potential) can internally ionize and heat a large neutral gas population.Time (the ionization process is not instantaneous), size (in association with time), and confinement efficiency are the critical factors. Using ion guns instead of gas puffers changes the dynamics considerably, but the final condition remains the same.

As for the ion energy. I think the original ionization of the neutral gas results in an ion also of ~ 100 eV (perhaps the original 10,000 eV electron creats ~ 50 ion + free electron pairs for a total of ~100 charged particles per high energy electron). Further ion- electron collisions are not very efficient at accelerating the ions. The ions achieve their high energy by falling down the potential well created by the excess high energy electrons confined by the Wiffleball state.

As for aluminum bar, etc considerations, I don't know the reasoning. I suspect it is not so much that the magnetic fields do not effect them much, but that the aluminum bars do not effect the magnetic fields much. The magnetic field is not distorted or dampened by the aluminum. . The cusp morphology that is the critical component of the electron confinement is different from the magnetic shielding of structures. The shielding is mostly a consideration of ExB drift and this is dependent on the gyror adius of the charged particle and collision characteristics of the plasma. This is a major difference between tokamaks and Polywells. Bussard's abandonment of magnetic shielding for the ions means the magnetic shielding (ExB) only needs to handle the gyro radii of the electrons. This allows for much smaller size and greater density(?). The subsequent scheme of ion confinement through electrostatic forces from the excess electrons is the major difference. Of course the Polywell has cusps and these are tremendous holes in the magnetic confinement. This is where the secondary consideration of improving magnetic confinement of the electrons through Wiffleball effect and recirculation becomes paramount. This results in confinement still much less than the cuspless tokamak (ignoring macroscopic instabilities), but once density considerations are added in, the advantage goes to the Polywell . What the Polywell lacks in raw confinement , it more than makes up for in achievable density.

Arcing due to too much ionizable gas and/ or charged particles is based on two factors. The Pashin breakdown at different pressures and the insulation against current flow. The charged particle are the charge carriers. They will ground the system. But, insulation will slow/ prevent this. Inside the maggrid the plasma is insulated from the "ground" -magrid case through magnetic insulation (primarily ExB drift concerns). Thiis applies fully to the electrons with their small gyro radii, and less so to the ions with their large gyro radii. But the ions are also held away from the magrid due to the potential well, so they only have the possibility of reaching the maggrid if they are up scattered. The cusps are holes, not failures in the insulation. Bussard claims that the effective insulation has to reach greater than 9,999 parts out of 10,000. This is apparently achieved inside the magrid. This is different outside the magrid. Supports, e-guns, shell, etc have very much less insulation. Thus once the charge carriers outside the magrid reaches some threshold the insulation is insufficient for the developed current drain.
This is why the Wiffleball is critical, not only from an input energy cost, but from the obtainable density inside the Wiffleball compared to the density/ pressure outside. If the exterior regions can tolerate ~ 1 Micron of pressure, the Wiffleball trapping factor of several thousand allows for internal densities / pressures of several thousand 1000 Microns, and as the fusion rate scales as the square of the density...

Improved insulation of exterior surfaces may allow for a corresponding increase in the pressure inside the Wiffleball, thus further gains in fusion verses containment costs. This is applicable in the WB7, WB7.1 work which may have improved the effective insulation of the nubs (possibly one of the most vulnerable external structures). This might be tremendously important. The Wiffleball trapping factor is a ratio. If you can raise the basline (external pressure) through better insulation, the final Wiffleball pressure will increase proportionately. Conversely, if vacuum pumping is the limiting factor, tolerance of a modest increased pressure plays dividends. Of course when considering the steepness of the Pashin breakdown curves, the gains may be minimal.

Pashin breakdown curves are determined originally with experiment with flat parallel electrodes. Other shapes changes the values. So you may find tables that give estimates, but books of experimental data will probably be needed for an accurate understanding.

In the Polywell I doubt arcing can be beneficial. In this instance the discharge is a high current, but decreasing voltage. As the fusion is dependent on the voltage at logarithmic scales and current squared(?), the net result is probably bad. Also, consider that arcing results in beam- target fusion. While this might be resonable in some schemes, in the Polywell, the beam- beam interaction is very important. With beam target you have one chance, with a recirculating beam- beam you may have many thousands of chances with the same input energy (or possibly at one half the energy as both particles are moving).

Another point. In the Polywell WB6 the current to ground during Wiffleball operation was ~ 45 Amps. But this current was bouncing around inside (and recirculating) ~ 100,000 times . So the straight current (like with arcing) was equivalent to ~ 4,500,000 Amps. Assuming the arcing current (mostly from the emmiting e-gun to the wall) was 5,000 amps, this would represent only 1/1000th of the recirculating working current within the Wiffleball while it survived.
Lets see, 100 times more input current verses ~0.001 times working current /input current = ~ 0.000001 input efficiency for the same fusion rate (ignoring density differences- further lost efficiency). These numbers may not be accurate but they reflect the scale of the effects. . This is reflected in theWB5 results (I think). They got some improvement over WB4, but when they tried to improve performance10 fold by inputting 10X the current, the measured improvement was only 2X. Any improvement in confinement far outweighs any increase in input current. Given enough horsepower you can push the fusion output up, but the critical input to output balance suffers tremendously.

PS: This confinement efficiency benefit does have a limit. At some point energy loss from other sources like Bremsstrulung may become dominate. For D-D fusion, my impression is that this consideration is trivial. For P-B11 fusion the story may be different.

Dan Tibbets
To error is human... and I'm very human.

hanelyp
Posts: 2261
Joined: Fri Oct 26, 2007 8:50 pm

Post by hanelyp »

mattman wrote:Question:

How can the D2 be an ion flying into device center at 12,500 eV, if it needs to be inside the device to become ionized?
The D2 is ideally ionized near the edge of the plasma, which is at a different electric potential than the center. It is accelerated by the same potential well that contains it.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

hanelyp wrote:
mattman wrote:Question:

How can the D2 be an ion flying into device center at 12,500 eV, if it needs to be inside the device to become ionized?
The D2 is ideally ionized near the edge of the plasma, which is at a different electric potential than the center. It is accelerated by the same potential well that contains it.
A possible shortcoming of my simulations... Not so sure how well it models ionization given that the ions and electrons are at different representation ratios, and ionization is a quantum physics effect whee as my model is classical / relativistic. To add that in I'd have to do it ad-hoc and that wouldn't be very realistic, not to mention highly prone to error.

By just from picturing it, it would be great for efficiency if low energy ions I the center nuetralized, as then they could "coast" up the electors tactic energy hill. That would help to reduce thermalization.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

happyjack27 wrote:
hanelyp wrote:
mattman wrote:Question:
....


By just from picturing it, it would be great for efficiency if low energy ions I the center nuetralized, as then they could "coast" up the electors tactic energy hill. That would help to reduce thermalization.
If you are saying that it would be good for low energy ions IN the center to to neutralize (recombine with electrons) and then coast up the potential well, then you are confused.
The whole point of the potential well, that is created by the energy input of the electrons, is to create a "hill". The ions introduced at the top of the well fall towards the center base of the hill, accelerating to high energy where fusion can occur. Even if the ions bounce around in this central dense and high speed region they do not gain much angular momentum due to the ge3ometry, they conserve angular momentum. If this high speed is maintained in the mantle region, there are higher velocities that can contribute to more angular momentum relative to the core. The net effect on the ion/neutral vectors diminishes any confluence/ central focus. Once an ion reaches the center it has maximum KE, as it climbs up the opposite side of the well (or bounces back up the near side of the hill/ well) this KE is transferred to the potential well. The KE is converted to potential energy. It is important to note that this is a very efficient process. If the ion is neutralized in the center, two things will occur. The neutral particles are not effected by the potential well and reaches the outside at the same speed as it had in the center. It is not electrostatically contained. Without a net charge it is also not effected by the magnetic field. It is not contained at all (unless it re ionizes just in time). Even if it was re ionized near the edge of the Wiffleball, it would have it' maximum, KE. The space charge / potential well might then be able to stop it and reverse it for another pass towards the center, then you could recycle the ion energy, but you are not recycling the energy fully. It is an open cycle, not a closed cycle.(?) It would cost you the full accelerating energy on each pass. Um... thinking about it the energy may recycle provided you can re ionize the neutral very efficiently on the edge.. but the ionization on each pass is not 100%. You might need efficiencies of as much as 99.9999%.

Lets see... If you ionize neutrals on the first pass at 99.9% efficiency and then trap and recirculate them 1 million times, the energy cost per pass is a lot less than ionizing neutrals on each pass at 99.9% efficiency. The comparative ratio is ~ 1:1000. A perhaps more enlightening view point would be that if every atom had to be reionized on every pass and the ionization efficiency was 99.9% then each ion would only have a lifetime of ~ 1000 passes before its quota of input energy was consumed. With needing 10s of thousands of passes or more before fusion, then the fusion output would have to overcome this shortfall. If you had to start over again for every 1000 ion passes (instead of of ~ every 1,000,000 passes), then you would need the fusion output to exceed this difference. Assuming ~ 100,000 passes per fusion you would need eg 80KeV per ion * two ions /fusion * ratio of ion confinement time (passes) to confinement time till fusion or ~ 160 KeV energy input per pair of potential fusion ions / 0.01 = 16 MeV. The confinement time may be up to ~ 1/10 of this example value, but it shows that with the ionization efficiency losses, a large chunk of the energy difference between the fuel ions and the fusion products is ate up. It is a major loss concern. By ionizing once and then containing for many passes~ the contribution of ionization efficiencies becomes a trivial portion of the total losses.

Another consideration is the scattering of the ions and/or neutral particles. Without the high speed to low speed gradient from the center to the edge, scattering collisions- especially angular momentum changing collisions would speed up the loss of steep orbits , central focus/ confluence would suffer.

Annealing is another perhaps critical consideration. As the ions pass outwards from the core, through the mantle and to the edge region, the Coulomb collisionality increases logarithmically till at the edge where they are turning around, annealing (thermalization around a low average energy point) occurs. Without this slowing as the edge is approached, the thermalization cannot be held in check, and this is critical, at least for P-B11 fusion. You want an elliptical/ parabolic shaped potential well. Your system would have a square potential well from the perspective of your ion- neutral- ion particle. It would result in fast particles in the center but at increased input costs due to less than 100% ionization efficiencies, Changes in thermalization properties, and also changes in Bremmsthrulung properties would occur.

Dan Tibbets
Last edited by D Tibbets on Wed Aug 29, 2012 6:32 pm, edited 2 times in total.
To error is human... and I'm very human.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

D Tibbets wrote:
happyjack27 wrote:
hanelyp wrote: If you are saying that it would be good for low energy ions IN the center to to neutralize (recombine with electrons) and then coast up the potential well, then you are confused.
you've explained below why neutrazilzation of high energy ions in the center is a bad thing. and i completely agree with that.

you say above, however, that it is a bad thing for low energy ions in the center to coast back towards the edge, where they will ionize before reaching the escape threshold, and thus restore their high energy state.
The whole point of the potential well, that is created by the energy input of the electrons, is to create a "hill. The ions introduced at the topof the well fall towards the center base of the hill, accelerating to high energy where fusion can occur.
my simulations show clearly that the ions spread to take on a spectrum of KE+PE energy levels. there will be ions that reach 0 ke halfway between edge and center, just as there will be ones that reach 0 ke exactly at the "edge". the former will pass much slower through the center than the latter. the slower ones are more likely to neutralize with the cold electrons.
Even if the ions bounce around, they conserve energy (the ions in the region all have about the same energy), so that the dominate effect on the ions energy is the potential well. Once an ion reaches the center it has maximum pKe, as it climbs up the opposite side of the well (or bounces back up the near side of the hill/ well) this KE is transferred to the potential well. The KE is converted to potenyial energy. If the ion is neutralized in the center, two things will occur. The neutral particleis not effected by the potential well and reaches the outside at the same speed as it had in the center. It is not electrostatically contained. Without a net charge it is also not effected by the magnetic field. It is not contained at all. Even if it was reionized near the edge of the Wiffleball, it would have it'smaximum, KE. The space charge / potential well migh then be able to stop it and reverseit for another pass towards the center, then you could recycle the ion, but you are not recycling the energy. It is an open cycle, not a closed cycle. It would cost you the full accelerating energy on each pass. Um... thinking about it the energy may recycle provided you can reionize the neutral very efficiently on the edge.. but the ionization on each pass is not 100%.

Lets see... If you ionize neutrals on the first pass at 99.9% efficiency and then trap an recirculatethe ions 1 million times, the energy cost per pass is a lot less than ionizing neutrals on each pass at 99.9% efficiency. The comparative ratio would be 99.9% *1,000,000 compared to 0.5% * 1. (represents the 0.5% lost because they were not reionized on the edge). A perhaps more enlightening view point would be that if every atom had to be reionized on every path and the ionization efficiency was 99.5% then each ion would only have a lifetime of ~ 200 passes befor its quota of input energy was consumed. With needing 10s of thousands of passes or more befor fusion, then the fusion output would have to overcome this shortfall. If you had to start over again for every 200 ion passes, then you would need ~ 500 times as much input energy. If the input energy of the ion was 80 KeV, and you needed to multiply this by 50 = 4 MeV, then the ~ 3 MeV from fusion of two deuterium ions, then the input would be 8 MeV (two ions) for ~3 MeV output. This is a losing proposition even if all other confinement and loss mechanisms were perfect.
presumably you will have very few neutralizations, and they will be predominately the low energy (i.e. small orbit) protons. the very few neutrals will then be passing rather slowly through the ionization region. so your loss due to failed ionization would be pretty low.

besides, if this was such a critical loss vector, could not the same argument be made for neutrals that failed to ionize during neutral gas puff injection?

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

Please re read m post as I have edited/ re edited to try to make it more readable.

There will always be down scattered ions not reaching the edge. But as the downscattered ios become slower their Coulomblomb collisisionality in then mantle and core regions mincreases. This places a brake on the core and mantle down scattering limits.

And assumeing some level of angular momentum for the down scattered ions they eventually will reach the edge regionwhere the huge collision rate will thermalize them along with all of the other low energy ions that are at the top of their potential well. I have wondered how annealing would apply to downscattered ions, I'm guessing the increased collisionality of these slow ions in the mantle region will knock them to the edgeannealing region at some predictable rate. IE: the low side hermalization curve dose form to an extent, but it is limed to some extent due to the inevidable transit to the edge region in this collisional plasma. This again goes with my didlike of the term monoenergetic plasma. It is a arrow energy distribution compared to fully thermalization curves. Add to this my impression that it is the high energy thermalization curve that causes most of the problems for the Polywell, and annealing works best on this population. The low energy side takes care of itself as mentioned above.

Dan Tibbets
To error is human... and I'm very human.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

D Tibbets wrote:Please re read m post as I have edited/ re edited to try to make it more readable.

There will always be down scattered ions not reaching the edge. But as the downscattered ios become slower their Coulomblomb collisisionality in then mantle and core regions mincreases. This places a brake on the core and mantle down scattering limits.
i think it is more efficient and effective to think of it as an orthogonal system than a classical lagrangian system.
And assumeing some level of angular momentum for the down scattered ions they eventually will reach the edge regionwhere the huge collision rate will thermalize them along with all of the other low energy ions that are at the top of their potential well.
yes, but again, it simplifies the whole picture if you look at it as an orthogonal system; that is, as a superposition of states.

these comments do not even mention neutralization or ionization effects, nonetheless the consequences of neutrals not seeing the electrostatic hill.

happyjack27
Posts: 1439
Joined: Wed Jul 14, 2010 5:27 pm

Post by happyjack27 »

What i mean by the superposition view is: pictures electrostatic equilibrium. Now consider all possible orbits of charged particles consistent with that distribution. The rules of quantumn electrodynamics say that any system, regardless of its size, will exist in all such consistent states simultaneously.

mattman
Posts: 459
Joined: Tue May 27, 2008 11:14 pm

Post by mattman »

Taken from: http://thepolywellblog.blogspot.com/201 ... ation.html

Not all the sources are listed. If there is a "-" the information was not found.

=======

We needed to estimate how much the rings cost. To do this we needed to know what material to make the rings from, how much material and what construction might cost. Picking the ring material was difficult. We needed a material that was good at allowing magnetic fields through. It may be desirable for the material to have a low electrical conductivity, to prevent arching. It had to be durable against a bombardment of neutrons and hopefully it could handle heat well. We also would prefer something low cost and easily machinable. There may be other concerns as well [7]. Listed below are candidate materials with the relevant properties included [1, 2, 3, 4, 5, 6, 7].


[Henrys/Meter] | [electron-volts] | [Kelvin] | [Watts/Meter*Kelvin] | [$/Kg] | [Siemens/Meter]

Relative Magnetic Permeability | Neutron Activation Threshold: | Melting Point: | Thermal Conductivity @~20C | Price: | Electrical Conductivity:

Graphite 1.5999999 < 1.00E-4 3925 112 $2.75 1.4E+05


316 Stainless Steel - < 1.00E-4 1672 15.16 $0.91 1.3E+06


Neodymium - 1.00E+04 1297 16.5 $340.00 1.6E+06


Molybdenum 1.0000901 < 1.00E+05 2896 138 $29.00 1.7E+07


Tungsten-Carbide 1.0000008 < 1.00E-4 3143 84.02 $0.48 5.0E+06


silicon carbide 0.9999989 < 1.00E-4 3003 3.6 $8.81 1.0E-06


Boron 0.9999794 <1.00 E-5 2349 27.4 $2,000.00 5.0E-02


Teflon - < 1.00E-4 600 0.25 $20.55 1E-25to 1E-23


Aluminum 0.9994930 < 8.00E+5 933 237 $2.19 3.50E+07


We want a material with a high relative magnetic permeability. This measures how hard it is for the magnetic field to get through the ring walls. This would be indicated by a number of one or higher on the chart above. When neutrons hit a material, they can make it radioactive. The neutron needs at least a certain amount of energy for this to happen. This is called the neutron activation threshold. Ideally, we want as high a value as possible here. Unfortunately, many of these metals have their value artificially lowered because they contain carbon. A high thermal conductivity tells us that this material lets heat out of the reactor – this is desirable. Rings which transmit heat better are less likely to heat up and melt. We cannot get an ideal material from this list. However based on this information there are two likely choices for an initial reactor: stainless steel and tungsten-carbide. Of these options stainless steel may be better because it has an electrical conductivity one fourth that of tungsten-carbide.

Sources:

1."Data Available from the Nuclear Information Service." T-2 Nuclear Information Service. Los Alamos National Laboratory, 20 June 2007. Web. 03 Nov. 2011. http://t2.lanl.gov/data/data.html.

2. "Product Data Sheet, 316 Stainless Steel." Ak Steel Corporation, 2007. Web. 3 Nov. 2011.http://www.aksteel.com/pdf/markets_prod ... _Sheet.pdf.

3.“World Steel Prices." Steel Prices, 2010, 2011, 2012, Steel Price Index, Stainless Steel Prices, Steel Price News, World Steel Prices, Current Steel Pricing, Global Steel Prices, Average Steel Prices. World steel prices, 24 Oct. 2011. Web. 24 Oct. 2011. <http://www.worldsteelprices.com/>.

4. "Aluminum Prices, London Metal Exchange (LME) Aluminum Alloy Prices, COMEX and Shanghai Aluminum Prices." Current Primary and Scrap Metal Prices - LME (London Metal Exchange). Metal Prices, 27 Oct. 2011. Web. 27 Oct. 2011. <http://www.metalprices.com/FreeSite/metals/al/al.asp>.

5."Metal-Pages - Tungsten Prices." Current Metal Prices | Historical Metal Prices | Metal News | Marketplace. Metal-Pages, 27 Oct. 2011. Web. 27 Oct. 2011. <http://www.metal-pages.com/metalprices/tungsten/>.

6."Molybdenum." Espimetals. Espimetals, 3 Nov. 2011. Web. 3 Nov. 2011. <http://www.espimetals.com/tech/molybdenum.pdf>.

7. Tibbets, Dan. "View Topic - What Is the Best Material for the Rings?" Talk-Polywell.org. Talk Polywell, 2 Nov. 2011. Web. 2 Nov. 2011. <viewtopic.php?t=3373>.


Source List (May be Duplicates):

http://www.northerngraphite.com/index.p ... ite-price/
http://biotsavart.tripod.com/bmt.htm
http://www.metal-pages.com/metalprices/neodymium/
http://www.azom.com/article.aspx?ArticleID=863
http://physics.info/conduction/
http://www.memsnet.org/material/tungstencarbidewcbulk/
http://aries.ucsd.edu/LIB/PROPS/PANOS/c.html
http://www.chemicool.com/elements/boron.html
http://www.imoa.info/molybdenum/molybde ... erties.php
http://www.chemicool.com/elements/molybdenum.html
http://hypertextbook.com/facts/2004/Afr ... rave.shtml
http://www.reade.com/Particle_Briefings ... ities.html
http://hyperphysics.phy-astr.gsu.edu/hb ... gprop.html
http://www.wolframalpha.com/input/?i=bo ... eptibility
http://www.wolframalpha.com/input/?i=ne ... ptibility-
http://www.bls.gov/oco/ocos223.htm
Last edited by mattman on Wed Sep 12, 2012 10:10 pm, edited 1 time in total.

D Tibbets
Posts: 2775
Joined: Thu Jun 26, 2008 6:52 am

Post by D Tibbets »

happyjack27 wrote:What i mean by the superposition view is: pictures electrostatic equilibrium. Now consider all possible orbits of charged particles consistent with that distribution. The rules of quantumn electrodynamics say that any system, regardless of its size, will exist in all such consistent states simultaneously.
I don't know quantum electrodynamics, but a possible consideration is that the Polywell is not in MHD (magnetic hydrodynamic) stability, so MHD rules can be missleading. Whether this applies by extension into quantum electrodynamics is the question.

Dan Tibbets
To error is human... and I'm very human.

Post Reply