WB-6 tests results: FOUR successful tests. Lab notes

Discuss how polywell fusion works; share theoretical questions and answers.

Moderators: tonybarry, MSimon

KitemanSA
Posts: 6179
Joined: Sun Sep 28, 2008 3:05 pm
Location: OlyPen WA

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by KitemanSA »

Not, it is not. The OP pdf was an illicit copy of a proprietary report that was posted without the copyrighter's permission. All responsible hosts removed it when asked.

crowberry
Posts: 672
Joined: Sun Sep 08, 2013 6:34 am

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by crowberry »

I guess that document is mostly of historical interest, because EMC2 has built WB7 to reproduce the WB6 results and WB8 to achieve significantly better performance than WB6. The interesting question is not how well WB6 performed, but rather how did WB7 and WB8 perform? That is of course not known. If General Fusion, LPP or Sorlox makes a significant step forward maybe that will encourage the Navy to show their cards?

prestonbarrows
Posts: 78
Joined: Sat Aug 03, 2013 4:41 pm

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by prestonbarrows »

My worry is I have still never seen a legitimate set of data showing neutron production rates and what sort of operating parameters produced them. There are just 2-3 mentions 10^9 n/s DD in whitepapers and presentations well after the fact.

There are a number of great papers from the Sydney group demonstrating well formation with electron injection. But, that is a much different thing than successfully fusing within that well at an appreciable rate.

It just is fishy that the only neutron-successful shots happened on the last day of the program's funding and then nothing published by anyone for 8 years...
I realize there are complications with private IP and the Navy etc. But, if things were going well you think there would at least be an official release that they are breaking neutron production numbers to drum up support for the project.

Unless I am missing something? Does anyone have legit data about neutron production in any polywell device? I would love to see it.

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

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by D Tibbets »

The WB6 report that was on EMC2's web site for a while gave some details about WB6 methods and measurements. The conclusion's have been presented in a number of sources, including patent applications.

The fusion of deuterium by electrostatic acceleration of the ions has been demonstrated repeatedly by many. It is a relatively simple methodology and the detection of neutrons is also straight forward. There are some pitfalls that have to be controlled, but overall it is doable by high school students with a couple of thousand dollars of equipment.
The pit falls are mostly related to measurements. Electronic Neutron detectors can be spoofed by electronic noise, but there are multiple ways to minimize or control for this.

But, neutron production and measurement is not the issue. WB6 results, if reportedly truthfully, did suffer from a small sample size, and operated in the margins of statistical confidence. Not so much due to the neutron production rate, but due to the sub millisecond time frames where the machine was operating in Wiffleball mode and with a deep potential well. I personally think that the detection of neutrons in earlier machines is more significant in terms of neutron production- the creation of conditions where relatively large proportions of neutrons were produced, even with ridiculously small potential wells of only a few hundred volts.

The real questions are two fold. First confirmation and expansion on Wiffleball formation and the consequences. Along with this is the depth and shape of the potential well. The second set of issues is not whether the Polywell can produce profuse fusion, but what is the price of doing so. All sorts of things relate to this, such as annealing, thermalization issues, and confluence towards the center. Magnetic and size and voltage scaling are all important.
But the most basic question (once the Wiffleball is accepted) is the electron costs, how they can be optimized, and what the scaling is. The output is of course important, and I understand that the output magnetic , size and voltage scaling is well accepted. What is less certain is how the input costs scale.

WB6 may have given an approximate baseline for both output and input numbers, but it is WB7 data that really gives any indication of the feasibility and lower level scaling of the machine. WB8 expands on this data and perhaps serves as much as a testbed for engineering issues as an expansion of the scaling laws and confidence.

So, the WB6 methods and results are enticing, but to make any predictions or analysis with confidence the WB7 data is needed. That is the information that is restricted to only a few. The WB6 data is out there, though in a piecemeal fashion, once the EMC2 report was withdrawn.

Mathematical modeling of the Polywell system is complex as there are competing processes that are dependent on a number of assumptions.These assumptions are more complex that a 'blob of plasma' assumptions. Position in the system is paramount. I'm far from a computer wiz or a plasma expert, but I suspect the predictive models are more complex than those of a relatively simple magnatized neutral plasma. The area where the modeling may be easier is with macro instabilities and possible edge effects, IF the assumption of always favorable magnetic surfaces is accepted. It mostly removes (can be ignored for simple modeling) these instabilities and pinches that confounds the Tokamak efforts .

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

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

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by mattman »

My worry is I have still never seen a legitimate set of data showing neutron production rates and what sort of operating parameters produced them. There are just 2-3 mentions 10^9 n/s DD in whitepapers and presentations well after the fact.

There are a number of great papers from the Sydney group demonstrating well formation with electron injection. But, that is a much different thing than successfully fusing within that well at an appreciable rate.

It just is fishy that the only neutron-successful shots happened on the last day of the program's funding and then nothing published by anyone for 8 years...
I realize there are complications with private IP and the Navy etc. But, if things were going well you think there would at least be an official release that they are breaking neutron production numbers to drum up support for the project.

Unless I am missing something? Does anyone have legit data about neutron production in any polywell device? I would love to see it.

I agree. I read Bussards IAF paper and I HATED IT.

It was sloppy. Sloppy wording, Sloppy results reporting and hastily written. There are basic standards for science publishing, and this paper failed them.

Here is the data presented:

Image

Image


It felt like pulling teeth -- trying to get benchmarks from this. It took weeks. I needed to estimate some distances based on counting pixels in photos. In fact, Bussard gives different information in his google presentation and his paper, and you must use math to check consistency. Also, you get ranges for some variables (like the # of ampturns). Sometimes I made educated guesses (I did this with the power supply) using Tom Ligons memory. Finally, the neutrons were only detected with two detectors. What convinced me was that only ~800,000 ions actually fused -- a tiny fraction of the gas puffed in.

The post: "Taking A Stab At Simulation" covers this.

Details are from the IAF paper, his Google talk, Tom Ligon and Mark Duncans work. If you want to double check my work, I welcome it.



https://github.com/ThePolywellGuy/Posts ... -6-13).pdf

After I wrote this, I found a mistake. I assumed there were electron emitters along the sides. There are no emitters in the photos along the sides.

Here is the section you might be interested in, neutron detection comes at the end:
===========================


The WB6 Experiment
This post examines simulating the Polywell. All good simulations begin with a benchmark. The code must first duplicate a real world test. The test to simulate - is Bussards' WB6 results from November 10th 2005.

Image
Figure 1: A blueprint of the ring structure in WB6. It is assumed these are meter distances.

Machine Geometry:
The ring geometry comes from a blueprint in Bussards’ Google presentation [14]. The tank volume comes from Bussards Google presentation [14] and Mark Duncan’s summary [16, page 15]. The tank had a two meter diameter and was 3.5 meters length. Inside the tank was a wire cage which is modeled as 3’ 8” side. The size of wire cage is an estimate; based on the machine photos. Such a cage would fit inside the tank comfortably.

Image
Figure 2: This shows the geometry of the machine used in this document. The red device represents the gas puffer tube. This is located at the corners of the rings. The yellow device represents the capacitor emitter this is located on the ring axis.

Ring Power Supply:

One long wire is wrapped through all six rings and was connected to 240 batteries [16]. These RV batteries will be modeled as having twelve volts and one hundred amps of current in each. This kind of battery is common [32]. These batteries were most likely wired in series. This makes a high voltage, low current source. This arrangement will use up electricity faster. But, this is ideal for a pulsed machine. This power source is superior for thin wire, many looped electromagnet [17]. Two hundred loops of No. 10 copper wire would fit easily inside the ring cross section [33]. The rings may have had between one and two hundred turns [31]. The rings were supplied with a current that ramped from zero to 4,000 amps. The number of amp turns was between 20,000 and 800,000. This is a critical number. The rings are mainly modeled here as having 20K amp turns.

The Emitter:

A capacitor electron emitter was used in the WB6 tests. In this work, the capacitor is placed halfway between the rings and the cage. The capacitors discharged between 0 and 40 amps of electrons.

Image

The Gas Puffer:

Four gas puffers were used in the test [photos]. These consisted of long tube spaced out from the corners of the rings. It is estimated here that they were 0.292 meters from the ring center [photos]. About 4.5E-6 moles of gas was puffed into the cage [Estimated below]. This is a tiny amount. The gas entered the machine at 0.04 Pa [2, page 11]. This gas was likely depressurized from a high pressure (tens of atmospheres) feedstock [24]. Bussard estimated that the gas had a number density of one atom in 1e-19 cubic meters [2]. The amount of gas used was estimated using the ideal gas law and listed tank pressures [2, page 12].

Image

The Neutron Detectors:
Two neutron detectors were used. These detectors could only record a portion of all the neutrons produced. This is related to the area they occupy on the chamber walls. Hence, the amount of neutrons produced in the test was extrapolated; from data for a single detector (Google presentation, 54:10).

The Experiment:
The experiment occurred in five distinct steps. First, the tank was pumped down to a starting pressure of 1.33E-5 Pa. About 6E-8 moles of air was in the tank at the beginning. Next, a potential of 12,500 volts between the cage and the rings was applied. Note: this voltage was described in the Google presentation as 12 kV. The capacitor emitters were switched on after this - at 40 amps for ~0.0005 seconds. The emitted electrons flew towards the rings. The magnetic Lorentz force overtook the electric Lorentz force and the electrons started following the magnetic fields [Estimated here]. The magnetic field generally pointed outward. The electrons recirculated along this field into the cusps. There, they hit magnetic mirror. This reflects them back into the machine. This trapped a cloud of electrons. This cloud generated a potential of 10,000 volts from the gas puffer to the center.

Roughly 1.2 to 1.5 E12 net electrons were trapped to create this drop [estimated here]. The electron lifetime was on average 1E7 seconds [Google presentation 44:50, 16, 2, page 12]. Bussard estimated that the electrons had a number density of 1E19 electrons/meters^3 [2]. This is the same as the gas. The electrons had a bell curve of energy. This was between 0 and 12,500 eV with an average of ~2,500 eV [2]. The electrons had an average velocity of between 1 and 4E7 meters per second [2, Google presentation (45:00)]. Bussard estimated that the electrons feel a 1,000 gauss containment field. These estimates were used in Bussards beta ratio calculation. Mark Duncan's work lists the WB6 strength as 1,300 Gauss, Dan Tibbets uses 1,000 gauss and the IAF paper said this is under 3,000 gauss [16, 31, 2 page 10]. However, it is unclear where in this field strength is calculated.

Image

In the last step of operation: uncharged deuterium gas is puffed in at room temperature (0.02 eV). Because it is neutral it is not affected by the cage voltage. It can therefore reach the rings without trouble. About 2.7E+18 molecules of deuterium gas were puffed into the chamber. Because the ring structure only represented ~0.002 percent of the tank volume, there is good reason to think most of this gas did not enter the rings. When the gas reaches the edges of the electron cloud it exchanges energy with the electrons. If this exchange causes the molecule to heat up past ~16 eV, they break apart into two electrons and two deuterium ions. The ions see the 10,000 volt drop and fly towards the center building up speed. The ion is 35,461 times the diameter of the electrons and 3,626 times the mass of the electron [8, 9, 28, 29]. If two ions hit in the center, at 10 KeV, they may fuse. The product contains energy on the order of ~10 MeV. It has too much energy to be contained by the fields. It rapidly leaves the ring structure.

Bussard reported a total of ~2E5 neutrons were generated over the 0.0004 seconds of the test [2, page 11]. Note: Bussard states the test was 0.00025 seconds in his presentation (53:50) and 0.0004 seconds in his IAF paper. 0.0004 is more consistent with other information. For every neutron detected, four deuterium ions fused and two fusion reactions occurred. This means that a rate of ~1E9 fusions per second was produced (Google presentation 54:03, IAF paper, page 11). This also means 800,000 deuterium ions were fused. This is a tiny fraction of the deuterium atoms injected.

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

Re: WB-6 tests results: FOUR successful tests. Lab notes

Post by D Tibbets »

I need to re- read your post several times. Much of your information is pertinent but there are errors. Some of them:

The electron emitters (E guns) were in some of the corners, spaced a calculated distance beyond the magrid. The electron emitter current was provided by batteries, not capacitors. The capacitors at up to 12,000 volts powered the surface of the magrid, not the electron emitters. The voltage of the magrid surface provided the volts to accelerate the electrons, but the electron source was the low voltage e- guns. The magrid surface was not the emmiters, at least not the e emitters during steady state phase of the tests. As the gas density outside the magrid climbed above ~ 10^-6 Torrs, arc breakdown occurred and Pashin arc discharge directly from the Magrid surface to the cage became dominate. In this situation the negative magrid surface was indeed an electron emmiter, but this was after the test proper. The thousands of amps of current at this time is also reasonable, considering there was a conduction path to ground at this time and the voltage driving the current was ~ 300-1000 times higher than the voltage to the car headlight electron emitters. That explains two of the graphs you present.

The voltage to the electromagnet wire was 12 volts to perhaps several dozen. I have estimated wire sizes closer to 6 to 8 gauge rather than 10 gauge. The length of the wire through all magnets added up to ~ 1 kilometer and the needed voltage for this presumably current limited wire length/ diameter dictated the voltage which I have roughly guestimated as 24 to 36 volts. This suggests that the 12 volt batteries were arranged in parallel banks of 2-3 batteries in series each. Note that this battery bank was presumably also used in other machines like WB4 and 5. WB4 operated at up to 3,000 Gauss. And for Beta testing the current may have been ramped up past these values so the excess capacity relative to the WB6 neutron counting tests are not unreasonable. Assuming 200 batteries arranged in 3 battery banks(in series) of ~ 70 parallel banks, and assuming you plan to use up to ~ 100 Amps from each battery gives a capacity of up to ~ 20,000 Amps. The numbers are reasonable.

Your number of fusions is way off. There was claimed ~ 500 million neutrons per second neutron production rates during the ~ 0.25 milliseconds of Beta=1 machine operation. For D-D fuel this equates to ~ 1 billion fusions per second as only one half of the fusions produce a neutron. Neutron counting is dependent on two things. The surface area of a sphere that is occupied by the detector tube, and the efficiency of the detector. Neutrons are assumed to be emmited isotropically (in all directions). Some reasoned numbers is that the detector area was ~ 1/1000 of the total sphere surface area at this given radius. Also, electronic neutron detectors may detect only ~ 1/1,000 to 1/10,000 of the neutrons that actually pass through it . The counter is tested with known neutron sources in order to calibrate for this. In this example if 3 neutrons was detected in under 1 millisecond then there would be 3*1000 * 1,000/0.001 sec = 1 billion neutron per second . Note that the detector available was limited to one millisecond (or was it 0.5 ms?) sampling intervals, so it was obvious that all of the neutrons were produced during this time, and presumably occurred during the ~ 0.25 milliseconds of machine operation. This is reasonable but not certain. Better equipment with WB7 would be more percise.

I'm not sure what your quoted potential well depths refer to. Certainly during Beta= one tests (where density was derived from Photo Multiplier Tube measurements) the voltage was set to arbitrary levels (limited on the high end by the rated voltage of the capacitors used of ~ 12,000 Volts). Beta values are based on the density and voltage(energy) of the charged particles divided by the magnetic field strength. The values of two of the perimeters can be set at any arbitrary value and the third value calculated. You do not need to only use the maximum available voltage, and actually a range of conditions is desirable to build the graph. The PMT data was collected at a given voltage, while the electromagnet current was ramped from zero to some maximal current (the resultat magnetic field strength could be calculated from the Amp turns or may have been directly measured). The PMT data gives the resultant changing density. That there was a peak when the magnetic field passed through Beta=1 conditions is consistent with Wiffleball performance. The magnitude of the PMT peak derives the Wiffleball trapping factor. Apparently this experimental data agreed with / derived the calculated Wiffleball trapping. This series of tests is independent of the test parameters set up of the final days where deuterium gas was used and neutron measurements were made. I'm guessing that Helium gas (often used in plasma test systems/ vacuum systems) or hydrogen gas was used for the Beta tests. Deuterium is less readily available and more expensive. Note that the PMT data is reasonable, but Nebel was unsatisfied with the reliability so he/ they used a different method to measure density during Beta tests of WB7 and that this independent method reportedly matched/ verified the PMT measurements.

As for the presentation quality that Bussard put out, I have the impression that he was not the best communicator. And, after WB6, as is obvious in his Google talk, he was ill, possibly dieing of cancer or some other terminal illness. Any omissions/ obtuseness can be blamed on this. Also, while Bussard threatened to release all of his data, this was a financing ploy , and he actually kept things pretty close to his chest. A situation that I assume was appreciated by the eventual funders (again the Navy).

I don't have any idea what run times, neutron counts, etc. that was achieved with WB7 and WB7.1 testing, but with the advantages of hind sight, and presumably better equipment, and persistent testing not ended by machine failure; it is reasonable that much more precise and predictive data was obtained.
The final measurements of neutron production terminated by machine failure and funding termination makes good drama, but it actually means nothing in regards to the data validity.

It would be awfully nice to have that WB7 data available .

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

Post Reply