Superb Web Site on Farnsworth's life: The Farnsworth Chronicles- http://farnovision.com
Here is an interesting Borderland Science Ftp file on Farnsworth: http://farnovision.com/chronicles/fusion/vassilatos.html

Best heard with 32+ bit wave table sound cards.

Return to Home Page

Below are the results of my search through the patents of Philo T. Farnsworth. While others have chronicled the life and times of P.T. Farnsworth, my interest lies in his inventions and patents. More specifically, how the design of those separate devices led up to the final development of the Fusor - the only fusion reactor ever designed and operated at better than break-even efficiency! Farnsworth's Fusor produced neutron counts measured in the billions-per-second. No other fusion device built, even 30 years later can match this performance. It is unfortunate that Farnsworth died before perfecting the device into a commercial product, but that does not detract from his accomplishment in the least. That one man could conceive of and design a working fusion reactor is a phenomenal personal achievement. It would be nice if someone would build another working device.

Please note that the patents were printed from microfilm, then scanned with some minimal clean-up of the resulting images. Enjoy - it's been a *lot* of work!

Click on the underlined patent numbers below for full patent information.
Note: Patents are currently being added to the Website from the newest (bottom) to oldest (top).

• 2,135,615Multipactor
• Multipactor design allows flexibility of use
• Utilizes cathode with many tiny apertures through which electrons may be accelerated to another cathode
• Theoretical gain of 1,460,000 can be had with this tube
• Aluminum can emit secondary electrons greater than unity.
• Symmetrical electrodes allow the tube to operate with AC instead of DC
• The Fusor would also use apertured cathode/anode structures, through which electrons would be accelerated and trapped in oscillating fields.

Full patent

• Interesting zigzag, multistage multiplier designs.
• Apparently it is immaterial what angle electrons strike an emissive surface - they always leave perpendicularly.
• This device uses accelerating screens in front of the emissive surfaces to accelerate the electrons. The screens may be biased at the same or higher voltage than the electrodes themselves. The Fusor uses biased screens.

2,217,860 Split Cathode Multiplier

Full patent

• Another cone shaped multiplier. Design talks about using anode voltages as high as 10,000 volts for very high power outputs.
• Fundamentally the same as the tube design in patent 2,204,479, but using dual cathodes, very high power for a single tube.
• It looks like this is the first multipactor that Farnsworth designed that could supply hefty power outputs.

Full patent

• CRT uses cells of ionizable gas, which are ionized by X-rays to fluoresce a screen to produce an image.
• Only device I'm familiar with that uses a scanning electron beam to generate a scanning X-ray beam.
• Probably no application toward the Fusor, but perhaps a possible immediate use for scanning lithography to be used in the manufacture of microchips?

Full patent

• Uses apertured cathode assemblies. Amplification of 100,000 to 1 million possible. Uses 200 MHz biasing oscillator, which allows for very small device.
• Within a multiplier, the positive ions have a mobility factor of almost precisely 1/500^th that of electrons
• This patent is specifically for an improved dissector tube design, and Farnsworth substantially reiterates his earlier patent claims on how multipliers can be operated to amplify RF, self-oscillate, etc.
• Nothing new in this patent as regards Fusor technology.

Full patent

• Zigzag multistage amplifier, much like 2,204,479.
• Placing fine wire grids in front of the emissive electrodes, and charged to high potentials, accelerate the electrons and improve operation.
• One design of this tube utilizes a magnetic field, which causes the path of the electrons to form a spiral, and he places the secondary emissive surfaces in a corresponding spiral so they will be impacted.
• No appreciable addition to Fusor technology in this patent over previous patents.

Full patent

• As best as I can tell, this is a traveling wave tube design, probably the first ever. A moving electron beam is played upon two curved plate targets. I am surprised that emissive surfaces were not employed in the design of this tube.
• No definite application to Fusor technology is evident.

Full patent

• Input electron flow controls number of secondarily emitted electrons, allowing high ration of mutual conductance value to dc flow in the device
• No Fusor contribution.

Full patent

• Rectifier has transformer secondary winding inside of tube structure, with transformer core and primary outside of tube, providing very high voltage isolation, allowing very high voltages to be rectified, without HV breakdown. High frequency AC (10 kHz) is used in the filament transformer.
• It is possible that the high voltage power supply for the Fusor utilized such rectifiers.

Full patent

• Multiple electrode multiplier where electrodes are in series - increases amplification of AC component without necessarily magnifying DC component.
• Interesting electrode design - it seems the more Farnsworth worked with multipactors, the more he used electrode shape and placement to control the action of the electrons.

Full patent

• Ion transport pump capable of vacuum of 10-10 millimeters of mercury. Glass envelope. Electrostatic fields are shaped to force ions into a restricted area then scavenged by a roughing pump. The vacuum tube should be operated as a multipactor to insure ionization is continuously done, via using a tungsten rhenium cold cathode, and an RF current, chosen to oscillate at the electron transit frequency between the two cathodes (dynodes).
• This ion vacuum pump is more difficult to build than his later pump, and this pump also has only one stage of pumping, versus his later two-stage pump.
• There is a specific note of the very high ion concentration in the center of the anode space - exactly what was going to be required in the Fusor.

Full patent

• Produces high-density electron beam with small cross section using multipactor technology. Uses 50 MHz oscillator as part of multipactor circuit. Cold cathode, no filament, used in CRT.
• Note the dished cathode, conical director electrodes, annular focusing rings - all these elemental designs were to be utilized in Farnsworth's patent 3,258,402 - Electric Discharge Device for Producing Interactions Between Nuclei.

Full patent

• Superior, two stage ion vacuum pump. Glass envelope. Uses multipactor design, simple construction techniques. Incorporates several reflector electrodes with spherical curvatures.
• Capable of 10-10 mm vacuum. Has magnetic solenoid coils, though a claim is made that the pump will work without them. Maximum DC voltage is only a couple thousand volts. Pump is about 18 inches long, and about 6 inches in diameter.
• His ion pumps really allow Farnsworth to gain critical experience with the controlled movement and directional flow of ions.
• Farnsworth had a knack for designing a complex machine to do a task, then he comes up with a simpler way - his two ion pumps are an excellent example of this simplification with experience.

Full patent 3,258,402

• An awesome patent, with incredible detail of construction. Farnsworth's pictorials start out by showing a pure spherical element design, based upon the 95% open mesh anode construction, then he describes a bizarre nested cup anode that is supposed to provide field shaping to create a true spherical electrostatic field. Both these designs appear to be experimental, and were test beds to provide him experimental data and design clues. Next, he t hrows in ion injection (personally, I'd modify the electron gun he used on the CRT patent) then he finally gets into serious practical designs using multiple spherical anodes used in conjunction with ion injection.
• This multi-concentric sphere design would be an absolute nightmare to design, machine, and construct, not to mention outrageously expensive!
• His early two reflector design required micrometer screw adjustments to precisely position the dynodes, so as to achieve focusing, but the final 12-gun design does not appear to have any way of making similar adjustments! I'd say he had nailed down the electrostatic manipulation and focusing of the beams to perfection by the time he built the 12-gun unit. The vacuum-tight electrode design is obviously the result of tried and true work.
• He modulates the anode with RF, but the anode has 140 KV DC on it, no mean task!
• A vacuum of 10-10 mm is required for operation, so his ion transport vacuum pump was undoubtedly used to pump this baby down!
• The cathode structures have a hole drilled in the center of the spherical reflector, which forms a "well", into which errant electrons fall and get lost. This is a definite sign of tweaking the anode current to perfection. An armchair design would never have this feature.
• There is an incredible wealth of information and formula in the text on electron lenses and how to direct and focus the electron beams very precisely.
• The "typical" dimensions are an invaluable aid in trying to replicate this device.
• Farnsworth has included many formula and operational directions, or "user instructions" which are immaterial to his receiving a patent, indeed, he was exceptionally wordy and descriptive. This extra information would be very valuable to anyone trying to construct a copy of his device. My feelings from reading the patent text are that Farnsworth may have deliberately included a lot of extra information in this patent to help others duplicate this device.
• The multipactor with the electron emissive surfaces, the curved cathodes, the nested spherical electrodes, shaped electrostatic fields, ion vacuum pump, conical beam shapers, ion guns, electron scavenging, etc., all of his past patents start to come together in this one device!

Full patent

• This was the last technical patent I could find by Farnsworth - the culmination of a lifetime of work!
• Practically every vacuum tube concept this man developed during 40 years of research came together and was utilized in this invention.
• First, it should be noted that he completely turns his previous fusion device inside out, making the cathode the center of the sphere, instead of the anode.
• This device not only produces huge amounts of heat from the nuclear reactions, but up to 40% of the energy is converted directly into electrical current which may be extracted from the anode/cathode connections.
• His original Fusor designs discuss the creation of a virtual anode in the center of the sphere, but now this device uses both virtual anodes and cathodes.
• He describes the bouncing of ions between his virtual electrodes, and coins the term "inertial containment" of the ions.
• The boundaries of the virtual anodes and cathodes must be kept well defined and separate.
• The emission of secondary electrons from the inner cathode surface is necessary to produce virtual anodes and cathodes within the inner cathodic space.
• Always run at lowest practical gas density to keep charge-exchange losses to a minimum.
• To reduce scattering losses, the electron current flowing in the cathodic space needs to be high. This is accomplished by having electron emitters within the cathode - are we surprised by the use of emissive surfaces to generate extra electrons?
• Many of the electrons traveling in the cathodic space have orbital paths, which sweep the inside of the cathodic space. Having more than 10% orbiting electrons is undesirable - they fuzz up the boundaries of the virtual electrodes. The orbiting electrons, on the other hand, reduce the probability of electron recombination.
• A series of alternating anodes and cathodes is made to exist within the space enclosed by the inner sphere and this creates a cathodic "singularity" or mathematical singularity at the center virtual electrode!
• If the ratio of the radius of the virtual anode and virtual cathode is much greater than 3 to 1, then the inner cathode may not form, and if this ratio is smaller, then the circulatory currents required to maintain the virtual electrodes becomes too large to sustain.
• The electron "mesh current", which is the result of secondary electron emission and the electrons which flow from the real cathode to the cathode apertures to the real anode, will initially be large while the virtual electrodes form. This current should immediately drop to insignificant values once the Poissor (virtual electrode structure) has formed.
• Once the virtual fields are established, and the space charge has stabilized, and the ions have been injected, the power consumed by the device is a minimum.
• Stainless steel (which happens to contains nickel) and molybdenum are listed as an electron emitting material for cathodes. Why did he not use his ubiquitous alloy of nickel and barium? Maybe the cathode does not become red hot? He talks about the difficulty in removing excess heat from this cathode - maybe the heat can be ignored by using his nickel with 1% barium alloy, as was used in his earlier patents on high power multipactors (patent 2,139,813 -Secondary Emission Electrode?)
• Figure 9 shows a working embodiment of the Fusor, using two concentric cathodic spheres within a spherical anode. The outer cathode is negatively biased, and also serves to remove ions from the non-space-charged field at turn-on.
• Losses within the Fusor are typically about 10% of the total power created (not power input!).
• Heat absorption by the inner cathode is a very serious matter, and Farnsworth deals with it by conducting heat away through the power supply line (pathetic heat transfer method, at best!) I suggest that replacing the electrical supply line with a suitably designed heat pipe (originally designed for space use by NASA) and external heat-sink (part of the water cooling system of the anode?) would work a hundred times better.
• Increasing the applied voltage will increase the neutron count. The current will increase by three halves power. Higher voltages decrease the heat losses within the system.
• Normal gas pressure during operation is 3.535x10^19 molecules or less, or 3.353x10^13 ions/cc at a normal operating pressure of 10-3 mm Hg.
• Fusor is initially pumped down to 10-9 to 10-10 mm Hg, high voltage is turned on, and deuterium and tritium gas (50/50 mixture) are introduced into the ion guns at a pressure of 10-6 to 10-4 mm Hg. The virtual electrodes (Poissor) will form.
• Within the larger virtual anode a small virtual cathode will form, and it is within this small cathodic space that fusion is most likely to occur.
• Neutrons will carry energy out through the walls of the device and represent a loss of energy at the core of the device, leaving 64% of the energy available to be converted directly into electricity.
• Only those ions which are created by the actual fusion process will have sufficient energy to produce electrical energy
• The system is self-regulating, in that a sudden increase in source gas will not produce a catastrophic surge in output energy - the virtual sheaths change size, modifying plasma density, performing a self-regulatory action.
• The ions leaving the core of the Fusor can have energy levels ranging from 0.82 meV to 14.7 meV. If the goal is to directly convert the energy into electrical power, it is most desirable to limit the production of ions to those with the lowest meV level (less heating of the outer anode.) If the desire is to produce heat which will then be converted into electrical (or other) power, then it would be best to produce only the faster (higher meV) particles.
• Modifying the gas density and using a high volume vacuum pump allows some leeway in controlling what elements get to enter the fusion process, but there are limits on what the operator can do (the highest meV particles will be created regardless.)
• In examining the text of the patent, I noted that Farnsworth did not expend much detail on his later, improved version of the Fusor, but instead lavishes descriptive detail on the earlier, unimproved version. This early design uses the weaker ion guns, and the unit does not have the electrostatic focusing rings that the final design incorporates. The early unit does not have (or possibly may not require) the ability to make micrometer adjustments of the inner anode sphere's position within the anode . The later Fusor design, which eliminates one of the inner concentric spheres, is also much, much easier to build. I think Farnsworth's strategy was to provide a competitor with detail plans for the early inferior device, and to provide just enough detail of his more advanced work so that he could defend his patent should the need arise.
• Examining Figure 23, you will notice that there are visible several screened holes arranged around the inside of the cathode sphere. The viewing port is aligned with one of these screens, allowing the user to look all the way into the core of the Fusor. No explanation of the particulars of the viewing port are given, but the drawing shows some detail of the viewing port which is not numbered or described. I believe that one of the undescribed it ems shown in the drawing of the port is a wire screen, which would be necessary to extend the anode surface across the viewing area. Immediately behind this undescribed line (screen) is a second item, which also appears to go across the viewing area. I conjecture that it could be a high-density filter for blocking the intense UV light produced by the core. A leaded glass or water filled port might be necessary to block the neutrons produced by the core so that the operator would not be injured by neutrons coming up t hrough the viewing port. Also, the design shown makes no provision for cooling the anode jacket, nor does any neutron shielding material surround the anode. Protection against high-speed neutrons would be mandatory for safe operation of this device.
• No description is given for the spacing between the holes for the screened ports and the smaller holes through which the ion jets pass.
• In Figure 27, It is unclear how the cathode sphere can be positioned by the micrometer adjustments of the 71a electrode assembly, since the 49a assembly is described as a "rigid" attachment capable of supporting the cathode sphere. If the mounting 110 consists of two concentric sliding cylinders, then I can see how the inner sphere could be moved up and down, but not side to side. The drawing shows an adjustment design that only al lows the sphere to be tilted a small amount , from side-to-side, and even then only in one axis. It is, however, possible that a second adjustment screw, for another axis, which is not visible to the observer in this drawing, might be hidden behind the high voltage bushing. It would also seem necessary that the physical mounting of the supporting structures, as shown, would necessitate an open bias screen port to maintain the symmetry of the spherical electrode and the resulting spherical virtual electrodes. The large, solid mounting, as shown, ruins the symmetry of the screened ports.
• In Figure 23, there is a lack of detail in the exact implementation of the high voltage terminations.
• Also in Figure 23, there is a ceramic feed-through bushing at the high voltage mount - this should be unnecessary, though it may have served a purpose in the design shown in Figure 11.
• Figure 26 has the electrical connection (wire 156) to be make from spring wire, and it looks like the design was to allow for the adjusting movement of the inner sphere. The patent describes the spring wire as touching metal plug 145, but the wire is inside a conducting metal tube 144, so it does not seem absolutely necessary for the wire to touch the plug.
• In Figure 27, the filament loop drops below the radial centerline 165, to allow ions to pass through the center of the filament? (In Figure 21 the filament is shown as a loop aligned with the center of the gun.) And there are unidentified plates or disks at the rear of the tube - are these plates possible ion emitters, as is shown in some other ion gun patent descriptions?
• The Fusor was quite small - the anode was a sphere 6 inches in diameter, and the cathode was slightly over 4 inches in diameter, and the electron guns were less than 3 inches in length.
• The cathode sphere is 60 mil 304 stainless steel; the anode is 93 mil 304 stainless steel.
• Farnsworth's cathode shell is only approximately spherical in shape - there are 12 flat areas where the biasing screens mount. It should be possible to form perfect hemispheres, then weld them together, or to have spheres made from 304 stainless (I know of an Internet Web-site, which is for a company that specializes in making custom balls.) The bias screens could have flanges machined to fit the radius of curvature, and the screens themselves could also be curved to help maintain a more spherical catho dic field.
• Farnsworth apparently is sacrificing two bias screen areas for the mounting flanges for the inner cathode - this should not be necessary! The mounting points can be relocated between the bias screen positions.
• Note in Figure 11, at one end of rod 58a, there is an undescribed cone shaped washer with a lip placed near the opening into tube 50. This is likely an electron shield, which can keep stray electrons in the anode space from traveling down tube 50. The cone shaped washer may instead serve as an electrostatic field shaper placed near the opening. Figure 23 does not show this important item.
• Referring back to patent 3,258,402 - Electric Discharge Device for Producing Interactions Between Nuclei, I have to note that this early design required a source of RF power, and the newer design apparently does away with this. Also the early design used a total of four spherical electrodes, one of which had deep cone-shaped ion guides, which would have been difficult and expensive to machine. The new design only has two sphe rical sheet-metal electrodes, which are relatively inexpen sive. What parts do require machining are relatively simple of design, and no exotic metals are used. Indeed, it should be possible to stamp the spheres out of stainless steel sheet metal, almost cookie-cutter style!

Full patent

• This patent is by Robert Hirsch and Gene Meeks, associates of Farnsworth. As you will see, this patent is a Farnsworth Fusor, but the patent was issued after Farnsworth's death. I know that Hirsch and Meeks did a lot of the Fusor work in the Farnsworh labs, and many of the successful Fusor innovations were theirs. I will save a detailed examination of the patent details for another time, but please study this patent and compare Meek's and Hirsch's changes to Farnsworth's own last patent.

Full patent

• This patent is by Robert Hirsch, an associate of Farnsworth. As you will see, this patent is a Farnsworth Fusor, but the patent was issued after Farnsworth's death. I know that Hirsch did a lot of the Fusor work in the Farnsworh labs, and many of the successful Fusor innovations were his. I will save a detailed examination of the patent details for another time, but please study this patent and compare Hirsch's changes to Farnsworth's own last patent.

Return to Home Page