High School Fusion Reactor: Project Overview and Progress Thus Far (As Of Last Summer)

13 min readSep 11, 2021

--

At a few of my mentors’ suggestion, I have decided to publicly document my progress designing and building an “inertially confined, electrostatically accelerated and focused, particle collider, within velocity space,”⁴ or a Farnsworth–Hirsch Inertial Electrostatic Confinement Nuclear Fusion Reactor (referred to colloquially as a fusor).

Note: This blog should not serve as a guide on fusion in any way. I am (as of now) inexperienced in this area and could be misinformed or entirely incorrect about the subjects discussed (including the incredibly dangerous high voltage and radiation). Those looking for the experiences, successes, and failures of an overenthusiastic, young, aspiring fusioneer are in the right place. Those looking for a fusor journey to emulate should refer to this post and those looking for more information should refer to my citations or fusor.net in its entirety.

Note 2: (Despite its length) This post is an overview of my fusor progress thus far. I plan to make more in depth posts soon.

Note 3: All terms used here are (I think) explained in this glossary.

Note 4: This post is not up-to-date with my progress or knowledge. It serves as a starting point where the project is explained.

Background:

A fusor is defined as “a device that uses an electric field to heat ions to nuclear fusion conditions.”¹ Such devices have been the object of billions of dollars of research over the approximately 70 years since the technology’s inception, due to their promise of clean, efficient, safe nuclear energy.² Despite this incredible scientific effort, self sustaining (or energy positive) “cold fusion” has not yet been achieved.

In parallel with industrial and research scale fusion efforts, a small, grassroots group of DIY enthusiasts has sprung up around the technology. These “fusioneers” build small, comparatively low-power devices in basements and workshops that, despite their limitations, achieve verifiable “hot fusion” (or energy negative) reactions.²

I first discovered this community when I read about the (then) youngest person to build a fusor³ when I was middle school. Despite my then severely limited resources and knowledge, I knew that I had to build one myself. The summer between my sophomore an junior year of high school, my chance came and I began my ongoing project to learn as much as I can about “desktop fusion” and eventually, hopefully build a fusor capable of significant, measurable fusion.

(Demo) Fusor 0:

After reading Richard Hull’s (a prolific mentor on the fusor.net forum) Beginners Beginnings for fusion,⁵ I realized that my goal is much farther away than I initially thought. After a bit more reading, I learned that I should first begin with a demo fusor, a device that demonstrates the same technologies as a fully-featured fusor without the complexity and danger of its most difficult aspects. A demo fusor does not, of course, achieve fusion.

My design, based upon a guide published in Make Magazine⁶, served as a test-bed for the two major parts of any IEC fusor design: high vacuum and high voltage (both unhelpfully abbreviated as HV). A “real” fusor creates plasma in a vacuum environment saturated with deuterium gas. These deuterons are accelerated together and eventually fuse. A demo fusor just creates plasma in a vacuum (usually with significantly lower voltage and a more primitive vacuum system).¹²

I was able to find an old 12kV neon sign transformer (NST) on eBay⁷ ⁸ that would connect to a variable AC source (variac) to supply high voltage DC via a homemade half-wave diode rectifier.⁹ ¹⁰ This high voltage (positively) charges a steel wire geodesic grid¹¹ suspended from an insulated feed-through in a “bell jar” style vacuum chamber¹² made of a borosilicate cylinder sandwiched between two circular aluminum plates. The voltage difference between the charged grid and grounded top and bottom metal flanges causes plasma to form in the center of the grid when the chamber is evacuated to vacuum and the transformer is activated.

Test setup for my NST, the right wire was secured to the table and the left one was taped to a wooden rod. At this point, I had no way to accurately measure voltage, so I watched to see if an arc formed between the wires. When one did, I assumed that the NST was working properly.

I created the grid out of three 1" diameter steel loops connected with silver solder. It was then attached to a steel screw which threaded into the feedthrough. The process was very time intensive as I had never silver soldered before and the process is very different from normal soldering.

I used copious amounts of vacuum grease to seal the chamber, this made disassembly and cleaning very difficult.

This was my full setup for fusor 0. The diode rectifier was sealed in mineral oil inside the PVC tee. For cost reasons (as this was a proof of concept and did not require instrumentation), I used an analog vacuum gauge that bottomed out and could not give accurate reading of vacuum levels.

An image of the first full test of fusor 0. Plasma was formed successfully around the grid.

A subsequent run done at night with the lights in my workshop momentarily turned off to better observe the plasma. The purple hue is due to the white balance on my phone camera. The actual color of the plasma in all tests was a much lighter bluish-purple.

During the test, a black residue was deposited on the borosilicate tube. This is called sputtering¹³ and, while useful in some contexts, required the full disassembly of the system for cleaning.

After the first successful test, I made an embarrassingly overconfident plasma club¹⁴ application on fusor.net (my first post) and was rejected (rightfully so) for my lack of any instrumentation.

During disassembly for cleaning, I replaced the grid entirely as I believed (potentially mistakenly) that the bulbous solder joints were a problem. The newly reassembled fusor was referred to as fusor 0.1.

I conducted three successful runs of 0.1, all with identical parameters (120Vac applied to the transformer, vacuum pump left on for 5 minutes before activation). The first two tests achieved similar results, so I decided to momentarily turn the lights in my workshop off during the third run to better observe the plasma (that test lead to the photo above). Videos of all tests will be provided in a separate post.

During the fourth run, the grid failed when a solder joint near the top of the grid disconnected, causing plasma to form solely around the disconnected point, instead of in the center of the grid. Additionally,
(what I believe to be) corona discharge¹⁸ occurred on the bottom plate.

Plasma formed solely around the failed solder joint (top) and between the feedthrough ceramic and bottom plate.

I believe that these two issues are not correlated and that their occurring at the same time was due more to my bad monitoring of the system then some common cause.

I believe that the grid failed because I used too little silver solder on the 0.1 iteration. I tried to limit my solder usage for this grid to avoid bulbous joints which I believed to be a problem (I still do not know if they are or are not an issue), and, due to my inexperience, might have created a seemingly secure joint out of mostly flux that melted away when the system was activated repeatedly. Examination of pictures taken of the grid between tests three and four show no visible weakening of the joint, but examination of the grid after the fourth test shows that there is a clear gap between the wire loop.

I do not currently have a comprehensive understanding of why plasma formed on the bottom plate, but it is my current belief that the sharp edge between the bottom plate and ceramic feed-through (created when I drilled the feed-through hole in the plate myself), acted as a breakout point for corona discharge¹⁸. I am unsure of what changed between the third and fourth tests that caused the issue to become so pronounced, but upon re-examination of video from the third test, the same discharge is visible to a lesser extent.

It is therefore my hypothesis that the sharp edge had been partially insulated from the system by the epoxy holding the feed-through in place. Over the four tests (one 0.0 test and three 0.1 tests) before the failure, the epoxy melted away, exposing the sharp edge to the voltage differential and allowing the discharge to occur.

Discharge from bottom plate visible to a lesser extent in images from the test preceding the failure. For subsequent tests, I will switch to using my mirrorless camera instead of my phone to avoid blown-out video.

As both discharge and grid failure occurred at the same time, I do not know which problem lead to the unsuccessful run (the images of tests three and four were taken with the same amount of voltage supplied to the system and test four shows a significantly lower amount of plasma). My hypothesis is that the grid failure was the more catastrophic of the two, as the discharge failure was present (to a lesser extent, as shown) in previous tests but I cannot be sure with my current information.

I plan to disassemble the chamber a second time to clean the sputtering residue on the chamber body and replace the failed grid with a new one manufactured with a different technique. I will continue using steel wire and silver solder for fusor 0, but will use a solder paste recommended on fusor.net¹⁹ instead of the silver bars and brazing flux I have been using. I will also further inspect the base plate and insulate any sharp points with epoxy. Hopefully, these repairs and improvements will bring fusor 0 back to its previous functionality for its 0.2 iteration.

Fusor 1:

Despite the continued potential of fusor 0, its ineffective vacuum system, weak power supply— less than half of what is required for fusion, steel grid (as opposed to tantalum or some other material)¹¹, and dangerously unshielded vacuum chamber combine to make it (despite my endless bragging at school) a primitive proof of concept at best. For this reason, I’ve decided to take what I have learned from fusor 0 and apply it to the construction of a much more advanced “research grade demo fusor.”¹²

This sort of fusor, also classified as a Proto Fusor,¹⁵ is a “fusor on its way to being a working, fusing, neutron producing fusor”¹² and will entail a significant amount of work and investment.

The clear first upgrade is the vacuum chamber itself: compared to its importance all other upgrades are peripheral. The ideal chamber is “two conflat flanges connecting two stainless steel, (SS), hemispheres to make a vacuum tight spherical chamber.”²⁰ These hemispheres would ideally form a sphere with a 6" diameter within which a 1.2" diameter²¹ tantalum¹¹ grid would be suspended. Such a chamber was projected to cost approximately $5,000 to build.

Such an expense is far outside of my budget, so I took to eBay to scrounge for serviceable backup options. After more research, and a seemingly never-ending cycle of finding an affordable chamber and then learning that it was either damaged or otherwise unusable for my purposes, I found a square chamber²². I had (and have) not seen a square fusor chamber, so I made a post on fusor.net requesting advice.

An image taken from the eBay listing of the square chamber that I considered using

I was met with quick, useful advice, and responses to my followup questions, suggesting (among other pieces of advice) that I use a “[4.5”] conflat 4- or 6-way cross chamber” instead of the aforementioned square one as the limited ports on the square would make conversion from a demo fusor into a fully featured one very difficult.²³

I then located a 2.75" by 4.5" conflat 4-way cross chamber on eBay for an attainable price that I thought would be the perfect option. I drafted a simple design schematic incorporating the chamber based upon two (frequently conflicting) ideals: The design should be modularly upgradable such that I could interface old and new parts to avoid buying everything at once; and the design should be a Proto Fusor, in that every new component that I buy (within reason) should not have to be replaced to do real fusion.

The first version of my Fusor 1 design. At this point, I still had significant questions regarding how my vacuum system and HV feed-through should work.

Shortly after completing my first design iteration, I stumbled across a post made on the fusor.net “trading post,” ²⁴ a place for fusioneers to buy and sell components. The post²⁵ advertised “Working Fusor Parts for Sale” and offered a thermocouple (TC) vacuum gauge and tanalum grid connected to an HV feed-through, two components that I was having difficulty finding used or surplus. Although the post’s writer, Stephen Haid, saying that he would not ship components, I emailed him and was able to arrange shipping of the components to my home in NC. Thrilled, I drafted a second fusor 1 design that incorporated his components.

This design incorporated Stephen’s feedthrough, tantalum grid, and TC gauge while being otherwise very similar to my V1 design

Shortly after finishing my second design, and while I was in the process of discussing the shipping of the TC gauge and feed-through/grid, I ordered the 2.75" by 4.5" conflat vacuum chamber. When the chamber arrived, I took its complete internal measurements and realized that I had made a significant mistake.

A top-down view of the 2.75" by 4.5" CF chamber that I ordered off of eBay. The 4.5" CF ports are the two large ports at the top and bottom. The two 2.75" CF ports are on the sides. All ports were wrapped in foil for shipping, I believe this is to preserve the “knife-edge” of the CF connection.

A view inside a 4.5" CF port on the first chamber that I ordered from eBay. The 1.5" grid would rest at the center of the cylinder, suspended from a feed-through inserted through one of the 2.75" ports.

Upon measurement, I realized that the inner diameter of the chamber was 2.35" and the inner diameter of the side ports was 1.37". This would have been clear had I looked at a conflat documentation table, but I did not have the foresight to check one. The issue arises from the fact that the tantalum grid that I was in the process of ordering had a diameter of 1.5 inches.

Disregarding the fact that it would be impossible to fit a 1.5" grid through a 1.37" tube, the grid would also be significantly too large for the chamber. According to Richard Hull, the ideal grid:chamber ratio is approximately 1:4 or 1:3.²¹ Stephen’s grid in this chamber would have a ratio of approximately 1:1.6. Additionally, a 4.5" diameter chamber is suggested for a 30kV supply, the minimum recommended for fusion. This all combined to show that, while the chamber that had just arrived could do fusion, it would be much more difficult — and impossible with the grid I had just bought.

Luckily, the chamber had a 30 day return window.

As the grid diameter was fixed (I could hypothetically make my own grid, but I do not have the equipment or expertise), I began the search anew for a vacuum chamber that not only fit my initial requirements, but also the 1:3 grid:chamber ratio laid out for me.

After some more searching, I located a 6" CF chamber that, I believe roughly fits the above requirements. With a 4" inner diameter, it has an unfortunate 1:2.7 grid:chamber ratio, and is only suited for 26.7kV according to Richard Hull’s post²¹. Although not perfect, the equation used to calculate my 26.7kV maximum voltage was designed with significant margin as a recommendation for new users and “rule follower, lock-step types.” Additionally, a 4" diameter has elsewhere been described as suitable — albeit very small — for serious fusion.²⁶

While the diameter of the chamber is not a significant issue alone, the 1:2.7 grid:chamber ratio is cause for concern. Unfortunately, I cannot replace the 1.5" grid on Stephen’s feed-through, and other than the grid diameter his feed-through is perfect. Hopefully by the time that the size becomes a serious worry (not until I begin to make the upgrade to actual fusion, I don’t think) I will have the capability to replace the grid or the size will reveal itself to be a non-issue.

After selecting my vacuum chamber, I made the decision to also upgrade my power supply from a 12kV NST unsuitable for fusion³⁷ to a 30kV x-ray transformer proven in a fusion system and also for sale by Stephen. With these changes made, I developed my latest fusor 1 design.

My latest Fusor 1 design incorporates my new 6" CF chamber and the TC gauge, feed-through, grid, and x-ray transformer from Stephen.

While this design is a huge leap forward in my path to fusion, it is not without its flaws. Besides the inner diameter issue, the actual vacuum pump system is still primitive compared to what it will eventually need to be for fusion (I can’t explain the depth of this issue concisely so enjoy a long list of citations to peruse).²⁷ ²⁸ ²⁹ ³⁰ ³¹ ³² ³³ ³⁴ ³⁵ Additionally, this diagram lacks voltage and current instrumentation. I will definitely be including measurement devices in this design, I simply did not include it to avoid complicating the drawing. A much more in-depth design will be created and posted here before Fusor 1 is completed.

In short, fusor 1 is intended to take me from demo fusor to Fusor, and to do so affordably and safely. All issues outlined above will have to be solved before I can even hope to add the complexity — and danger — of a deuterium feed-through and fusion measurement, not to mention the worrying reality of x-ray radiation, but I am confident (this time, hopefully beyond the Dunning–Kruger curve³⁶) that such a goal is achievable.

As fusor 1 begins to move from paper to my workshop, I will continue work on fusor 0 (making the repairs and upgrades mentioned above) for two reasons. The first is that I would like to figure out what caused the failures exhibited in my last run. Secondly, once fusor 1 reaches its final iteration, no parts from fusor 0 will carry over, so I will have a self-contained, rough demo fusor that I can use for demonstrations (as even a proto fusor can create dangerous x-ray radiation) and as a physical record of my fusion journey.

This (longer and more in depth then originally imagined) post is intended to serve as an overview of my fusor progress thus far and an outline of my immediate aspirations. I intend to update with new posts whenever significant developments occur or I make progress. Future posts will probably be a combination of in-depth “report style” additions documenting experiments and fusor runs and more broad progress updates. If you have any questions or suggestions, please comment them below!

Citations

¹ https://en.wikipedia.org/wiki/Fusor

² https://fusor.net/ Note: fusor.net is generally very informal in the presentation of its information. Despite this, I find the information cataloged there to be extensive and generally correct.

³ https://www.cnet.com/news/13-year-old-builds-working-nuclear-fusion-reactor/ Note: These types of articles are often sensationalized at best or actively incorrect at worst. For actual fusion information, refer to fusor.net or more credible journals. That said, this kid has a significantly more advanced system than I currently do and I’m definitely not jealous.

⁴ https://www.fusor.net/board/viewtopic.php?f=42&t=10775

⁵ https://www.fusor.net/board/viewtopic.php?f=24&t=3247

⁶ https://makezine.com/projects/nuclear-fusor/ Note: This guide has been criticized at length on fusor.net for incorrectly implying that the demo fusor described could achieve fusion. Despite this, it is the first demo fusor that many build and is the highest quality, most comprehensive guide I could find.

⁷ https://www.fusor.net/board/viewtopic.php?f=29&t=8710

⁸ https://www.fusor.net/board/viewtopic.php?f=29&t=10333

⁹ https://www.electronics-notes.com/articles/analogue_circuits/diode-rectifiers/half-wave-rectifier-circuits.php

¹⁰ https://en.wikipedia.org/wiki/Rectifier

¹¹ https://www.fusor.net/board/viewtopic.php?f=24&t=8660

¹² https://www.fusor.net/board/viewtopic.php?f=24&t=2674

¹³ https://en.wikipedia.org/wiki/Sputtering

¹⁴ https://fusor.net/board/viewtopic.php?t=13&f=7&sid=c03634687c8297889a4d61478c773b2d#p512

¹⁵ https://fusor.net/board/viewtopic.php?f=45&t=14105

¹⁶ https://www.fusor.net/board/viewtopic.php?f=29&t=4267

¹⁷ https://www.fusor.net/board/viewtopic.php?f=29&t=4266

¹⁸ https://en.wikipedia.org/wiki/Corona_discharge

¹⁹ https://www.fusor.net/board/viewtopic.php?f=24&t=12142

²⁰ https://www.fusor.net/board/viewtopic.php?f=24&t=2978

²¹ https://www.fusor.net/board/viewtopic.php?f=24&t=8765

²² https://www.ebay.com/itm/224580357520?hash=item344a084d90:g:moAAAOSwjLBhIFai Note: As of the writing of this post, the chamber is still for sale. I, of course, do not recommend using it as a fusor chamber.

²³ https://fusor.net/board/viewtopic.php?f=46&t=14150

²⁴ https://fusor.net/board/viewforum.php?f=73

²⁵ https://fusor.net/board/viewtopic.php?f=20&t=14154

²⁶ https://www.fusor.net/board/viewtopic.php?f=24&t=13508

²⁷ https://fusor.net/board/viewtopic.php?f=25&t=9291

²⁸ https://fusor.net/board/viewtopic.php?f=25&t=8626

²⁹ https://fusor.net/board/viewtopic.php?f=25&t=13525

³⁰ https://fusor.net/board/viewtopic.php?f=25&t=3878

³¹ https://fusor.net/board/viewtopic.php?f=25&t=13706

³² https://fusor.net/board/viewtopic.php?f=25&t=3909

³³ https://fusor.net/board/viewtopic.php?f=25&t=4213

³⁴ https://fusor.net/board/viewtopic.php?f=25&t=4142