Biconic cusp

Teller noted that any design that had the plasma held on the inside of concave magnetic fields would be naturally unstable.

The concept was independently presented in 1954 by both Harold Grad at the Courant Institute in New York and James L. Tuck at Los Alamos.

This sort of "bad curvature" was part of almost all designs of the era, including the z-pinch, the stellarator and the magnetic mirrors.

Nevertheless, a number of researchers began considering new concepts that did not use this sort of field arrangement and would thus be naturally stable.

The cusp concept was independently developed in 1954 by James L. Tuck at Los Alamos and Harold Grad at New York University.

[4] A single-cusp version was seen as a simpler device to test the concept, and a magnet assembly for one such machine was built at Los Alamos.

[3] Calculations at Los Alamos noted that the plasma would escape the reactor because the magnetic lines were "open" and ions following a certain trajectory would be free to leave the core.

This meant the picket fence would lose plasma at a fast rate, no matter how stable it was, and it would not be useful as a power-producing reactor.

[5] Before the system was considered further, results from newer versions of the other designs all seemed to be suggesting Teller's issue was simply not being seen, or was at least far below predictions.

Among them, the pinch concept had been demonstrating serious problems, but Tuck and others had continued studying the system and were introducing new solutions.

Months later, they were forced to publish a retraction, noting that the neutrons they saw were not from fusion events, but a new type of instability that had not been previously seen.

The power supply for the early machine at Los Alamos had been sitting in a warehouse for years, and was then taken out of storage and used to build the single-cusp Picket Fence I.