Construction Classics

The Oak Ridge 86-Inch Cyclotron

Why this document can be hosted

ORNL-1196 was produced by Oak Ridge National Laboratory under AEC contract W-7405-eng-26 and issued by the AEC; AEC contractor reports are treated as US Government works, which carry no copyright, and DOE/OSTI hosts the report openly today. Public distribution before 1978 without a copyright notice stands as an independent public-domain basis. Verified against the scan’s own header and the OSTI record; determination made August 2026.

Reading guide

This is the fullest construction-and-commissioning record in the open cyclotron literature: a machine that went from excavation in September 1949 to a 1.2 mA, 23 MeV internal proton beam by mid-1952, with every subsystem described, photographed, and costed. Page references are the report’s printed numbers, which match the hosted PDF’s pages.

The theory section, pp. 17–19, earns its place by explaining threshold dee voltage and the brutal arithmetic of the machine center — electric-field defocusing loses roughly 90% of starting ions to the dee surfaces (p. 18). Performance, pp. 20–27, is honest in a way later reports rarely are: sparking, not any meter reading, set the beam-current ceiling (p. 24), and measured wall-plug-to-beam efficiency ran 2.6–9%. The magnet chapter, pp. 29–37, covers a yoke welded from 2-inch plate, oil-cooled coils with winding-temperature alarms (p. 33), 65-turn control coils for steering the beam onto the median plane (p. 35), and azimuthal shims machined from 2¼-inch plate on a boring mill, giving a field that falls 1% per 13 inches of radius (p. 35, Fig. 9).

For fabrication substance, three chapters stand out. Vacuum, pp. 38–46: a 450 ft³ chamber sealed with continuous square rubber gaskets in grooves sized for metal-to-metal flange contact (p. 41), dry-air backfill in place of a refrigerated trap (p. 44), ion and Penning gauge practice (pp. 45–46), and a rate-of-rise tightness benchmark (p. 46). RF, pp. 47–61: dees cooled by furnace-brazing flattened copper tubing to 1/8-inch plate (p. 50), the shorting “spider” clamped to silver-plated stems with water-cooled split rings (p. 53), the oscillator sited where stray field stays below ~60 oersteds with copper-lined cabinets and insulated water leads (p. 59), and a plate supply built from paralleled, individually fused units (p. 61). Ion source, pp. 62–67: a graphite arc chamber and two independent servo loops holding arc voltage and current (p. 66). Targets, pp. 68–74, show grazing-incidence water-cooled designs that absorbed 41.7 kW. The development chapter, pp. 97–109, is the collection’s best commissioning narrative — eight months of ion-loading trouble before reliable high-current running, and exactly what was changed. Appendices give the itemized cost analysis (pp. 112–113), field-measurement equipment (pp. 114–117), the RF design worked through a full-scale model (pp. 118–130), and proton energy measured with copper foils (pp. 131–133).

What scales down: the joint and seal discipline, gauge practice, the dee-power budget method, and the commissioning sequence — find beam, authenticate it, then chase current. What does not: the 200-ton magnet, oil-cooled coils, milliampere beam handling, and the target metallurgy, which exist because this machine ran 30 kW of beam power.

Design rules extracted from this document

27 Design Guide rules cite ORNL-1196. Representative examples:

• dg-172 — gentle radial field decrease for axial focusing
• dg-313 — dee excitation power budget from P ≈ 2πfCV²/2Q
• dg-317 — clamped, silver-plated, water-cooled RF joints
• dg-361 — perforated dee walls and graphite facing
• dg-477 — rate-of-rise as the chamber tightness benchmark

Used by the magnet design, ion sources, beam measurement, and beam quality deep dives.