Construction Classics

Design of the Radio-Frequency System for the 184-inch Cyclotron

Why this document can be hosted

UCRL-64 is a UC Radiation Laboratory report produced under AEC contract W-7405-Eng-48 — the contract number appears on the title page and again in the closing acknowledgment — and issued through the AEC’s standard 80-copy distribution list, which is bound into this scan. AEC contractor reports are treated as US Government works, which carry no copyright; pre-1978 distribution without notice is an independent public-domain basis. The manuscript was later published in the Journal of Applied Physics, but this copy is the laboratory report, not the journal reprint. Verified against the scan and the OSTI record; determination made August 2026.

Reading guide

Written by the four engineers who built it, this is the design history of the RF system that drove the first large synchrocyclotron — and it says so plainly: much of the reasoning is given “as it is felt this may be useful to others designing frequency-modulated oscillators for cyclotrons.” Page references are the report’s printed numbers; the hosted PDF runs three pages ahead, with LBNL library cover sheets in front.

The general-features section (pp. 5–6) explains the architecture: a single dee and a separate rotary capacitor joined by a half-wave transmission line, chosen because the 37-inch cyclotron had already proven it — the “exceptional ease of operation” of the 184-inch is credited directly to copying a tested system. Seven numbered reasons for the choice follow, from keeping the rotary capacitor out of the magnetic field to the ability to bias the dee. The resonant-system design (pp. 6–10) carries the fabrication detail: four parallel lines to reach roughly 10 ohms, copper-plating on every steel surface exposed to RF, and door seals made of soft copper strip backed by foam rubber — the RF-gasket trick still used on homebuilt amplifier enclosures. The insulator test rig on p. 10 held 50 kV RF at 13 Mc across the 15-inch feed-through insulators before any were committed to the machine, and parallel breakdown tests set the rotary condenser blade gap: 0.080 inch holds 50 kV at 5 × 10⁻⁶ mm.

Scale-model tests (pp. 10–12) show how the quarter-scale model set the line length and capacitance range, and taught one transferable lesson: bare steel condenser blades halved the Q, so everything got copper-plated. The oscillator chapters (pp. 13–19) detail the grounded-grid 9C21 circuit — coupling loops slotted into the transmission lines, a grid grounded through a ring of fifteen vacuum capacitors, a 1400-ohm grid leak built from twenty-eight 200-watt resistors, and a steel oscillator house that cuts the stray field at the tube from 140 to under 20 gauss. Operating experiences (pp. 19–22) are the candid part: a cavity mode between the pole pieces cured by strapping them together, a cumulative two-inch model error that forced removal of one transmission line, RF discharges around the dee (resonant-electron and multipactor) suppressed with perforated shields, a grounded dummy dee, and negative dee bias, and warped stainless-steel condenser disks replaced with copper. Running power: about 40 kW for 20–30 kV on the dee, close to the model prediction.

The scale is wrong for a tabletop machine, but the discipline is not: model the resonant system before building it, test insulators to destruction outside the machine, plate the steel, and expect the discharge physics — the dummy-dee and negative-bias cures here are the standard answers still. The hosted copy is an OSTI scan of an LBNL library print with a rough but searchable text layer; several figure plates at the back are faint.

Design rules extracted from this document

13 Design Guide rules cite UCRL-64, concentrated in RF fabrication craft and discharge suppression. Its companion UCRL-476, by co-author Wouters, remains the most-cited hosted document in the collection. Representative examples:

• dg-658 — demountable RF joints from soft copper strip backed by foam rubber
• dg-659 — copper-plate every steel surface exposed to RF fields
• dg-662 — budget vacuum RF gaps from bench data, then derate for surface roughening
• dg-667 — cure multipactor discharges with perforated shields, a dummy dee, and bias
• dg-669 — mount ceramic insulators in pure tension or compression, never shear

Used by the dee RF coupling deep dive.