Construction Classics

Commissioning practice

Almost every construction report ends at first beam. These three cover what comes after: two consecutive 1952 quarterlies from Oak Ridge that read as a running log of three cyclotrons being commissioned side by side, and a 1966 study of retrofitting an axial ion source into a machine that already worked. Commissioning detail is the scarcest content class in this collection, because it is the part nobody writes up for publication.

Why these documents can be hosted

The two quarterlies are Oak Ridge National Laboratory reports produced under AEC contract W-7405-eng-26, which appears on the title page of ORNL-1339. AEC contractor reports are treated as US Government works and carry no copyright, and pre-1978 distribution without a copyright notice is an independent public-domain basis. ORNL-1345 is the released (Del.) version and is marked Unclassified throughout; neither report carries classification markings beyond the routine distribution controls of the period.

UCD-CNL-49 is an AEC Research and Development Report from Crocker Nuclear Laboratory, stamped “released for announcement in Nuclear Science Abstracts” — an explicit public-release marking. It also carries the conference number CONF-660906-1, indicating it was prepared for a September 1966 conference. That is an AEC accession number, not a publisher imprint: the hosted copy is the laboratory report, and no copyright notice, publisher marking, or proceedings imprint appears anywhere in the scan. Any version of this work later published in commercial conference proceedings would be separately copyrighted and is not what is hosted here. Verified against the scans and the OSTI records; determinations made August 2026.

Three machines, one quarter at a time

The value of the ORNL quarterlies is that they were written to report progress to a sponsor, not to persuade a journal — so they record what was actually happening to the 86-inch, 63-inch and 22-inch cyclotrons in the same weeks, including the parts that were not working. Each report opens with a summary before the per-machine sections.

ORNL-1339 (quarter ending 31 March 1952) carries the single most quotable commissioning number in the pair: the newly assembled 63-inch held a 75 kV dee-to-dee voltage under vacuum for long periods without failures, against a 60 kV design figure — a measured margin over specification, reported plainly. On the 86-inch, relocating the ion source increased output, and operation was interrupted in January 1952. The 22-inch section covers an investigation of the z-wise proton beam distribution — that is, beam extent parallel to the field axis, which is the measurement an amateur builder is most likely to skip and most likely to need.

ORNL-1345 (quarter ending 30 June 1952) picks the same three machines up three months later. The 86-inch had a 41 kW beam calorimetered — beam power measured as heat, the honest way to do it. The 63-inch heavy-particle machine was in operation producing N³⁺, with a new ion source ready for test. The 22-inch had moved on to the problem of getting ions from the source into the dees, which is the same central-region transport problem the Davis report below attacks fourteen years later with better tools.

Read consecutively, the pair shows something a single report cannot: the rhythm of commissioning. Machines are interrupted, sources are relocated, a design figure is beaten, and the next quarter’s work is set by what broke in the last one.

Retrofitting a running machine

UCD-CNL-49 is a short, focused study of what it would take to put an axial ion source into the Davis 76-inch — a machine that already worked with an internal source. The central-region orbit calculations run at comparatively low dee voltages of 20–30 kV in order to centre the orbits, and the report is candid that particle beams were not obtained for all the cases discussed. What grounds it is the empirical check: start-up data from 42 MeV alpha particles, and an internal beam of 21 MeV hydrogen actually obtained on the machine.

For a small-machine builder the interest is in the framing rather than the specific geometry. An axial source moves the source out of the median plane, trading a complicated injection path for an uncluttered centre and a source you can service without opening the chamber — the same trade a tabletop builder faces, argued here with orbit calculations and confirmed against a real machine.

Design rules extracted from these documents

36 Design Guide rules come from these three reports — ORNL-1339, ORNL-1345 and UCD-CNL-49 — and they came out as predicted: measurement practice and what-to-expect margins rather than design formulas. That is the scarcer half. Representative examples:

• dg-797 — establish beam energy by at least three independent methods before quoting it
• dg-783 — condition the RF past its design dee voltage and hold it there
• dg-785 — build a self-calibration into beam calorimetry — an electric boiler of known wattage
• dg-787 — ion-source axial position is a first-order energy and beam-quality knob
• dg-801 — prove first beam with a radiation signature plus a physics argument, not probe current alone
• dg-804 — make ion-source position adjustable from outside while the machine runs
• dg-818 — sequence commissioning around your shielding, using a heavier, slower species first

Used by the ion sources and beam measurement deep dives.