Construction Classics

The Rochester 27-inch

Most of this site about building a cyclotron ends at first beam. These five come from one 27-inch machine in a university physics department across eight years, and between them they cover both halves of its life: getting the thing to run, and then getting work out of it. Two are machine reports — starting an oscillator that will not start, and making more ions from less filament. Three are what came after — getting the beam out and analysed, building a spectrograph around it, timing neutrons with it.

The pairing is the point. The same laboratory that published a broad-range spectrograph also had to publish a note about a spark gap, because the machine would not come up to voltage. Very few institutions wrote the second kind down.

Why these documents can be hosted

All three are University of Rochester reports in the AEC’s NYO- series (New York Operations Office), NYO-3823 under contract AT(30-1)-875, each carrying the standard US Government sponsorship disclaimer. AEC contractor reports are treated as US Government works and carry no copyright, and pre-1978 distribution without a copyright notice is an independent basis.

Two of the three carry something better than inference. NYO-3823 and NYO-9360 are stamped “This document is PUBLICLY RELEASABLE” with an authorizing officer — an explicit release determination on the document itself rather than a rights argument built from its series and date.

Checks recorded: the “IEEE” strings a text search finds in NYO-3823 are OCR garbage from its figure regions, not a publisher marking; the “thesis” and “Proceedings of” hits in NYO-3823 and NYO-9683 are reference-list entries; and NYO-9360’s “to be published in Nuclear Instruments and Methods” describes a privately circulated monograph it cites, not this report. Verified by reading the scans; determinations made August 2026.

A small machine is a beamline problem

The Rochester 27-inch is the size class this site is about, and these reports show what a department does with one over a decade. That trajectory is worth seeing whole, because the builds census records no amateur machine that has yet produced an external beam — so the amateur record stops precisely where this one starts.

NYO-3823 (1954) is 53 pages on getting the beam out and making it useful: a focusing and analyzing system for the extracted beam. It is the longest of the three and the most directly relevant to a builder who has beam inside a chamber and wants it somewhere else. Pair it with the extraction hardware cluster, which covers getting the beam past the septum; this covers what happens after.

NYO-9683 (1961) is a broad-range spectrograph designed around the machine — an instrument sized to a specific small cyclotron rather than a general design. Instrument papers rarely state the constraints their host machine imposed; this one has to.

NYO-9360 (1962) is a fast-neutron time-of-flight system, and it is the most transferable of the three in one specific way: timing. A time-of-flight measurement needs a start signal referenced to the accelerating RF, which makes it a diagnostic of the machine as much as of the reaction. Its title says “for use with cyclotrons” in general, not just this one.

What these are not

None of the three is a construction report. They will not tell you how the 27-inch was built, and the machine itself is documented elsewhere — Rochester’s 8 MeV variable-energy cyclotron report NYO-6541 remains in the queue. Read these for the layer above the machine: what external beams, spectrographs and timing systems demand of a small cyclotron, which is often how a builder discovers a requirement they should have designed for.

Design rules extracted from these documents

53 Design Guide rules come from these reports. Representative examples:

• dg-965 — architect an external beamline as condenser → shielded slit → analyzer
• dg-966 — put the beam-defining slit inside the shield wall — the slits intercept most of the beam
• dg-967 — prefer a strong-focusing quadrupole pair over a sector magnet: about ten times less space
• dg-970 — hyperbolic quadrupole pole profiles are not worth precision machining at this scale
• dg-988 — a cyclotron needs no beam sweeper for time-of-flight — the beam is already RF-bunched
• dg-989 — derive the timing reference from the oscillator, not from a beam-intercepting pickup
• dg-993 — time-resolution budget honesty: 2 ns FWHM at best, 2–3.5 ns typically
• dg-1276 — a third multipactor cure: shock-excite the dee past the top of the loading band
• dg-1280 — the decay envelope of a ringing dee maps the multipactor band edges
• dg-1284 — a passive magnetic mirror from a 1/8-inch steel bearing ball at the top of the arc
• dg-1287 — filament life scales strongly with the emission the arc demands of it

Used by the beam measurement and beam extraction deep dives.