Design Guide › level 5
Cyclotron design rules, level 5: large, unusual, or professional machines only
29 of the guide’s 1878 rules sit at level 5: large, unusual, or professional machines only.
The level ranks how early and how universally a rule binds a cyclotron design — breadth,
never weight. It is not permission to skip a rule whose trigger a machine has, and safety
rules are never skippable on level alone; how the levels were assigned and audited is on
the methodology page. Each rule keeps its
formula where the source gives one, a verbatim quote, a page-level citation, and a stable
identifier (dg-NNNN) that resolves here, on its
subsystem page, and on the
all-in-one guide. Where an editorial note says “the
reference machine”, its parameters are on the
guide’s front page.
This level’s rules by subsystem — each link opens just that subset, in the all-in-one guide’s filters: Extraction (11) · RF (9) · Beam dynamics (5) · Magnet (5) · Safety (5) · Fabrication (3) · Project management (3) · Cyclotron general (2) · Materials (2) · Shielding (2) · Targets (2) · Vacuum (2) · Beam measurement (1) · Coils (1) · Dee (1) · Detectors (1) · Ion source (1) · Modeling (1). A rule carrying several tags is counted under each; a subsystem’s complete rule set, across all levels, is on its own page in the subsystem directory. To add a search term or a second subsystem, open this level in the all-in-one guide with filters, which carries every rule and filters in the browser.
Verify before use. Every rule here is a source extract in the vocabulary of the editorial methodology — faithful to its cited page, not an independently validated engineering requirement. Re-read any rule that drives a real design decision at the cited page before committing metal, money, or high voltage to it. The editorial note under each quote is this site’s extrapolation to a tabletop machine, not something the source said: an editor’s judgement, audited for overreach, never a citation.
-
A commercial-class 10 MeV PET-cyclotron power budget (CYC2016 design): 1.5 kW internal PIG ion source against 26 kW magnet coil and 14 kW RF consumption; simulated beam after the third accelerating gap ~197 uA at 190 keV from a 40 kV gap voltage.
P_ion_source ~ 1.5 kW (commercial); ~4% of machine wall powerSource quote & editorial note
Coil Consumption Power [kW] 26 ... RF Consumption Power [kW] 14 ... Ion Source Power [kW] 1.5 (Table 1) ... Cavity loss power was calculated 12.7 kW to generate an electric field with 40 kV gap voltage ... Beam energy and current was checked 190 keV, 197 uA after third accelerating gap
Editorial note, tabletop extrapolation: Context datum, not a scaling law: the reference machine's ~0.1-0.2 kW source budget is a deliberate derating of this class of design, but beam current does not scale with source power - capture, acceptance and extraction losses dominate - so estimate current from measured source output and capture efficiency.
-
In the cited independently-tuned two-dee system, unneutralized dee-to-dee capacitance coupled the two tuning servos so strongly that stability was, in the source's words, insuperable - power flows dee-to-dee through the high-Q resonator, and shielding skirts and time-constant tweaks did not fix it; transmission-line neutralization between the stems did.
dee-dee neutralizing line load condition Vn = Va*w*CDD*Zo*sin(beta*l)Source quote & editorial note
The most serious objection to the dee-to-dee capacitance is the coupling between servo systems which it provides. The problem of servo stability becomes insuperable.
Smith, A Three-Phase Radiofrequency System for Cloverleaf Cyclotrons — UCRL-3153 (1955) — p. 15
Editorial note, tabletop extrapolation: A single-dee next machine dodges the coupled-servo problem entirely - the design lesson. Any two-dee or dee-plus-tuned-dummy variant with separate tuners should measure the coupling matrix and analyze loop stability first: neutralizing lines are one narrowband remedy, and common tuning, coordinated (MIMO) control or reduced bandwidth are others.
-
Verify neutralization by exciting one dee stem at a time and measuring the voltage induced on the others; the 20-inch achieved coupling coefficients below 3%. The adjustment was done with the machine vented to air, because at low pressure the low-level test drive multipactors.
N_ij = e_j/e_i; 20-inch achieved < 3% (their result, not a universal pass number)Source quote & editorial note
It was necessary to do this while the machine was down to air, in order to avoid multipactoring. ... The coefficients for the 20-inch cyclotron were below 3%. In order to adjust the coupling loops of the neutralizing lines the dee stems were excited one at a time
Smith, A Three-Phase Radiofrequency System for Cloverleaf Cyclotrons — UCRL-3153 (1955) — p. 18
Editorial note, tabletop extrapolation: Two transferable habits: quantify RF isolation as a measured coefficient (set the pass number from your own loop-stability analysis), and remember low-level RF in vacuum can sit in a multipactor window - the reference machine has seen multipactor-like loading. Venting for the test sidesteps multipactor but is not blanket safety: corona, heating and hazardous RF voltage remain, so keep monitoring and interlocks in place.
-
Derive extraction-element timing from turn separation: with ~0.1 in radius gain per rf cycle and deflector bars 1 in apart, ions cross the bar aperture in ~10 rf cycles, so the pulse must fire within +/-5 rf cycles.
aperture transit ~ (bar spacing)/(radius gain per turn) rf cycles; at ~10 Mc, 10 cycles ~ 1 us, so the firing window is ~+/-0.5 us; rise time is budgeted separately from the allowable field transient while ions occupy the deflectorSource quote & editorial note
the increase in radius of the burst of ions per rf cycle is approximately 0.1 inches and the deflector bars are spaced one inch apart, the pulse must occur within +/- 5 rf cycles.
Kerns et al., High Voltage Pulser for 184-inch Cyclotron Electric Deflector — UCRL-95 (1948) — p. 8
Editorial note, tabletop extrapolation: Synchrocyclotron-specific hardware (a CW deflector needs no pulse), but the requirements chain - turn separation sets element aperture sets timing budget - is the template for sizing ANY extraction element, including a next machine's septum entrance.
-
The 184-inch pulsed electric deflector needed ~75,000 V/cm - about 190 kV across bars spaced one inch - to shift the center of rotation of its full-energy beam into the lowered-field magnetic channel.
E ~ 75 kV/cm; V = E*d ~ 190 kV across 1 in (2.54 cm)Source quote & editorial note
The electric field between the deflector bars necessary to shift the particle center of rotation enough to allow it to pass through the magnetic channel is about 75,000 volts per centimeter.
Kerns et al., High Voltage Pulser for 184-inch Cyclotron Electric Deflector — UCRL-95 (1948) — p. 8
Editorial note, tabletop extrapolation: Scale perspective without a false law: the required field follows from the integrated kick theta ~ q*INT(E dl)/(p*v) - beam rigidity, electrode length, needed displacement and septum clearance all enter, which is why UCRL-10654 quotes 150-200 kV/cm at 50 MeV while this machine used 75 kV/cm at 200 MeV. Compute a sub-MeV machine's requirement from its own kick integral; it will come out modest, but earn the number.
-
Fast-pulse transformer lore: keep leakage inductance down by paralleling coils and minimizing core cross-section (1.5 x 1.5 in Hipersil, 2-mil laminations); at 6 kV/turn, interlaminar insulation arcs - splitting the core into two segments halves per-segment voltage and halved those losses; ~90% of input power ends up as core heat (500 W, cores reach 200-300 C), demanding non-shorting water-cooled jackets; vacuum-fill the lucite case with de-aerated oil to kill corona. Result survived 300 kV = 3x rated output.
2:17 turns, 6 kV/turn, two coils paralleled halve leakage L; core split halves interlaminar V; tested 300 kV vs 100 kV service (verified on page image)Source quote & editorial note
Approximately ninety percent of the total power input to the system is eventually dissipated in the transformer cores as heat. At rated operating levels this loss is approximately 500 watts.
Kerns et al., High Voltage Pulser for 184-inch Cyclotron Electric Deflector — UCRL-95 (1948) — p. 12
Editorial note, tabletop extrapolation: Transferable craft here is the failure mode (interlaminar voltage at high volts-per-turn) and the de-aerated-oil practice - which reduces bubbles and partial discharge, not corona from bad geometry. The 300-kV survival was that transformer's result, not a portable 3x proof-test rule: overvoltage testing at these levels is itself hazardous and can leave latent damage, so test to an applicable HV standard's waveform, duration and partial-discharge limits, remotely, with discharge provisions - not to a generic multiple.
-
DC resonance charging through the pulse capacitors' voltage reversal gave a step-up beyond the textbook maximum: 11,000 V at the thyratron plates from a 2,750-V supply - four to one against the usual two to one - because each shot leaves the capacitors reversed. The report adds that the ratio depends on losses in the entire system, with step-up ratios as high as ten to one observed.
cited circuit: V_plate/V_supply = 4:1 (vs 2:1 classical resonant charging), via post-pulse capacitor reversal; loss-dependent, up to 10:1 observedSource quote & editorial note
a plate voltage of 11,000 volts on the thyratrons can be maintained with a power supply voltage of 2,750 volts, a step up ratio of four to one ... Step up ratios as high as ten to one have been observed.
Kerns et al., High Voltage Pulser for 184-inch Cyclotron Electric Deflector — UCRL-95 (1948) — p. PDF p.15 (printed p.-12-)
Editorial note, tabletop extrapolation: Pulsed-modulator craft, not CW-deflector material; file under 'if a next machine ever needs a kicker'. The ratio is topology- and loss-dependent - simulate or measure the actual waveform before sizing a supply on it, and rate every capacitor, switch and insulator for the real reversal stresses.
-
Design a regenerator by the source's seven-step procedure - from nonlinear equations of motion on the measured field through amplitude-dependent tunes to the required momentum kick and its field perturbation (the step contents and gain expressions summarized here are the report's derivation - re-read queued for the equations and variable definitions).
a = -sin(wr*th1)/sin(wr*(th2-th1)), wr = wr(r>R); delta(p') = -p0''*sin(wr*th2)/ sin(wr*(th2-th1))Source quote & editorial note
The determination of the required perturbation for extracting the beam of a synchrocyclotron is made in seven steps.
Stubbins, Extraction of Synchrocyclotron Beams Near the Maximum Energy — UCRL-3476 (1956) — p. 6
Editorial note, tabletop extrapolation: The workflow (measured field -> amplitude-dependent tunes -> impulse-matrix tracking -> element strength) is exactly the CYCLOPS-lite pipeline planned for a next machine; the peeler-regenerator field shapes themselves are synchrocyclotron machinery and need not transfer. Treat the sine-ratio gain coefficient as branch- and model-specific once the re-read pins its definitions - it is singular near its denominator zeros, so no monotone smaller-interval-more-gain rule survives unqualified.
-
Express regenerator strength as integrated field-times-angle - with B0 in gauss and angle in radians the perturbation integral INT(dB dtheta) reads in gauss-radians, the source's unit convention (its formula, worked table and B0 are the report's data - re-read queued).
integrated perturbation = INT(dB dtheta) [G-rad]; conversion: 1 kG-deg = 17.45 G-radSource quote & editorial note
When B0 is in gauss, B-theta is in gauss-radians.
Stubbins, Extraction of Synchrocyclotron Beams Near the Maximum Energy — UCRL-3476 (1956) — p. 17
Editorial note, tabletop extrapolation: The gauss-radian bookkeeping is a handy unit for ANY azimuthally localized field bump (harmonic coils, shims, channel compensation) on a next machine. What fraction of B0 an effective bump needs is geometry-dependent - the same integrated strength over 20 or 60 degrees is a 3x different local field - so compute the integral for the actual bump, without a stock percent anchor.
-
In the cited regenerator calculation, the disturbance to axial motion at 1-in axial amplitude was about twice the corresponding radial disturbance from the same field perturbation - large-axial-amplitude particles were the vulnerable population in that analysis.
delta(z') ~ 2x radial disturbance at 1-in axial amplitude; d(axial)/dr of Br from curl B = 0 -> Br = (dBz/dr)*zSource quote & editorial note
For a 1-in. axial amplitude this disturbance is about twice as strong as that occurring in the radial motion from the same field perturbation.
Stubbins, Extraction of Synchrocyclotron Beams Near the Maximum Energy — UCRL-3476 (1956) — p. 18
Editorial note, tabletop extrapolation: The transferable warning: any radial-field-gradient extraction element has an off-midplane Br ~ z*dBz/dr whose vertical effect can focus or defocus depending on gradient sign and trajectory - include the deflector fringe and any field bump in the next machine's 3-D tracking rather than assuming the sign or which particles go first.
-
Include the magnetic channel's own field in the orbit calculation: the computation lets one modify the regenerator field to account for the channel effect, so that maximum-effort corrective shimming of the channel is not required.
treat channel fringe as a fourth orbit region; adjust regenerator to compensateSource quote & editorial note
The computation enables one to modify the regenerator field to account for the channel effect, and, thus, the maximum effort of corrective shimming for the channel is not required.
Stubbins, Extraction of Synchrocyclotron Beams Near the Maximum Energy — UCRL-3476 (1956) — p. 20
Editorial note, tabletop extrapolation: Direct analog for a next machine: the septum and exit-channel iron (or deflector entrance fringe) perturbs the last internal turns - model that perturbation in the tracker and consider compensating upstream (harmonic coil, shim, or the bump program) as ONE option alongside local shielding or shimming of the channel itself; co-optimize rather than nulling one element in isolation.
-
The cited swept-RF system split its deflector trigger into two stages - a frequency-sensitive circuit that GATES and a phase-sensitive circuit that TRIGGERS - so a coarse condition opens the window and the RF itself supplies the firing phase.
Source quote & editorial note
The frequency-sensitive circuit "gates" the phase-sensitive circuit, and the phase-sensitive circuit triggers the deflector.
Foss et al., Cyclotron Component Design Technical Reports — TID-454 (1952) — p. 168
Editorial note, tabletop extrapolation: The architecture for a timed kick against an extraction gap when frequency and phase are not derived coherently from one reference: coarse condition (frequency, turn count, integrated field) gates, RF phase triggers. A modern phase-coherent synthesizer can supply both from one reference - then a single measurement suffices.
-
To detect when a swept RF reaches a chosen frequency, do not build a stable tunable RF filter - heterodyne the RF against a crystal local oscillator and detect the transient through a low-frequency band-pass filter, making the trigger point adjustable via the low-frequency side and crystal switching (the report used a 1-1.25 Mc filter with switched crystals to cover 19-21.5 Mc).
trigger when |f_dee - f_LO| = f_IF - BOTH sign branches respond, so the unwanted (image) crossing must be gated out or rejected; report's implementation: f_IF ~ 1-1.25 Mc, crystals switched across the bandSource quote & editorial note
Block 4 is a local oscillator with a frequency about 1 megacycle below the desired deflection frequency P. When the dee-oscillator signal sweeps through point P, the 1-megacycle filter passes an a-c transient.
Foss et al., Cyclotron Component Design Technical Reports — TID-454 (1952) — p. 168-169
Editorial note, tabletop extrapolation: Classic measurement doctrine - move the precision problem to a low frequency where stability is cheap. A modern mix-down marker inherits the crystal's stability only for the LO term: the IF filter's center drift, bandwidth and threshold timing all enter the marker's error budget, so build that budget rather than expecting crystal accuracy from junk-box filters.
-
A swept signal peaks in a band-pass filter LATER than the moment it crosses the filter's center frequency - so trigger timing calibrated at one sweep rate silently moves when the sweep rate changes; the cited system provided a per-repetition-rate bias adjustment (reported detail, scan re-read queued).
peak delay depends on filter bandwidth and instantaneous df/dt of the sweep - characterize against the actual chirp rate and filter responseSource quote & editorial note
the time at which the transient is at a peak is somewhat later than the time at which the dee-oscillator frequency is in the center of the filter pass band.
Foss et al., Cyclotron Component Design Technical Reports — TID-454 (1952) — p. 171-172
Editorial note, tabletop extrapolation: Matters wherever a resonant pickup watches a changing frequency - a synchrotron RF ramp or an FM-tuned marker on a cyclotron: if the ramp rate changes, re-characterize the timing rather than assuming the old calibration.
-
Match design effort to field class: the source says power and maximum current density 'become important factors and more complicated designs are useful' for fields of 1e5 gauss and above - elaborate minimum-power current distributions (j ~ sin(theta)/r^2 kernels) are particularly motivated in that regime. [Corrected 2026-08-23: earlier text inverted this into a claim that near 1 kG air-core coil power is small and optimisation 'seldom worthwhile', which the quote does not say.]
ideal minimum-power distribution: j = k*sin(theta)/r^2 inside boundary r^2 = k'*sin(theta) (Eqs. 3-4) - relevant only in the high-field regimeSource quote & editorial note
For fields of 10^5 gauss and above, the situation is quite different; the power and maximum current density become important factors and more complicated designs are useful.
Foss et al., Cyclotron Component Design Technical Reports — TID-454 (1952) — p. 122
Editorial note, tabletop extrapolation: Locates amateur work far below the exotic regime: for sub-kG correction coils, steering windings and test solenoids a simple winding is usually adequate - but 'usually' is earned by computing NI, resistance, I^2R heating, temperature rise and current density for every coil, since a small high-duty coil can be power-limited at any field. [Note revised 2026-08-23: earlier note said 'sophistication buys nothing'.]
-
Derive an FM (frequency-vs-time) program from the constant-ion-phase condition and measured oscillator data rather than seeking an exact law - in the cited synchrocyclotron design, the required capacity-vs-time variation was 'not very critical'; and cycle dead time taxes average beam current directly, so minimize the return-to-start time.
df/dt from constant-phase relation integrated numerically against measured f-vs-C of the model oscillator (Eqs. 1-3); t_return <= t_accel for best duty cycleSource quote & editorial note
The operation of the oscillator determines the variation of capacity with time which will keep the ion phase constant. This variation is not very critical.
Foss et al., Cyclotron Component Design Technical Reports — TID-454 (1952) — p. 152
Editorial note, tabletop extrapolation: For a small synchrotron's RF ramp the pattern maps with its own tolerances: f(t) and the allowable phase/frequency error come from the magnetic ramp, synchronous orbit and RF-bucket acceptance - the cited looseness belongs to that FM oscillator, not to synchrotron ramps in general; the duty-factor lesson (reset time is pure tax, minimize it within hardware limits) transfers as stated.
-
Different structural metals leave different residual-nuclide inventories under the same irradiation - the report's survey at its energy: aluminum yielded no long-lived activities they detected, iron essentially pure 300-day Mn54, stainless adds 27-day Cr51 and 71-day Co58 from its Cr and Ni, copper gives 12.8-hr Cu64 and Co58.
long-lived residuals at 730 MeV: Al -> none; Fe -> Mn54; SS(10%Ni,20%Cr) -> Mn54 + Cr51 + Co58; Cu -> Co58; yield ratio Cu64/Na24 ~ 50/1 (factor ~2)Source quote & editorial note
Aluminum yields no long-lived activities, while Co58 is produced from copper ... Stainless steel produces two long-lived isotopes, Co58 and Mn54, while only Mn54 is induced in iron.
Boom, Toth & Zucker, Residual Radiation of the LRL 184-inch Cyclotron — ORNL-3158 (1961) — p. 18
Editorial note, tabletop extrapolation: ENERGY SCOPE: nuclide-by-material bookkeeping from 730-MeV spallation; these channels are closed at sub-MeV proton energy. The durable pattern is inventory-follows-alloy-content (Ni -> Co58, Cr -> Cr51) - worth knowing when reading other labs' surveys. The report's list is what their instruments saw, not an exhaustive table: modern data adds Be-7 and Na-22 from high-energy aluminum, so treat any such inventory as survey-specific.
Cited in: Shielding a Small Cyclotron
-
Quarantine the operations that activate the machine hardest (at Crocker, deuteron runs) into scheduled windows — end of week, mandated 25-30 min cooling, longer for prolonged runs — so the activation decays over the idle period instead of irradiating the next shift.
Source quote & editorial note
They are run only on Friday evenings and Saturday, and - if demands are high - on Sundays. For these bombardments a longer cooling time is required. A nominal time of 25 to 30 minutes is set
McWalters et al., Radiation Exposures of Personnel at the 60-inch Cyclotron — UCRL-8276 (1958) — p. 10
Editorial note, tabletop extrapolation: ENERGY SCOPE: Crocker's figures are (d,n) activation at tens of uA and ~20 MeV. Sub-MeV deuterons still make neutrons - D(d,n)3He and 9Be(d,n) are exothermic - so the analogue exists the day deuterium enters the machine (dg-121, dg-545). The scheduling pattern transfers either way: batch the nastiest operations (HV conditioning, any future deuteron or neutron work) into planned windows with a defined stand-down, rather than interleaving them with routine bench time.
-
Hold the magnet gap to a relative tolerance of order a few parts in 1e4 of the gap when the field must satisfy an isochronism/focusing spec across the pole.
gap tolerance +/-0.004 in. on 8 in. gap = 5e-4 relativeSource quote & editorial note
Gap tolerance +/- 0.004 in.
Oak Ridge National Laboratory, A Proposal for the Mc² Isochronous Cyclotron — ORNL-3540 (1963) — p. 137
Editorial note, tabletop extrapolation: 810-MeV provenance: the spec belongs to a 53-ft isochronous magnet, and even the RELATIVE number (5e-4 of the 8-in gap) is that machine's, not a scale-free constant - the transferable content is the framing: derive the next machine's gap tolerance from its allowed field error via a magnetostatic model or measured dB/dg, then stack machining, assembly and thermal terms; sanity-check the result against the 5 G / 5-deg-phase budget.
-
Set seal policy by radiation dose and replaceability, as the cited design did: elastomer seals only where the predicted 10-yr dose was below 1e8 rad AND the seal is easily changed; elsewhere their choice fell to metal seals, with the interspace of large double seals pumped (design practice reported around the quoted criterion).
elastomer allowed where 10-yr dose < 1e8 rad AND easily changed (their criterion)Source quote & editorial note
elastomer seals are used only where the predicted 10-yr radiation dose is less than 10^8 rad and where the seal is easily changed.
Oak Ridge National Laboratory, A Proposal for the Mc² Isochronous Cyclotron — ORNL-3540 (1963) — p. 157
Editorial note, tabletop extrapolation: The decision MATRIX (dose x replaceability -> seal type) is scale-free and becomes relevant the moment a next machine makes real beam current near median-plane seals; the 1e8-rad line is that design's criterion for its compounds and lifetime - qualify the actual elastomer against predicted dose, temperature, compression set and access before borrowing the number.
-
A Buechner-Bainbridge 90-degree broad-range spectrograph (uniform field; source and focus each one characteristic radius outside the field boundary) covers a wide energy band in one exposure; the practical top of the band is set by chamber size - beyond ~1.3 E0 the exit chamber grows unreasonable - and single-focusing solid angle punishes the high end (detailed range/resolution figures report-attributed - scan re-read queued).
energy scales as (B*R)^2 for similar optics; the cited instrument: R = 50 cm at 14 kG for 33 MeV protonsSource quote & editorial note
an extension of the energy range much beyond 1.3 E0 requires an unreasonably large vacuum chamber at the exit of the magnet.
Editorial note, tabletop extrapolation: SCALE-HONEST only when the scaling is done: the geometry fixes E/E0 ratios, but reaching a given E takes B*R. For ~170 keV protons, B-rho ~ 0.06 T-m, so an R ~ 5-10 cm bench version needs roughly 0.6-1.2 T - iron-pole territory, not a few hundred gauss. Still compelling as a teaching-lab focal-plane instrument; copy the optics and size the field honestly.
-
Long-lived photoproduced isotopes in shielding cannot be waited out: 'if large quantities of this isotope build up, it will be necessary to physically remove the activated shielding, so plans for this contingency should be made in the design of the walls' - the isotope's identity and production threshold are the chapter's context (Na-22-class, multi-MeV photons; cite current nuclear data when used).
above-threshold gamma flux + years of operation -> Na-22 inventory -> removable-wall contingency in designSource quote & editorial note
If large quantities of this isotope build up, it will be necessary to physically remove the activated shielding, so plans for this contingency should be made in the design of the walls.
Editorial note, tabletop extrapolation: Scope closed for the machine's photon energies - photoproduction needs multi-MeV photons far beyond any dee. The design principle transfers as prudence rather than prohibition: prefer enclosure designs that could be dismantled selectively (block walls, the sacrificial inner course of dg-1050) over monoliths - cheap to choose now, expensive to regret.
-
Track radiological teardown work against a plan and a target: Nevis's ten-week, 2000-ton teardown held all workers below 100 mrad/week averages, most below 25 - the quoted record; the cooling delays and strip-down sequencing are the report's account of how (scan re-read queued).
cooling delay + staged strip-down + weekly per-worker dose trackingSource quote & editorial note
with all workers averaging below 100 mrad/week, and most below 25 mrad/week for the 10 weeks of this activity
Editorial note, tabletop extrapolation: Dose scale is irrelevant to a 150 keV proton machine, but the pattern — cooling time, planned sequence, measured-not-assumed exposure — is the template for any future activated-hardware work and for the plan's licensing narrative.
-
Adiabatic RF manipulation at the Nevis parking point: with the beam parked where df/dt ~ 0, a slow linear reduction of RF amplitude spreads the phase angle near-adiabatically - DEbunching the beam: phase width grows while energy-oscillation amplitude shrinks (the duration and the ~3x figure are the report's numbers - re-read queued).
slow linear V_RF turn-off at df/dt ~ 0 "parking frequency" -> ~3x reduction in phase-oscillation dESource quote & editorial note
a slow linear reduction (turn off) of the RF amplitude there will result in a near adiabatic spreading out of the phase angle
Editorial note, tabletop extrapolation: Swept-frequency machinery, not fixed-frequency CW territory - the design space it illustrates (adiabatic capture, slow parameter ramps judged against the phase-oscillation period, not a fixed microsecond count) belongs to any future synchro- or synchrotron-class RF program.
-
The Nevis synchrocyclotron modification report's shielding datum (Rainwater et al., 1971): about 6 inches of iron 'or the equivalent' per factor-of-two attenuation for the forward high-energy neutron cone (>100 MeV, charge-exchange), while at 90 degrees or more the neutrons are mainly below 100 MeV and the same 6 inches buys closer to a factor of 10 - shield thickness is budgeted per direction. [2026-09-06 erratum, scan re-read: previously attributed to 'Moyer', a name that appears nowhere in the 81-page report; and the report says 'or the equivalent', never naming concrete.]
~6 in Fe (or equivalent) per x2, forward cone >100 MeV; same 6 in ~ x10 at >=90 deg (mainly <100 MeV) - per-direction budgetingSource quote & editorial note
Forward cone > 100 MeV neutrons ... require ~ 6 in. Fe (or the equivalent) for each factor of 2 attenuation. At 90 deg or more ... closer to a factor 10 attenuation.
Editorial note, tabletop extrapolation: Pure high-energy datum — no tabletop relevance except as a worked example of directional shielding budgets, but it anchors the energy scaling.
-
Design components in activated regions for remote replacement: Nevis designed all dee SUPPORT INSULATORS to be removable and replaceable by remote handling tools - the quote; the wider behind-shields work practice is the report's context.
activated-region components = pin-located, tool-accessible, removable without entering the chamberSource quote & editorial note
all support insulators have been designed so that they can be removed and replaced by remote handling tools.
Editorial note, tabletop extrapolation: At tabletop energies the driver is vacuum hygiene and downtime rather than dose, but the same design habit — most-likely-to-fail parts (insulators, filaments, septa) replaceable without major disassembly — is what the reference machine's filament-change experience already argues for.
-
Utilization benchmark from a mature research cyclotron: 1,680 hours of operation in one quarter - about 18 hours per day - with the reported unscheduled losses being three days to a cold-trap refrigerator failure and one day to low diffusion-pump oil.
1680 h / 92 days ~ 18.3 h/day operatingSource quote & editorial note
the cyclotron was in operation 1680 hours, or about 18 hours per day. Three days were lost due to failure of the refrigerator for the cold-trap above the diffusion pumps. Another day was lost because of inadequate oil levels in the diffusion pumps.
Harvard University Cyclotron Laboratory, Quarterly Progress Report, 1 June – 31 August 1964 — p. 1
Editorial note, tabletop extrapolation: An upper anchor for what sustained cyclotron utilization can look like - useful against the NRL 60% figure only as a rough contrast, since the two reports account time differently (Harvard reports operating hours and lost days; NRL a full scheduled-time ledger). Note both of Harvard's losses were vacuum-auxiliary failures - cold-trap refrigeration and pump oil - not accelerator physics.
-
Rotate the target when average power exceeds what a static foil stands: Folger (GSI) ran 9 sector ("banana") targets covering 59.6% of a 97.4 cm circumference at 15.5 cm radius, spun at 666 rpm phase-locked to the beam macropulse (20 degrees per 5 ms pulse) so successive pulses hit different targets; ~1e17 particles were integrated without significant radiation damage.
wheel synchronization; 666 rpm = 20 deg per 5 ms macropulse (25% duty, 5 ms in 20 ms)Source quote & editorial note
The wheel thus had to be rotated at a velocity of 666 rpm (equal to 20 deg in 5 ms or during one macropulse).
Editorial note, tabletop extrapolation: The design move transfers whole: spread the duty over many target areas, and if the beam is pulsed, phase-lock the rotation so no spot sees consecutive pulses - it scales to a bench wheel behind any external beamline. The 1e17-particle survival belongs to GSI's target, beam and cooling; a tabletop wheel's achievable dose comes from its own thermal, stress and deposited-dose arithmetic.
-
Actinide-alloy targets for in-beam work were arc-melted into cubic non-paramagnetic host intermetallics and mounted on thick brass holders serving as heat sinks; the Stony Brook fission-isomer team chose the UIr2 host to defeat paramagnetic relaxation, with radiation damage from recoil implantation the other standing obstacle (paraphrase only — journal reprint, no quotation).
Source quote & editorial note
NO QUOTE — paper IV-1 is reprinted from Nucl. Instr. and Meth. 206 (1983) 361-366 with North-Holland permission; finding paraphrased, cite the journal article.
Editorial note, tabletop extrapolation: NOT a tabletop construction example. Actinide targets - arc-melting the alloy, mounting it, putting it in a beam - are licensed radiological-laboratory work: contamination control, shielding, dosimetry, fission-product and activation handling, and radioactive-waste management, on top of the licence itself. Keep this rule as a literature example of matching host-material physics (crystal symmetry, conductivity, heat sinking) to what a measurement needs, and as the rights-boundary marker for this volume; the transferable idea is the matching, never the material. Cite the journal article (Nucl. Instr. and Meth. 206 (1983) 361-366) for the paraphrased finding. [Corrected 2026-08-23: earlier note called this merely "marginal technically" for a proton machine, which understated the hazards that actually decide it.]