Construction Classics

R.F. System for Frequency Modulated Cyclotron

Why this document can be hosted

MDDC-1045 belongs to the Manhattan District Declassified series — AEC-published, declassified June 16, 1947, printed and priced (10 cents) at Oak Ridge — and such documents are US Government works, which carry no copyright. Pre-1978 distribution without notice is an independent public-domain basis. The title page states the copy “is reproduced as a project report” whose manuscript was submitted to the Review of Scientific Instruments: what is hosted here is the AEC report, not the 1947 journal article. Verified against the title page; determination made August 2026. The DTIC scan had no text layer; this copy was OCR’d in August 2026 and is searchable.

Reading guide

This eight-page report documents the RF system of the machine that proved synchrocyclotrons work: the Berkeley 37-inch, shimmed in 1946 so its field fell 13% from center to the 18-inch radius to imitate the planned 184-inch, then driven with a frequency-modulated oscillator. Deuterons reached 7.5 MeV on 15 kV of dee voltage — after which “the successful acceleration of deuterons to 7.5 Mev was sufficient basis for converting the 184-inch” (p. 3). Everything in it happens at a scale an amateur recognizes: a one-dee system, 5–6 kW of RF, and commercial triodes.

The introduction (pp. 1–2) records a decision worth knowing: electronic frequency modulation was tried first and produced 10% frequency swings — but only at 500 V of dee voltage per kilovolt of modulator, so mechanical modulation by rotary capacitor won. The rotary-capacitor arrangement (pp. 2–3) sits outside the magnetic field on its own vacuum system, coupled through a 14-foot transmission line operating in the half-wave mode; minimum capacitor capacity is matched to the 300 µµf dee capacity so the capacitor never sees more voltage than the dee. The oscillator section (pp. 3–5) compares three grounded-grid builds on the same circuit — four 304TL triodes in parallel, an 889, and the 3X2500A3 that the authors say would have made the others unnecessary — with the practical details: zircon replacing a heat-cracked porcelain dee insulator, a housing of copper sheet nailed to a wood frame, sponge-rubber-backed joints, a tuning vane giving ±6% of frequency trim, and the finding that all-steel rotary condenser surfaces were tolerable at 10 Mc but had to be copper-plated at 20 Mc.

The phase-correction analysis (pp. 7–8) is the report’s most transferable engineering: a grounded-grid oscillator picks up 20° or more of plate-to-filament phase shift, worked here as a vector diagram and corrected by choosing the filament bypass capacitor — the same 220 µµf capacitor doubles as the correction element. The closing pages cover amplitude modulation over the FM cycle and the discharge phenomena that appear when an RF cavity, a magnetic field, and 10⁻⁵ mm of gas share a volume. Page references are the report’s printed numbers; the hosted PDF carries two DTIC cover pages in front.

Design rules extracted from this document

13 Design Guide rules cite MDDC-1045 — a remarkable yield for twelve pages, nearly all of it oscillator and discharge practice at small-machine scale. Representative examples:

• dg-671 — dee-voltage-per-watt benchmark: 15 kV at 10 Mc from 6 kW input
• dg-676 — keep the grid-leak RC below a tenth of the resonant system’s time constant
• dg-679 — a series emission limiter in the oscillator HV lead
• dg-680 — the sub-500 V electron-oscillation discharge and its bias cure
• dg-682 — a tickler oscillator to drive the dee through the low-voltage region

Used by the dee RF coupling deep dive.