Construction Classics

X-ray output below 100 kV, measured

Any evacuated gap with kilovolts across it can make X-rays, and the question for a builder is how many. These four documents are the measured layer of the answer for the voltages a small accelerator’s RF system reaches: a 1970 survey of what classroom gas-discharge tubes emit between 15 and 85 kV, a 1978 compendium of tube outputs and spectra at 20 to 60 kVp, the 1961 handbook that set out the barrier-sizing arithmetic, and the 1972 standard for analytical X-ray installations. All four are United States government publications.

Why these documents can be hosted

BRH/DEP 70-1 and FDA 79-8071 are reports of the Bureau of Radiological Health, a unit of the US Public Health Service and later of the Food and Drug Administration, written by its own staff; works of the United States Government carry no copyright. Neither scan carries a copyright notice. The FDA copy is the National Technical Information Service reproduction, which says so on its cover.

NBS Handbook 76 is NCRP Report No. 26 issued by the National Bureau of Standards as a government handbook, printed by the Government Printing Office and sold by the Superintendent of Documents. The 64-page scan carries no copyright notice, and a work published in the United States in 1961 without one entered the public domain on publication. The committee that wrote it was not itself a government body, which is why the basis stated here is the absence of notice on a government publication rather than government authorship. NBS Handbook 111 is ANSI N43.2-1971 issued the same way in June 1972; its verso carries the ANSI consensus statement and a Library of Congress card number and no copyright notice, and the same no-notice basis applies. Both were read page by page before this determination was made.

Two companions from the same mining wave are cited but not hosted. NBS Handbook 114 (ANSI N543-1974) reproduces a lead-sheet table in its Appendix C credited Permission: Lead Industries Association, Inc. and From NCRP Report No. 34, which makes it a public-domain wrapper around a licensed inclusion, the class this collection first met in AD-755510; its rules link to the NIST copy. The 1970 Radiological Health Handbook carries third-party charts throughout and stays cite-only. Determinations made October 2026.

What a classroom discharge tube emits (BRH/DEP 70-1)

The Bureau’s X-ray Exposure Control Laboratory bought the cold-cathode tubes sold to schools, ran them from the supplies schools used, and measured. The operating points were approximate by the report’s own account: about 25 kVp and 0.01–0.09 mA on a hand-held Tesla coil, 50–65 kVp and 0.03–0.18 mA on a 3-inch induction coil, 73–85 kVp and 0.15–0.20 mA on a 6-inch coil, with kilovoltage read by a spark gap in parallel with the running tube so that the loading of the supply was included (dg-1925, dg-1936). At 1 ft from the anode the X-ray tubes gave 6.7–15 R/h on the 3-inch coil and 42–217 R/h on the 6-inch coil (dg-1926). Along a radial from the best-emitting tube the rate fell from 240–281 R/h at 1 ft to 0.53–0.77 R/h at 16 ft, which the authors judged in general agreement with the inverse-square law (dg-1927); the hemisphere behind the anode read lower by a factor of up to several hundred (dg-1928).

The more useful half of the report, for anyone whose apparatus is not meant to be an X-ray tube, is section 2.3.2: the tubes not intended to produce X-rays. A heat-effect tube ran at about 15 kV with nothing detectable on correct polarity; reversed, it ran at about 55 kV and emitted, and kept emitting when reconnected forward, which the authors attribute with an apparently to internal getter action having hardened the vacuum (dg-1930). A magnetic-effect tube read about 1 R/h at 1 ft at 50 kV before its voltage collapsed (dg-1933). The authors’ reading of their own results is that excessive gas pressure was the predominant factor keeping most demonstration tubes quiet, and that pumping them to high vacuum could be expected to make them emitters (dg-1931). Their conclusion is the general one: X-ray production will exist wherever there is an electron source, a target, sufficiently high voltage and gas pressure in the proper range (dg-1932). A pumped chamber with an RF gap satisfies every clause.

Appendix B is measurement craft. Induction and Tesla coils radiate enough RF to upset a survey meter; the laboratory sewed a bag of silver-metalized nylon marquisette around a CS-40A and found the shield changed its response by at most about 5 percent (dg-1935). The 1970 recommendations on pp. 39–40 (shield to 10 mR/h at 1 ft, keep students 8 ft from an unshielded incidental emitter) are the Bureau’s, for its era, and are reported as such.

Tube output and spectra at 20–60 kVp (FDA 79-8071)

A mammography handbook is here because it is the one public-domain source that tabulates what a tungsten-anode tube actually puts out at the kilovoltages a small accelerator’s RF reaches. Read from its Fig. 6, a tungsten tube with about 0.64 mm Al inherent filtration and no added filter gave roughly 1.6 mR/mAs at 1 m at 30 kVp, about 3.3 at 40, 5.7 at 50 and 8 at 60 kVp (dg-1937). At 30 kVp the four tubes measured ranged from 2.84 to 9.26 mR/mAs at 77 cm with half-value layers of 0.41–0.60 mm Al (dg-1938). Sections 4.5.2 to 4.5.4 (pp. 22 onward) are the numerical spectra, primary and exit, tube by tube.

Two methodological points travel further than the numbers. Exposure in mR overweights soft photons, so comparing tubes by mR/mAs is unfair to a harder spectrum, and the authors supply photon-fluence conversions for that reason (dg-1939). And at the low tube currents used for spectra, leakage current was a significant part of the measured current, enough that one system’s mR/mAs rose threefold between 0.4 and 4.0 mA; the output data were therefore taken at high current (dg-1940). The authors’ prescription for making a spectrum reproducible is kVp, filtration and a measured HVL (dg-1941).

The barrier arithmetic (NBS Handbook 76)

Handbook 76 is where the shielding method the later handbooks assume was written down. Its premise is a measured regularity: at a given kilovoltage and with minimum added filtration, any X-ray machine’s exposure rate is nearly constant in roentgens per milliampere-minute at 1 m, and the barrier thickness is essentially independent of the machine’s own filtration because that filtration is small next to the barrier (dg-1946). The design outputs it tabulates run from 0.05 R/mA-min at 1 m at 50 kV to 280 at 2,000 kV (dg-1947).

Appendix C (pp. 42–51) is the procedure. A primary barrier is sized by K = P d² / (W U T), with P the weekly permissible exposure, d the distance in metres, W the weekly workload in mA-min, U the use factor and T the occupancy factor; the thickness is then read where the transmission curve for that kilovoltage reaches K. The handbook’s worked example at 100 kVp and 10 m comes out at 1½ in of concrete or 0.4 mm of lead (dg-1949). Other wall materials are converted to concrete by density alone, which the handbook notes errs toward extra protection for materials of higher atomic number (dg-1950). Secondary barriers count leakage and scatter separately, with 90-degree scatter taken as about 0.1 percent of the incident exposure, and the two thicknesses are combined by their HVL difference (dg-1951). The lead transmission curves of Fig. 3 are the data behind all of it, with the handbook’s own caution that the 50 and 70 kVp curves were interpolated and extrapolated from Braestrup’s 1944 measurements (dg-1952).

What is superseded here is the regulatory layer. The dose limits, the shall and should rules and the occupancy factors are 1961 recommendations, revised many times since; the handbook’s own preface says it is not a legal document. Current law for a given jurisdiction is on the Legal pages. The transmission curves, the barrier equation and the leakage-versus-scatter logic are what to read it for.

Leaks, rectifiers and small beams (NBS Handbook 111)

The 1972 standard governs diffraction and fluorescence sets, which are the laboratory apparatus closest in kind to a small accelerator: a sealed high-voltage source in a cabinet that people work next to. Its numbers are the standard’s own and should be read as such: a primary-beam trap 5 cm from the focal spot equivalent to at least 1.5–1.7 mm of lead at 50 kVp or 7.7–7.9 mm at 100 kVp (dg-1958), enclosed-system walls limited to 0.25 mrem/h at 5 cm (dg-1959).

Three of its observations are the kind that get learned the hard way. Gassy rectifiers in the high-voltage supply can radiate at an effective potential twice that applied to the tube, producing very penetrating radiation at any time, and the standard’s only countermeasure is to shield the rectifier assembly and check it for leakage at least twice a year (dg-1960). A survey instrument calibrated in a uniform field reads low on a beam smaller than its sensitive area, roughly in the ratio of the two areas, and beams as small as 0.01 cm² can need correction factors of 6,000 or more for a 3½-inch chamber; the standard recommends averaging over 1 cm² near ports and using film to locate leaks (dg-1962). And lead parts should be inspected after every reassembly for damage or distortion that could open a leak, and mounted so they cannot cold-flow under their own weight; the standard reports that most severe injuries occur during non-routine operations such as repair and alignment (dg-1963).

Why these four belong together

They answer one question in sequence. The BRH survey shows that apparatus never meant to emit will emit once its vacuum is good enough and its voltage high enough. The FDA handbook says how much a tube of known kVp and filtration emits, and how to characterise the beam so the number can be matched. Handbook 76 turns an output and a workload into a wall. Handbook 111 says where the number will be wrong in practice: at the rectifier, at a pinhole too small for the meter, and after the last repair. None of the four was written about accelerators, and the Design Guide’s notes on these rules keep the numbers scoped to the tubes that produced them; the method, from spark-gap voltage reading to barrier equation, is what transfers.

Design rules extracted from these documents

31 Design Guide rules come from these four documents. They sit in the safety, shielding and detectors domains, with the FDA handbook also feeding materials and the diffraction standard feeding fabrication. Representative examples:

• dg-1932 — X-rays are produced wherever an electron source, a target, enough voltage and the right gas pressure coincide (the BRH conclusion)
• dg-1927 — measured falloff from a 6-inch-coil tube: 240–281 R/h at 1 ft to 0.53–0.77 R/h at 16 ft
• dg-1936 — measure the operating voltage with the spark gap in parallel with the running device
• dg-1935 — an unshielded survey meter can be upset by the RF of the high-voltage source
• dg-1941 — characterise a beam by kVp, filtration and measured HVL
• dg-1949 — size a primary barrier by K = P d² / (W U T) and read the thickness off the transmission curve
• dg-1951 — count leakage and scatter separately for a secondary barrier, then combine by HVL difference
• dg-1962 — a survey instrument reads low on a beam smaller than its sensitive area

Every rule, by document:

• BRH/DEP 70-1 (12): dg-1925, dg-1926, dg-1927, dg-1928, dg-1929, dg-1930, dg-1931, dg-1932, dg-1933, dg-1934, dg-1935, dg-1936
• FDA 79-8071 (5): dg-1937, dg-1938, dg-1939, dg-1940, dg-1941
• NBS Handbook 76 (8): dg-1946, dg-1947, dg-1948, dg-1949, dg-1950, dg-1951, dg-1952, dg-1953
• NBS Handbook 111 (6): dg-1958, dg-1959, dg-1960, dg-1961, dg-1962, dg-1963