hasanukegoneoff.com logo ← Back to the live monitor

The Bhangmeter

A Documentary History of British Nuclear-Burst Yield Measurement, 1952-1969

Compiled from nineteen declassified files held by The National Archives, comprising some 165 individual reports, letters and minutes, transcribed and rendered searchable in July 2026.

Contents

About this report

This report reconstructs the development history of the bhangmeter — the British instrument for estimating the yield of a nuclear explosion from the time-profile of its light flash — using only the evidence contained in nineteen declassified archive files. The files divide into three broad groups: formal technical reports of the Admiralty Research Laboratory (ARL), the Atomic Weapons Research Establishment (AWRE) and the Home Office; trial reports of the Defence NBC School; and four thick War Office/Ministry of Aviation correspondence files recording, letter by letter, the ten-year effort to put an Army instrument into the field.

Citations are given in brackets using the short keys listed in Appendix B, e.g. (T5/64) for AWRE Report No. T5/64 or (WO-1: Minutes, 30 Sep 1960) for an item in the first War Office correspondence file. Quotations are verbatim from the source documents, including their inconsistent spellings (“Bangmeter”, “Bhangmeter”, “B Meter”). Where the record is incomplete or self-contradictory, this is flagged in the text and summarised in Appendix C. Nothing in this report draws on any source outside the nineteen files; readers should be aware that the files themselves record only part of a larger programme, and that several referenced reports (notably ARL/C/R13, DNBCS R16/65 and AORE Report 6/63) are not among them.

1. The instrument and its principle

A nuclear explosion announces itself with light, and the light behaves in a way no other phenomenon imitates. The output climbs to a first maximum within a few hundred microseconds, falls away to a minimum, and then rises again to a second, slower maximum before decaying. The Home Office described the effect as “the odd nature of the phenomenon. Instead of falling gradually from its first peak to zero, the light falls to a minimum and then rises again” (SA/PR 91). The interval between detonation and that first minimum — the “time to first thermal minimum”, written Tm — grows with the yield of the weapon: roughly 20 milliseconds for a nominal (20 kiloton) burst (ARL/R7/C), rising toward hundreds of milliseconds at megaton yields. Measure Tm, apply a calibrated scaling law, and the yield follows.

The bhangmeter is the instrument built around that measurement. In every British variant the architecture is the same: a photocell watches the sky through a filter; its output, amplified logarithmically to compress an enormous brightness range, deflects the beam of a cathode-ray oscilloscope whose single-shot time base is triggered by the steep leading edge of the flash itself; time markers are written along the trace; and a camera photographs the result. Because the method rests on a time measurement rather than an intensity measurement, it “has the merit of being independent of distance from ground zero and of atmospheric conditions” (ARL/R7/C) — with one crucial proviso, discovered the hard way at the weapon trials, that the observation be confined to a narrow spectral band, since the minimum arrives at different times at different wavelengths and the atmosphere filters the flash unevenly. Every British instrument therefore observed in a green band around 5,500–6,000 Å, the band in which the calibration data had been won and which, as ARL later insisted, could not be abandoned without repeating the nuclear trials themselves (ARL/C/N6).

The Home Office summarised the instrument’s worth in a sentence: “It is perhaps the simplest device available for measuring this quantity with a reasonable accuracy, say ±25 per cent, up to distances of about 200 miles” (SA/PR 91). Both halves of that claim — the simplicity and the 200 miles — would be tested repeatedly in the fifteen years these files cover.

On the name itself the record is laconic. ARL’s 1962 report carries a footnote: “This term is of American origin” (ARL/R7/C). The British services never quite settled the spelling; in 1964 a War Office branch ruled that “BHANGMETER, rather than BANGMETER” was “the more correct nomenclature” (WO-1: nomenclature minute, 9 Jun 1964), but catalogue entries, packing cases and the 1966 trials directive went on saying “Bangmeter” to the end.

2. Origins, 1952–1959

Two separate needs produced the British bhangmeter: the airman’s need to know what his weapon had done, and the defender’s need to know what the enemy’s had.

The offensive line came first. The instrument “was first developed in this country by the M.O.S. [Ministry of Supply] (now M.O.A. [Ministry of Aviation]), and was intended for use in the attacking aircraft to verify the yield of a weapon which had been dropped” (ARL/R7/C). An aircraft instrument could be forgiven much: it could be pre-set for the expected yield, aimed at the burst point, and switched on moments before the event. The pioneer of this work was D. Stone of the Ministry’s NAV A2 branch — ARL’s trials report credits “D. Stone (M.O.A.) who pioneered the bhangmeter in the U.K.” (ARL/C/R14) — and his 1959 note “A Weapon Yield Measuring Device (the Bhangmeter) with Special Reference to its Calibration” (M.O.S. NAV A2 Note 1/59) remained the standard calibration reference for years. AWRE’s Antler report likewise records that the instrument “has been under development by the Ministry of Supply (NAV A2) for the Air Ministry for some time” (T5/64). The design and manufacture of the hardware was in the hands of a single contractor throughout the whole story: Aeronautical & General Instruments Ltd (AGI) of Lansdowne Road, East Croydon.

The defensive line began, curiously, not with the bhangmeter at all. In the early 1950s the Home Office “made a detailed study of all the possible means of locating a nuclear burst” and, for reasons of economy, backed the simplest imaginable technology: pinhole images burned by the heat flash. Early models were tested by G. Stanbury at the first British trial at Monte Bello in 1952 — where “water vapour arising from a sea-surface burst masked much of the thermal radiation” — and the work matured into the Ground Zero Indicator (GZI), a four-hole pinhole camera giving bearing and elevation of the burst, installed at Royal Observer Corps posts on a grid roughly ten miles apart across the United Kingdom by about 1960. From about 1955 it was joined by the Bomb Power Indicator (BPI), “a copper bellows activated by blast pressure, design[ed] by A.W.R.E. for us” (SA/PR 99). NATO partners mocked the crudity of these devices — “in spite of the fact that they have then proceeded to buy them in considerable numbers!” — but the philosophy was deliberate: “a multiplicity of simple devices, independent of electronics”, enough of which would survive any attack (SA/PR 99).

Civil defence knew of the bhangmeter “about 1956, as a requirement for the armed services”, but the price put it out of reach: “we in C.D. saw these devices rather as a beggar in a fairy story sees the Prince’s golden coach” (SA/PR 99). How the beggar came by the coach is told in section 6.

3. The proving grounds: Buffalo, Grapple and Antler, 1956–1957

The bhangmeter earned its place at the British weapon trials in Australia and the Pacific. At Operation Buffalo (Maralinga, 1956) “an instrument gave clear records at over 20 miles range” (T5/64). Reports from Operation Grapple, the hydrogen-bomb trials of 1957, “indicated a fair measure of success” (T5/64), and Grapple and Antler data would later anchor the definitive calibration (ARL/C/R14). What Buffalo and Grapple began, Operation Antler finished: the first deliberate user trial of the instrument in Service hands.

3.1 The Antler user trial

The Atomic Weapons Trials Executive agreed on 13 March 1957 to a user trial of land-based and airborne bhangmeters at Antler. In the event only two land instruments could be spared — none were available for the airborne trial “owing to commitments at Operation Grapple” (T5/64). In early May 1957 Major F. M. Birnie, RA, of the Target Response Group, visited AGI in Croydon with AWRE and Ministry of Supply representatives for instruction; the two instruments reached AWRE in the first week of June and Maralinga village, in specially developed packing cases, in early August (T5/64).

The 1957 pattern was a valve instrument on a wheeled tubular frame, 135½ lb ready for use, drawing half an ampere at 230 V AC plus 24–28 V DC. Two photocell units looked through Perspex-rod optics with a 140° conical field of view, one behind a green filter (5,500–6,000 Å) and one behind a red (7,750–10,000 Å); triggering was taken from the green channel. Six preset range settings covered everything from 0–10 kt to a nominal 1–3,300 kt. The trace was photographed on 35 mm film in a clockwork Shackman Auto Camera Mk 3, developed wet in a mobile darkroom and read, still wet, on a microfilm reader — a quarter of a millisecond legible by eye (T5/64). A mains-driven flash simulator, whose flash “did not have the same characteristics as those of a fireball” but was bright enough to trigger the instrument, gave an end-to-end check; on trial days it was fired automatically by the range synchronising system at minus ten seconds, proving both instruments live moments before each burst (T5/64).

The three Antler rounds were fired at Maralinga in the spring of 1957: Round 1 (predicted probable yield 0.8 kt, 100-ft tower, Tadje) at 1435 CST on 14 September; Round 2 (predicted 6 kt, tower, Biak) at 1000 CST on 25 September; Round 3 (predicted 30 kt, suspended from a balloon at 1,000 ft over Taranaki) at 1615 CST on 9 October (T4/58). The bhangmeters stood 8, 8½ and 9¾ miles away respectively, in front of building FC2 at Roadside (T5/64).

3.2 Results and lessons

Judged as hardware, the trial was a success with one unexplained wound. Instrument No. 5 read within +9.6 and −1.7 per cent of AWRE’s reference timing on Rounds 1 and 2 and agreed on Round 3; instrument No. 7 read +4.1 and −4.4 per cent on the first two rounds and then, on Round 3, “did not produce a record” — the cause was never discovered (T5/64). Everything fell within the maker’s design limit of ±10 per cent on time, about ±20 per cent on yield, which Birnie judged “adequate for Service use”, though he noted honestly that AWRE’s own comparison data came from instruments working on the same principle, so the check was not fully independent. The Director had his yield estimate 29, 25 and 30 minutes after the three firings — a delay Birnie thought could be halved “by a very simple method indeed” (T5/64).

The failures were as instructive as the successes. After Round 3 the assessor used the wrong one of the two traces, and nobody noticed until the records were re-examined in England — hence the later insistence that the two traces be visibly different. Peculiar patterns on the films were traced, with Ilford’s help, to electrostatic discharge from winding film too fast in Maralinga air at under 30 per cent humidity; such marking over a real record “would make it unreadable”. The recommendations Birnie set down in his report read, with hindsight, like the specification of the next decade’s instrument: operation from Service batteries; larger wheels; visibly distinguished traces; a time base starting accurately at zero; a light-tight camera taking commercial daylight-loading cassettes with a way “of telling from the outside of the camera that the film is passing” — the lack of which he called “the most serious fault of all”; and better knowledge of the yield law band by band (T5/64).

One oddity of the record deserves note: although the trial took place in 1957 and interim results were reported immediately, AWRE Report T5/64 describing it was not published until April 1964 — six and a half years later, just as the Army programme described in sections 7–9 was struggling to define its production instrument. The trials establishment that made the measurement and the procurement machine that needed it were, for most of the story, barely connected.

3.3 The instrument among its rivals

The bhangmeter was never the only yield meter at the trials. AWRE’s 1957 guide to Maralinga operations lists the establishment’s methods: radiochemical analysis of bomb debris collected by aircraft, high-speed photography of the fireball’s growth (about 140 photographs at one-microsecond exposure), thermal instruments of all kinds, blast gauges, and remote-reading stations telemetering fallout gamma dose-rate (T43/58) — with rapid weapon-diagnostic telemetry recording events to fractions of a microsecond (AWRE-1957; T51/58). Gamma dose-distance measurements were explicitly weighed as a competing yield method at Antler before being deferred (T34/58). Against this battery of establishment science the bhangmeter’s virtues were different in kind: it was cheap, self-contained, needed no forewarning of burst time or place, and gave its answer in minutes rather than days. Those virtues are exactly what later recommended it to the soldier — and its ±20 per cent, one-parameter answer is what later disappointed him.

4. Reading the flash: the evolution of the scaling law

The instrument was only ever as good as the law relating Tm to yield, and the files record that law being remade three times.

At Antler the assessors worked from the American formula t = 0.0027 W½ (t in seconds, W in kilotons), taken from the US Department of Defense handbook Capabilities of Atomic Weapons (1957). It failed the test: yields computed from it ran “60 to 70% high” when the times were measured in the green band (T5/64). The reason, established by AWRE afterwards, is the spectral effect already noted — a broad-band or wrongly-banded measurement does not see the same minimum. AWRE’s revised analysis for the green band and kiloton range gave W = 0.078 t1.95, “or more approximately W = 0.07 t2” with t in milliseconds (T5/64).

The definitive calibration came from Operation Dominic, the American test series at Christmas Island in 1962, at which British bhangmeters recorded all twenty-four rounds (section 5). Combining Dominic with the earlier Grapple and Antler data — thirty-one shots spanning 1 to 8,000 kilotons — ARL fitted by least squares:

W = 0.045 Tm2.24 (W in kilotons, Tm in milliseconds)

with a standard deviation of 12.5 per cent and a worst error of +39 per cent across all burst types; on the Dominic air bursts alone the scatter fell to 8 per cent. ARL judged the single formula “adequate for general Service use” across half a kiloton to fifty megatons, while noting that separate laws for tower and air bursts would do better (ARL/C/R14). The same report confirmed a useful second string: the time to the second maximum followed the American relation W = 0.98 × 10−3 Tmax2 tolerably well for yields up to about 200 kt.

For training the Army later simplified again: the 1966 nomogram issued with the Service instrument was “based on the law Tm = 0.0025 W½”, with a rule of thumb — multiply the yield by 1.2 for an atmospheric burst, divide by 1.2 for a ground burst (WO-2: nomogram note, 22 Jul 1965; WO-3: nomogram, Jun 1966). Even in 1964 the question was not regarded as closed: the Defence NBC School’s trial report formally asked that “a competent authority should be requested to produce the best ‘Time to First Thermal Minimum/Yield’ graph or table” for 1 kt to 10 MT (R4/64). The calibration, in other words, remained live science for the whole life of the instrument — and, as section 9 records, its classification caused real trouble when unclassified training equipment had to be built around it.

5. The naval instrument, 1960–1965

5.1 A defensive problem

The Admiralty’s requirement inverted the aircraft designer’s assumptions. A warship cannot know when, where or how big: “The naval application, being defensive, requires an instrument which is able to stand in readiness for an event, at any time, of unknown yield, bearing and range” (ARL/R7/C). B. W. Allwood of the Admiralty Research Laboratory, Teddington, took the problem, and AGI — again — built the answer, described in ARL/R7/C of March 1962.

Three design choices defined the naval prototype. First, an approximately logarithmic single-shot time base of 500 ms, assembled from three sequential square-wave generators integrated through non-linear “Metrosil” elements, so that Tm from 1 to 500 ms — nominally 0.1 kt to 50 MT — could be captured “with roughly equal percentage accuracy at all points” and with no pre-setting of range whatever. Second, two logarithmic amplifier channels of different sensitivity writing simultaneous traces, covering an amplitude range of 1 to 105 without adjustment. Third, an always-ready philosophy: the CRT spot biased beyond cut-off (no shutter, no fogging even after 24 hours), a Polaroid Land camera delivering a readable record ten seconds after the event, and a trigger circuit — a 5 kc/s tuned filter feeding a Schmitt circuit — that fired on the flash itself and locked out retriggering for 600 ms (ARL/R7/C).

The optics served the ship’s motion: a two-inch polythene sphere within a six-inch opal globe gave 360° azimuth coverage and 240° vertical, holding the 90° elevation requirement under ±30° of pitch and roll. The detector was a Cintel VTA2 vacuum photocell behind a Chance OY2 filter centred at 5,600 Å; to protect the cell from a lifetime staring at the sky, its standing current was restricted to two microamperes (ARL/C/R14; ARL/C/N6) — a decision with consequences. A bridge network in the mast head compensated changes of ambient light. The mast unit was built to full Service environmental specification, with results the Christmas Island climate would shortly vindicate (ARL/R7/C).

5.2 Operation Dominic, 1962

The prototype got a calibration campaign no British instrument could have expected: the United States’ Operation Dominic at Christmas Island in 1962. Six British instruments deployed: the ARL prototype and three Ministry of Aviation-type instruments at “C site”, 16 to 36½ miles from ground zero, and two of the Home Office’s modified units, borrowed back for the trial, on Fanning Island about 200 miles away for the first twelve rounds. “Bhangmeters were successfully used on all twenty-four rounds”, and beyond calibration they were worked in earnest: yields declared in near-real time from the bhangmeter records matched the official American figures with a standard deviation of 8 per cent (ARL/C/R14).

The trial also delivered three sharp lessons. The photocell ambient-compensation circuit degraded the system time constant and “removed soon after the commencement of the trial”. The receiver proved “hardly sensitive enough” — the two-microampere protection had cost too much — and removing the inner diffusing sphere raised the light sensitivity about six-fold, after which nineteen rounds showed the sensitivity “to be near that required” (ARL/C/R14; ARL/C/N6). And Fanning Island failed absolutely: the two Home Office instruments stationed there for the first twelve rounds failed to operate on every one of them, though an observer saw the flash on all but one. ARL offered three explanations — plain lack of sensitivity; preferential atmospheric degradation of the green band over 200 miles; and geometry, the fireball lying below the horizon until after the first maximum, so that the sharp rise the trigger needed never arrived (ARL/C/R14). The Home Office’s cheerful “up to distances of about 200 miles” (SA/PR 91) did not survive contact with the experiment: the demonstrated maximum remained the 100 miles achieved once at Grapple-Z (ARL/C/R14). One incidental triumph: the mast-head unit “suffered no corrosion or any other form of deterioration” through a trial in which “other equipment was seen to suffer considerably” (ARL/C/R14).

5.3 The 1965 design study

Allwood distilled the accumulated experience into ARL/C/N6 of February 1965, “Design Information for a Naval Bhangmeter” — in effect the specification of a second-generation Service instrument, written “with the possible introduction of a bhangmeter into Service use” in view and in consultation with E. G. Odell and R. A. Fothergill of AWRE. Its prescriptions: transistors throughout; the vacuum photocell replaced by a silicon solar cell (a Ferranti MS11 type) that could stare at the sun without ageing, filtered back to the sacred green band because the calibration was “unlikely to be repeatable in any other waveband”; triggering sensitivity at least ten times the 1962 design, but not so high as to fire on “the natural background flicker of the sun”; a 700 ms sweep, quietly conceding that the 500 ms prototype had not truly reached 50 MT; crystal-controlled time markers replacing free-running oscillators; multiple detector heads to cover full azimuth from imperfect sites; an automatic reed-relay scheme to clamp out ambient-light wander at the instant of triggering; and a built-in test suite from flasher to crystal monitor. Two details reversed earlier practice with a historian’s candour: the two traces should now deflect in the same direction, “contrary to the statement” in the 1962 report, to spare unpractised observers; and a footnote formally corrected the trials report’s account of the sensitivity change (ARL/C/N6). Whether the second-generation naval instrument was ever built lies outside these files.

6. The beggar and the coach: civil defence, 1961–1965

In 1961 the golden coach arrived unbidden: “the Prince — in the unlikely guise of the Ministry of Aviation — said they had some bhangmeters which were no longer required” (SA/PR 99). These were aircraft-programme instruments now surplus. Six were modified “to read MT as well as KT” and in 1962 “incorporated in the U.K. W & M [Warning and Monitoring] system” (SA/PR 99), operated within the Royal Observer Corps structure alongside the GZIs and BPIs.

The Home Office’s operating notes (SA/PR 91, November 1961) describe what the ROC actually got: a two-part instrument — light receiver on a mast, recorder on a wheeled trolley — running from 230 V mains, with a linear single-sweep time base, a two-trace display timed by a free-running 250-pip-per-second oscillator, a Polaroid camera, a daily test-flash routine, and a warning lamp with a socket for a remote alarm. The notes are frank about its habits: lightning triggers it (the record closely resembles the daily test-flash record, so it can be told apart), and two bursts in quick succession write over one another so that it is impossible to tell from the record which came first. Pavry’s 1965 NATO paper adds the systemic limitations: no yield above about 15 MT, no direction, no time-of-burst record, a 140° field of view, and the standing problem of correlating a yield with a position found by other means — “difficult problems if a large number of explosions occurred in a short time” (SA/PR 99). (ARL, incidentally, put the same instruments’ field of view at 120° (ARL/C/R14) — one of several small disagreements catalogued in Appendix C.)

By early 1964 the Home Office wanted better. Stanbury “initiated exploratory discussions with A.W.R.E.” towards an instrument giving yield over the whole range plus direction and time of burst; AWRE agreed to study it, and Fothergill presented the results alongside Pavry’s paper at the NATO Scientific Working Party meeting of 29 June–2 July 1965 (SA/PR 99). That study is the visible seed of the automatic detection system the Army files later call “AWDREY”, whose characteristics — 75 ft cable, 12/24 V operation, installation in an APC or office truck — BAOR was commenting on by September 1966 (WO-3: BAOR letter, 14 Sep 1966). The bhangmeter, conceived for the attacker and matured for the defender, was by mid-decade the acknowledged ancestor of the national burst-detection network.

7. The Army programme I: definition and contract, 1960–1961

The Army came to the bhangmeter with a doctrine, not just a curiosity. Yield was the key that unlocked fallout prediction, casualty assessment and the crucial first NBC report; and the alternative — a sentry timing cloud growth by eye and submitting NATO Form NBC 1 — took ten minutes. “The Bangmeter provides the yield and time of the detonation within 15 seconds of burst” (WO-2: Trials Directive, 24 Jun 1966).

The programme opened formally in mid-1960: a War Office initiating letter of 18 May, Military Characteristics of 30 June, and — in between — the first General Meeting on the “Development of Army-Type B Meter” at AGI’s Croydon works on 20 June 1960. The cast assembled there ran the project for the next nine years: J. A. Peartree of the Ministry of Aviation’s TL1 branch in the chair for most of it; Allwood representing the Admiralty art; the War Office equipment branches (EP4/AEP4); the Joint School of Nuclear and Chemical Ground Defence at Winterbourne Gunner (later the Defence NBC School) as user; and AGI’s designers, among them L. V. F. Russell. The meeting decided to base the Army instrument squarely on AGI’s existing ARL work, and to lend the School an overhauled Ministry of Aviation instrument with a Polaroid camera and a flash simulator in the meantime (WO-1: Minutes, 20 Jun 1960).

The second General Meeting, on 30 September 1960, fixed the shape of the thing: a cylindrical, resiliently mounted unit in a tubular trolley frame; a detector head recording in the green waveband with roughly 240° hemispherical coverage, mountable up to 100 ft from the unit; two traces of different sensitivity; logarithmic amplifiers to reach megaton yields. The experimental “A” model would carry a manual Polaroid Land camera while an automatic camera — two-second wind-on, a built-in 24-hour watch photographed beside each trace, a frame counter warning at the eighth shot — was developed for the production “B” models. The programme optimistically attached dates: A model by 31 March 1961, three B models late 1962 (WO-1: Minutes and School note, 30 Sep 1960). The Ministry told the War Office that August there were “good prospects of getting equipment into service in 1962/63” (WO-1: MOA letter, 17 Aug 1960). It went into service in 1966.

The development contract, KM/3C/036/CB.21(d), was placed with AGI on 27 February 1961 for the A model. Then the record goes strangely quiet: apart from masts issued for detector-head trials — of which, when the question was asked in February 1963, the School could find no record, “the technical officers here during 1961 have all been posted away” (WO-1: mast query, Feb 1963) — the files preserve almost nothing of 1961–62.

8. The Army programme II: six instruments as insurance, and the troubled A model, 1963–1964

The programme that re-emerges in February 1963 is in a hurry. The War Office was “very anxious” to give BAOR six bangmeters within twelve months, and proposed buying copies of the prototype rather than waiting for proper development, AGI’s ability to copy being “more certain than its ability to complete development to the production stage”. The minute put production cost at £750–£1,000 an instrument, prototypes at about twice that, and stated the strategic frame: “War Office policy is that Bangmeters will be required as part of a nuclear surveillance system. The probable requirement is about fifty” (WO-1: loose minute, 14 Feb 1963). The Ministry agreed, estimating £10,000–£12,000 for the six; the financial case that went up in March said £12,000–£15,000 and described the instrument’s pedigree in one line: “designed by the Admiralty and used successfully in recent atomic trials” (WO-1: EP4 loose minute, 6 Mar 1963). AGI quoted on 21 May 1963: six modified-A-model bangmeters at £1,500 each, spares, packaging, test specification and documentation bringing the total to about £12,000, delivery nine months from order (WO-1: AGI quotations, 21 May 1963).

The A model itself meanwhile stumbled from bench to bench. Delivered to ARL for engineering tests in late August 1963, it went back to AGI for adjustments before the tests finished in November. ARL’s verdict was double-edged: “the instrument performs its overall function of measuring ‘time to minimum’ with the required accuracy… The automatic camera feed mechanism is ingenious in design, but needs improvement before it can be used for operational purposes” (WO-1: ARL report quoted, 13 Dec 1963). In November 1963 AWRE’s Applied Physics Division formally took over as research-and-development authority for nuclear-burst-indication instruments, and in the first week of December, on AWRE’s bench, the A model failed — overloaded diodes in the 400-volt supply (WO-1: AWRE takeover minute, 20 Nov 1963; progress papers, Dec 1963). When AWRE, the War Office sponsor and the School gathered around the machine on 4 December, the sponsoring branch was seeing the instrument for the first time: “This is a pity since several obvious user requirements have not been included” (WO-1: minute, 12 Dec 1963).

December 1963 also surfaced the problem that never went away: power. The design treated the 24-volt battery input as a standby, but a field instrument on continuous watch would draw “180 amp/hr for 24 hours, which in terms of weight would be in the order of 300 lbs of batteries” a day; full transistorisation was identified as the cure (R4/64, Annex A; WO-1: modification minute, 20 Dec 1963). AWRE recommended cutting the detector cable from 120 to 50 ft, the School concurring; the sponsor overruled — “not acceptable to reduce… to less than 100 feet” (WO-1: EP4 ruling, 31 Dec 1963).

A sixteen-point modification programme was settled at AGI on 30 January 1964 — protected supplies, one consolidated power unit, simplified time markers, a confirmed 100 ft cable, investigation of detector-head microphony among them — with six B models promised, “two at least by August, 1964” (WO-1: progress minutes, 30 Jan 1964). The funding requisition P9/CS/Z8/61071935 followed in April; the instruments were catalogued under the interim designation Z8/00000-01770 “Bangmeter”, to be produced “as an interim measure under a development contract”, without formal drawings, and packed to Trade Export standard only (WO-1: MOA letter, 20 May 1964).

8.1 The School passes judgment: R4/64

On 21 April 1964 the Defence NBC School put the repaired A model through a limited evaluation — the trial written up as Report R4/64, the fullest surviving description of the machine. What had been a 135½ lb trials instrument in 1957 was now a 211 lb prototype: oscilloscope and electronics in a metal drum on a tubular frame, automatic eight-frame Polaroid camera photographing trace, watch-face and frame number together, detachable detector head on 100 ft of cable weighing 22 lb, inverter for 24 V working, two solid rubber wheels. Two servicemen worked it from the maker’s interim handbook — “adequate but too lengthy and involved compared with normal service instructions”. It triggered reliably from its simulator out to 25 ft (and from a thunderflash at 33 ft), read the simulator’s fixed 4–5 ms “yield” consistently, changed film in a minute and a half, and changed power source in ten minutes, most of them spent on the eleven screws of the power take-off cover (R4/64).

The School’s conclusion was measured: the A model “functions satisfactorily”; two men can carry it, one “intelligent soldier of the type chosen for NBC Centres” can operate it; it fits a Land Rover. But the B model must be lighter; it needs a proper head bracket, tool kit and waterproof PVC cover; Type 47 (3000 ASA) Polaroid film must be standard, the slower Type 42 being useless; the simulator must offer variable yields instead of one; and the false-alarm question — lightning “may be very similar to a nuclear burst”, gun flash at night, shell bursts, headlights — needed trials of its own. Doctrine settled on issue “down to Brigade Headquarters NBC Centres”, the head sited high within 100 ft of the set, not necessarily line-of-sight, since “reflected light from a burst can trigger the mechanism” (R4/64).

9. The Army programme III: B models, film and the simulator, 1965–1966

The first B model reached the School on 19 May 1965, and the programme immediately acquired three new problems: schedule, film and secrecy.

Schedule first. At the May 1965 progress meeting AGI declared the design “final electronically” but slipped the remaining deliveries to the end of September; the sponsor noted glumly that “deliveries would now be too late for use of the bangmeter in Army exercises in 1965”. By October — AGI having lost staff — four of the six were promised only “by the end of December 1965”, a slip the minutes call “very disturbing” (WO-2: progress meetings, 25 May and 15 Oct 1965). Five instruments finally reached Ordnance stores on 21 April 1966, and the first was in Germany by July (WO-2: delivery note, 21 Apr 1966; WO-3: BAOR letter, 11 Jul 1966).

Film second. AWRE tests in 1965 showed the standard 3000 ASA Polaroid film fogged by gamma doses of 3–5 roentgens; the School’s own irradiation trials confirmed 1–2 r acceptable, 5 r “barely readable with difficulty”, while the slower 200 ASA stock survived to 25–30 r but produced traces too faint to use without modifying the instrument. The sponsor’s alarmed observation — “These sort of intensities will be common in General War” — earned the reply that in Service film supply “no special precaution is taken to prevent ‘fogging’ from fall-out radiation” (WO-2: film correspondence, Jul–Oct 1965). AGI investigated a 200 ASA conversion and, in December 1966, gave the final answer: “the CRT trace cannot be brightened enough to give satisfactory photographs with 200 ASA film… various factors such as uncertainty of success, time and cost may well rule out any further work” (WO-4: film letter, 13 Dec 1966). The Service bhangmeter went to war-station, in other words, with film its own quarry — fallout — could blind.

Secrecy third. The training simulator (built by AWRE itself under “a crash programme”, seven fully engineered Simulator Flash Stations to AWRE drawing WA/102/76) had to present selectable “yields”; but the true yield/Tm relationship was classified. Rather than alter timing circuits, AWRE simply relabelled the panel: “the association of the original simulated yields with the tmin values built into the equipment would reveal the classified relationship” (WO-2: security minute, 23 Mar 1966). Training material was rebuilt on an unclassified law, and the panel settled at 2½, 20, 55, 170 and 430 kilotons (WO-2: yield minutes, 25 Mar and 7 Apr 1966). The simulator had teeth of its own: switching it on flashed the head and could trigger the bangmeter, so the interim operating instructions of June 1966 prescribed a strict sequence — bangmeter off first, four minutes of flash-tube warm-up, yield knob behind an interlocked flap (SIM-OI).

Through the first half of 1966 the pieces converged on Germany: the Service names settled (“Bangmeter” and “Simulator Flash Station”); design information downgraded from Confidential to Restricted; 120 packs of Type 47 film ordered at about sixteen shillings a roll (WO-2: order, 1 Mar 1966); consignments redirected from a single Corps address to the six formation headquarters — 1 (BR) Corps, 1, 2 and 4 Divisions, 4 Guards and 7 Armoured Brigade Groups; and REME in BAOR made responsible for inventing vehicle mountings and head cages, the Ministry confessing it had “NO information on Vehicle Installation” (WO-2: DEME letters, May 1966). One anomaly in those DEME letters — the instruments are twice called “D Models”, a designation appearing nowhere else — is noted in Appendix C. The Trials Directive issued to HQ BAOR on 24 June 1966 set the aims (tactical value, allocation, functioning, drills), warned in advance that lightning might counterfeit “the important Initial Form NBC 1”, and required a final evaluation by the end of the year — a schedule later revised to an interim report by December 1966 and a full report by mid-1967 (WO-2: Trials Directive, 24 Jun 1966; WO-3).

10. The BAOR trials, 1966–1967

The field trial got off to a stumbling start — the first set reached 15 Army Base Ordnance Depot “just too late to be used in an exercise at the end of June [1966]”, and with no film (WO-3: BAOR letter, 11 Jul 1966) — but by mid-September five of the six sets were in Germany, the first installed by 1 (BR) Corps in the office truck of the Corps NBC cell behind locally made brackets, cable drum and head cage (up to £50 a vehicle was authorised for six vehicles). One instrument deployed on Exercise FALLEX that autumn, an outing chiefly memorable for what was missing: the handbooks in theatre lacked their photographs, external alarms “consisted only of the plugs”, and no technical manuals had arrived — “TL1 has been pressing for completion and publication of the manuals for over a year” (WO-3: FALLEX points, 3 Nov 1966). The Ministry of Aviation meanwhile waived environmental testing of the seventh, retained development model as pointless, and marked the old A model for cannibalisation (WO-3: letters, 5 Oct 1966).

The Corps’ initial report of 29 November 1966 was the first field verdict, and it set the tone. On the credit side: twenty hours’ continuous running without defect, a head that sat out thirty-six hours of dull, showery weather and “seemed to be weather-proof”, a successful cage, easy installation in a 3-ton office body. On the debit side, three sentences that defined the rest of the programme: at low kiloton settings readings disagreed with the simulator by 20 per cent, “In fact, it is at the lower yields that the greater accuracy is needed”; “If several strikes occur within a period of one minute, the Bhangmeter (and its operator) cannot cope”; and, fatally, “Information of yield is useless without at any rate a rough indication of ground zero” (WO-4: initial report, 29 Nov 1966). MGO Nuclear’s technical staff pushed back in January 1967 — the 20 per cent conflated instrument error with simulator error, and “The only way that the accuracy of the bhangmeter can really be measured is by testing it with a nuclear explosion” — while drawing the strategic moral that yield-measuring art must be kept alive for the successor system (WO-4: MGO minute, 10 Jan 1967).

Through the first half of 1967 the brigades trialled their sets on command-post exercises (BATTLEAXE 671 and 672, FIRST PARADE) and live-firing ranges. The range results made the false-alarm file thicker: at Sennelager a WOMBAT anti-tank gun’s backblast left it alone until the instrument was repositioned, after which the camera fired three times “although there was no weapon flash at the time”; at Hohne a deliberate test against Centurion 105 mm muzzle flash at 250–300 yards drew no response at all, yet with the head raised on a tower balcony it triggered on nearby tank fire, and again the next day from a vehicle roof — while an actual thunderstorm on 12 April 1967 had “no effect on equipment” (WO-4: 20 Armd Bde annex). The consolidated BAOR report of 24 August 1967 pulled no punches: “The Bangmeter in its present form is not reliable”; “too heavy and too bulky considering the very limited information which it provides”; triggered by “lightning, flashing headlights etc… a great waste of film”; one brigade’s set “set itself off on several occasions for no apparent reason”, lost a valve (reducing the millisecond graph to a straight line) and refused the 170 kt simulator setting. The brigades’ recommendations asked for exactly what the instrument did not have: a bearing capability (±3°), flash-to-bang ranging, an operator alarm, man-portability (WO-4: BAOR trial report, 24 Aug 1967). One late discovery escalated beyond the trial itself: a monitoring team found that speech into the BID 150 secure-speech equipment modulated the bangmeter’s voltages, which the instrument could then radiate — “the possible breach of security of such radiations must be further investigated” — awkward, since every brigade wanted the bangmeter in the same vehicle as the BID 150 (WO-4: BAOR letter, Sep 1967).

The School’s review in October 1967 added perspective and one correction: a brigade’s complaint of wrong yields dissolved on inspection — its photographs “do in fact provide the correct result in each case”; the nomogram had simply been used incorrectly (WO-4: DNBCS review, 3 Oct 1967). The pattern of 1957 — sound measurement, human error at the reading step — had reproduced itself in Germany a decade later.

11. Successors and closure, 1967–1969

By late 1967 the sponsors had conceded the B models’ nature: they “were issued mainly for trials purposes and it was realised that they were not very suitable for operational use in time of nuclear war” (WO-4: Ministry of Technology letter, 10 Nov 1967). The forward programme had in fact been running alongside all along. The June 1964 R&D forecast already pointed past the interim buy — six bhangmeters in 1964/65, perhaps ten ever — to a “Nuclear Burst Indicator” under General Staff Requirement 3147: six in 1966/67, twenty-four thereafter (WO-1: AEP4 minute, 9 Jun 1964). MGO Nuclear’s 1967 minute gives the NBI’s intended reach — sub-kiloton bursts at 20 miles, 1–10 kt at 35, over 10 kt at 50 — and insists the yield-measuring art be preserved for it; the Ministry of Technology set AWRE to re-examine radar for the ground-zero half of the problem, the half the bangmeter never had (WO-4: MGO minute, 10 Jan 1967; MinTech letter, 10 Nov 1967). On the civil side the same convergence produced AWDREY (section 6).

The doctrinal testament is the United Kingdom’s January 1968 paper to the NATO NBC Defence Panel, “The Requirements for Nuclear Burst Information”. It defines the four essentials — yield, ground zero, height of burst, time — and accepts yield “within a factor of two”. It names the bhangmeter’s principle as the national method: “Time to Thermal Minimum… is the principle used in the UK Bangmeter, at present issued to HQ’s of brigades in 1 (Br) Corps”. And it draws the lesson of the whole file: observer-and-signals reporting under STANAG 2103 is “probably as good as can be expected” yet “liable to be overwhelmed by massed strikes”; a future system should give every headquarters from battalion upwards its own automatic burst data, independent of communications (WO-4: NATO paper, Jan 1968).

The archive’s last word is an anticlimax perfected by delay. On 22 July 1969 the Ministry of Technology circulated, “at last”, ten copies of AGI’s Handbook No. 87 — the definitive technical handbook for the B models “supplied in 1965/66… which has been awaited for so long” — with a note that it remained to be decided “whether or not it can be accepted in fulfilment of the contract item” (WO-4: covering letter, 22 Jul 1969). The instruments it described had by then been in Germany, and out of favour, for three years.

12. The instrument in evolution

Five distinct British configurations appear in the files. All share the AGI pedigree, the green waveband and the two-trace logarithmic display; the table shows what changed.

Instrument Date / source Distinctive features Fate
MoS/MoA aircraft-type; Home Office modified units late 1950s; SA/PR 91 Linear single-speed time base; free-running 250 pip/s markers; manual Polaroid; 140° (or 120°) field of view; mains-powered; six units modified KT→MT Six into UK Warning & Monitoring system 1962; two failed at 200 mi on Dominic; superseded by AWDREY studies from 1964
Antler trials pattern 1957; T5/64 Two photocells (green + red), 140° Perspex-rod optics; six preset ranges; clockwork Shackman 35 mm camera, wet development; 135½ lb Two-round success, one unexplained failure; recommendations shaped all later models; reported only in 1964
ARL naval prototype 1962; ARL/R7/C Logarithmic 500 ms single-shot time base (no range setting); 1–105 amplitude on two traces; always-ready, no shutter; Polaroid (10 s); 360° azimuth sphere-in-globe optics; Service-spec mast head Starred at Dominic 1962 (all 24 rounds); sensitivity ×6 fix mid-trial; fed the 1965 design study
AGI “A” model (Army prototype) 1963–64; R4/64, WO-1 Automatic 8-frame Polaroid with watch and frame counter; detachable head on 100 ft cable; 24 V inverter; 211 lb Failed at AWRE Dec 1963; repaired; School evaluation Apr 1964 defined B-model changes; cannibalised from Oct 1966
AGI “B” model (Service, 6 built) 1965–66; WO-2/3/4 Traces reading left-to-right (user request); simulator flash station (AWRE-built, 5 preset “yields”); Type 47 3000 ASA film; vehicle-mounted, REME cage Issued to six BAOR formation HQs 1966; damned by 1967 trial reports; awaiting handbook until Jul 1969; pointer to GSR 3147/NBI

12.1 Scaling laws in sequence

Law Basis and band Source Recorded performance
t = 0.0027 W½ (t s, W kt) US DoD, Capabilities of Atomic Weapons (1957); band unstated used at Antler 1957 (T5/64) Yields 60–70% high when t measured in green band
W = 0.078 t1.95 ≈ 0.07 t2 (t ms) AWRE re-analysis, green band, kiloton range T5/64 (1964) Corrects the Antler discrepancy
W = 0.045 Tm2.24 (Tm ms) Least squares, 31 shots (Grapple, Antler, Dominic), 1–8,000 kt, green band ARL/C/R14 (1963) s.d. 12.5%, worst +39% (all burst types); s.d. 8% on Dominic air bursts; recommended ½ kt–50 MT
W = 0.98 × 10−3 Tmax2 (second maximum) US DoD (1957); checked on Dominic ARL/C/R14 (1963) Fair agreement up to ≈200 kt
Tm = 0.0025 W½ (training nomogram) Unclassified training law; ×1.2 atmospheric, ÷1.2 ground burst AEP4 nomogram, Jun 1966 (WO-2, WO-3) Issued with Service handbook; incorrect use of the nomogram caused false “inaccuracy” findings in BAOR

13. Assessment

As physics and engineering, the bhangmeter was a quiet success. A cheap photocell-and-oscilloscope instrument, standing unattended, measured the yield of nuclear explosions to better than ±20 per cent across four orders of magnitude, at ranges of tens of miles, within seconds — and did it on every one of the twenty-four Dominic rounds. Its calibration law, refined from a borrowed American formula to a thirty-one-shot least-squares fit, held to a 12.5 per cent standard deviation. The naval prototype survived Christmas Island weather that wrecked lesser equipment, and the 1965 design study reads as a competent, self-critical piece of instrument engineering that corrected its own earlier claims in print.

As a procurement story, the files read less happily. An instrument promised for service in 1962/63 reached Germany in 1966, at 211 lb where the user wanted man-portability, drawing battery weight no unit could carry, photographing its evidence on film that fallout would fog, its technical handbook arriving in 1969 — three years behind the hardware and two behind the verdict. The sponsoring branch first saw the prototype at the end of 1963, after the key design decisions were frozen; the trials establishment’s definitive report on the 1957 user trial appeared in the same season the production design was being settled, too late to inform it cheaply.

Yet the sharpest judgment in the record is not about lateness or weight but about information. “The priorities of information regarding a nuclear strike are ‘where’ and then ‘how big’. At present the Bangmeter provides the answer to the second factor” (WO-4: 11 Inf Bde report, 1967). The instrument answered the second question superbly and the first not at all, and no accuracy in Tm could fix that. What the programme actually purchased, the later documents suggest, was not six instruments but a national competence: a proven measurement principle, a calibrated law, a trained user community and a clear statement of requirements — all of which flowed directly into GSR 3147, the Nuclear Burst Indicator, and AWDREY. The 1968 NATO paper, which enshrines “time to thermal minimum” as the UK method while demanding automatic burst data at every headquarters, is less an epitaph for the bangmeter than its will.

Appendix A: Chronology

Date Event
1952 Monte Bello: Stanbury tests heat-flash shadow instruments; origin of the Ground Zero Indicator (SA/PR 99)
c. 1955 Bomb Power Indicator (AWRE copper-bellows design) work for Home Office begins (SA/PR 99)
c. 1956 Bhangmeter known in UK “as a requirement for the armed services”; MoS (NAV A2) developing aircraft instrument for the Air Ministry — D. Stone the UK pioneer, AGI Ltd the contractor (SA/PR 99; T5/64; ARL/R7/C; ARL/C/R14)
1956 Operation Buffalo, Maralinga: bhangmeter “gave clear records at over 20 miles range” (T5/64)
13 Mar 1957 Atomic Weapons Trials Executive approves bhangmeter user trial at Antler (T5/64)
1957 Operation Grapple: “a fair measure of success” with bhangmeters (T5/64)
14 Sep – 9 Oct 1957 Operation Antler, Maralinga: Rounds 1–3 (predicted probable yields 0.8, 6 and 30 kt; two towers, one 1,000-ft balloon). Two bhangmeters at 8–9¾ mi; ±10% timing; yield to Director in 25–30 min; one unexplained failure on Round 3 (T5/64; T4/58)
1958 Operation Grapple-Z naval bhangmeter measurements (reported in ARL/R1/R254, Apr 1960): instrument operates at 100 miles (ARL/C/R14)
1959 Stone’s calibration note, M.O.S. NAV A2 Note 1/59 (ARL/R7/C; ARL/C/R14)
18 May / 20 Jun / 30 Sep 1960 War Office initiates Army programme; 1st and 2nd General Meetings at AGI Croydon define the Army instrument; A model promised 31 Mar 1961, service entry forecast 1962/63 (WO-1)
27 Feb 1961 Development contract KM/3C/036/CB.21(d) placed with AGI (WO-1)
1961 Ministry of Aviation declares aircraft bhangmeters surplus; six modified KT→MT for Home Office (SA/PR 99); HO operating notes SA/PR 91 (Nov 1961)
1962 Six bhangmeters incorporated in UK Warning & Monitoring system (SA/PR 99)
Mar 1962 ARL/R7/C describes the naval prototype (ARL/R7/C)
Apr–Jul 1962 Operation Dominic, Christmas Island: 6 UK instruments record all 24 rounds at 16–46 mi; live yield declarations s.d. 8%; both instruments at ~200 mi (Fanning Is.) fail on all 12 rounds (ARL/C/R14)
Feb–May 1963 War Office case for six interim instruments (“probable requirement is about fifty”); AGI quotes £1,500 each (WO-1)
May 1963 ARL/C/R14 publishes W = 0.045 Tm2.24 from 31 shots (ARL/C/R14)
Aug–Dec 1963 A model at ARL (verdict mixed), AWRE takes over R&D authority (Nov), A model fails at AWRE (Dec); battery problem quantified at ~300 lb/day (WO-1; R4/64)
Early 1964 Home Office opens advanced-surveillance discussions with AWRE — seed of AWDREY (SA/PR 99)
30 Jan 1964 Sixteen-point modification programme agreed at AGI; six B models ordered against requisition P9/CS/Z8/61071935 (WO-1)
Apr 1964 AWRE T5/64 (Antler bhangmeters) finally published; DNBCS evaluates A model 21 Apr (R4/64, issued 30 Jun 1964)
19 May 1965 First B model delivered to DNBCS (WO-2)
1965 Film-fogging trials (3–5 r fogs Type 47); simulator programme at AWRE; classified yield law forces relabelled simulator panel (WO-2)
Feb 1965 ARL/C/N6 second-generation naval design study (ARL/C/N6)
Jun–Jul 1965 Pavry/McAulay NATO paper; Fothergill presents AWRE advanced-system study (SA/PR 99)
Apr–Jul 1966 Five B models to Ordnance stores (21 Apr); Trials Directive to BAOR (24 Jun); first set in Germany, no film (Jul) (WO-2; WO-3)
Oct–Nov 1966 Ex FALLEX deployment; missing manuals; Corps initial report: yield-without-GZ “useless”, 20% low-yield discrepancy, cannot cope with massed strikes (WO-3; WO-4)
Feb–Aug 1967 Brigade trials (BATTLEAXE 671/672, FIRST PARADE; Sennelager and Hohne firing tests). Consolidated BAOR report: “not reliable… too heavy and too bulky”; false triggers from muzzle flash, headlights, and spontaneously; BID 150 radiation hazard found (WO-4)
Oct–Nov 1967 DNBCS review (wrong nomogram explains “inaccuracy”); GSR 3147 study revived incl. radar GZ investigation; B models acknowledged as trials equipment (WO-4)
Jan 1968 UK paper to NATO AC/225 Panel VII: “The Requirements for Nuclear Burst Information” (WO-4)
22 Jul 1969 AGI Handbook No. 87 for the B models arrives “at last” — the file’s final document (WO-4)

Appendix B: Sources

All sources are files of scanned documents from The National Archives, transcribed into searchable form in July 2026. Archive file identifiers below are the filenames of the scanned PDFs; the key is the citation used in this report.

Key Archive file Document(s)
SA/PR 91 259709392513068 Home Office Scientific Adviser’s Branch, “Notes on the Operation of the Bhangmeter”, Nov 1961
SA/PR 99 259763139597201 F. H. Pavry, “The Use of the Bhangmeter in War-Time Nuclear Surveillance Systems”, 25 Jun 1965 (NATO Scientific Working Party)
ARL/R7/C 261241160284435 B. W. Allwood, “A Naval Bhangmeter for the Determination of the Yield of a Nuclear Weapon”, ARL, Mar 1962
ARL/C/R14 261463382627723 and ADM 2042281 (two copies of the same report) B. W. Allwood, “The Use of Bhangmeters for the Determination of the Energy Released in Nuclear Explosions”, ARL, May 1963
ARL/C/N6 260807346342654 B. W. Allwood, “Design Information for a Naval Bhangmeter”, ARL, Feb 1965
T5/64 261904764000130 Maj. F. M. Birnie, “Operation Antler: The Functioning of Bhangmeters”, AWRE Report T5/64, Apr 1964
T4/58 273778688604309 R. A. Siddons & D. Sams, “Operation Antler: Theoretical Predictions”, AWRE T4/58, Mar 1958
T51/58 262649616721806 J. W. Jarvis, “Operation Antler: Weapon Telemetry”, AWRE T51/58, Feb 1959
T14/59 263239258815558 Kendall, Johnson, Johnson & Nelson, “Operation Antler: An Attempt at Measuring the γ-Flash Spectrum”, AWRE T14/59, Dec 1959
T34/58 263534589065010 J. A. Carr, “Operation Antler: Gamma Dose-Distance Measurements”, AWRE T34/58, 1958
T43/58 260309789997655 D. F. Godward, “Operation Antler: The Remote Measurement of the Variation with Time of Gamma Dose-Rate from Fallout”, AWRE T43/58, Nov 1958
AWRE-1957 271709057032864 “Instrumentation and Operation of a Land Based Nuclear Weapon Test”, AWRE, 22 Jan 1957; with “VOLCANO — Summary Statement”
R4/64 272036990780990 Maj. D. N. Forrow, “Bhangmeter ‘A’ Model (Prototype) — Limited Evaluation Trial”, Defence NBC School Report R4/64, 30 Jun 1964
WO-1 273235010187377 War Office / Ministry of Aviation / AGI correspondence, 54 documents, 1960–64
WO-2 268441476408874 Trials directive, B-model and simulator correspondence, 66 documents, 1964–66
WO-3 267054295439487 BAOR trials, FALLEX, costs and installation correspondence, 17 documents, 1966–67
WO-4 266653792352011 BAOR trial reports, NATO requirements paper, Handbook No. 87, 12 documents, 1966–69
SIM-OI 267170679385344 “Interim Operators Instructions for Simulator Flash Station SIW/A10276”, MOD AEP4, Jun 1966

Appendix C: Gaps, anomalies and disagreements in the record

The following limitations of the source files bear on the account above.

Gaps. The Army file is nearly silent from mid-1961 to the end of 1962, and the 1961 detector-mast trials were undocumented even to contemporaries. Several referenced reports are absent: ARL/C/R13 (the Dominic data volume), ARL/R1/R254 (Grapple-Z), DNBCS Reports R16/65 (B-model evaluation) and R2/59, AORE Report 6/63 (concept of use), and the G Tech final report of 1967. Two brigade trial reports (4 Guards and 12 Infantry Brigades) are referenced but not present. Physical damage removed some text: pages 2–3 and Annex A of the consolidated 1967 BAOR report are missing from the scan; a 1960 AGI meeting note ends mid-sentence in both carbons; the simulator operating instructions lose their final page; and one 1966 BAOR letter survives only as a fragment.

Anomalies. DEME letters of May 1966 twice call the BAOR instruments “D Models”; every other document says “B”. The AGI contract number appears in several variants (KM/3C/036/CB.21(d), C.B.21(b), CB.16(c)), evidently one contract cited carelessly. The signature block of the 1957 AWRE instrumentation guide reads “22nd January, 1956”, almost certainly a typist’s error for 1957.

Disagreements between documents. The Home Office gives the modified instruments a 140° field of view; ARL says about 120°. The 1961 Home Office notes claim usefulness “up to about 200 miles”; the 1962 Fanning Island deployment failed on all twelve rounds at that range. ARL 1962 preferred oppositely-deflecting traces; ARL 1965 reversed this “contrary to the statement” in the earlier report. ARL 1962 claimed the 500 ms sweep covered 50 MT; ARL 1965 states a 700 ms sweep is needed to reach 50 MT (and the Home Office put its own units’ ceiling at about 15 MT). A 1965 ARL footnote formally corrects the 1963 trials report’s description of the sensitivity modification. These are recorded here as found; no attempt has been made to adjudicate beyond what the documents themselves say.