1. Pursuant references (a) and (b) the authors of this report, Vice Admirals K. S. Masterson and L. M. Mustin, USN (Retired), have inquired informally into the subject explosion, to assist the Chief of Naval Material and other officers and officials as requested, in matters related to it.
2. The casualty resulted from the explosion of an 8” high capacity projectile Mk 25-2 in the bore of the center gun of turret 2. Explosive filler of the projectile was 21.3 pounds Explosive D. It was fitted with a base detonating fuze (BDF) Mk 48-4, a nose auxiliary detonating fuze (ADF) Mk 55-0, and either a steel nose plug or a point detonating fuze (PDF) Mk 29-3. (The latter uncertainty is discussed further in the later paragraph.)
3. Our inquiry included consultation 5-9 October at Naval Base Subic Bay, Republic of the Philippines, with the formal investigating officer (Rear Admiral P. P. Cole, COMSERVGRU THREE); and during the same period, discussions with the Commanding Officer NEWPORT NEWS (Captain W. F. Zartman) and various of the ship’s company; examination of the turret and the ammunition recovered from it; and on-scene consultation with the ordnance technical team sent to Subic by the Chief of Naval Material in connection with the casualty. Our inquiry also included consultations in HQ NAVORDSYSCOM; review of information and informed opinion available in the headquarters and from other ordnance-related technical commands: and examinations 13 and 30 October at NWL Dahlgren, Virginia, of 8” projectile detonation experiments done in relation to this casualty, and of the parts of the damaged gun and exploded projectile recovered from NEWPORT NEWS and returned to Dahlgren for analysis.
4. We received the most complete cooperation, assistance, and support in all phases of our inquiry, from every fleet and Navy Department level concerned, notably COMNAVORDSYSCOM and COMSERVGRU THREE.
5. We noted with satisfaction the evidence that throughout this deployment the ship’s performance had been outstanding, and that this had extended to the emergency actions taken in this casualty.
6. In-bore projectile explosions have been occurring ever since explosive-loaded types came into use, and have been experienced with every recorded type of explosive filler adopted for service. Reference (c) notes some 298 such explosions since 1888 in Navy guns of 3” and larger calibers (the total including many which were deliberately induced at proving grounds for various purposes). Four have been reported since reference (c), including that in NEWPORT NEWS, the latter being the first in that caliber in the nearly fifty years it has been in service. Army experience is believed comparable.
7. During World War II the Navy fired some 3,104,000 rounds in 3” and larger calibers, out of which 10 shipboard in-bore prematures were recorded. Much of that ammunition reflected design philosophy and technology dating from long before the war. The new-generation designs which appeared late in the war have been with us ever since, so afford a coherent basis for consideration. The record of shipboard in-bore explosions since World War II, 3” and larger, is as follows:
Number in bore prematures
Total Rounds Fired
1946–1950 - 0
Total rounds fired: 850,000 (est)
1951–1954 - 0
Total rounds fired: 1,005,000
1955–1965 - 1
Total rounds fired: 5,077,000
1965–1972 - 23
Total rounds fired: 6,154,000
Total - 24
Total rounds fired: 13,086,000
8. In general, low-order in-bore bursts in guns of modern construction have done relatively little damage. This has ranged from internal marring of the bore, sometimes not realized or discovered until later, up to externally-visible bulging or swelling of the barrel. On the other hand, high-order in-bore detonations burst the barrel catastrophically, and blow out high-energy fragments capable of penetrating decks and structures, and causing various further damage and personnel casualties.
9. Explosive D has been in Navy use as a projectile filler since 1907. It is the least sensitive to shock or friction, and has the lowest explosive energy, of any explosive filler in use today. It has successfully withstood the impact of test firings against armor plate in all calibers of projectiles thus tested. Nonetheless, it has figured in a number of premature projectile explosions. Reference (c) includes D-loaded high-order in-bore prematures which occurred as soon after propellant ignition as “at the origin of rifling”, and others in proving ground firings in which the projectile included no fuzes, or inert fuzes. Over-all, causes to which D-loaded projectile in-bore prematures have been attributed (sometimes on the basis of little more than on-the-spot opinion) include “lose detonator firing on setback”; premature fuze action; cracked projectiles; and propellant flame entry via faulty gas check seal in the projectile base. In-bore explosions of projectiles loaded with the more sensitive Composition A-3 have shared similar attributions, and in addition, two other causes which have been reproduced experimentally: faulty ADF threads allowing the ADF to slam to the rear in an oversize fuze cavity under firing set-back, and gaps or porosity in the explosive, discussed in paragraph 10.
10. A series of proving ground firings were done in 1969-70 to investigate premature explosions in A3-loaded 5”/38 and 5”/34 projectiles, over a range of loading densities, when gaps of various dimensions existed between filler and projectile base. The gaps investigated ranged from zero up to 1”. More than 800 rounds were fired, producing 30 high-order and 2 low-order in-bore prematures. Correlation of premature probability was direct with size of gap and inverse with loading density. No comparable tests appear to have been done for Explosive D in any caliber.
11. Several series of proving ground test firings for exploration of projectile prematures have produced a mix of apparently normal projectile flights, out-of-bore prematures at various muzzle distances, and in-bore prematures. Thus, for some failure modes at least, an association is apparent between observed out-of-bore prematures and an impending in-bore. In February 1971 NEWPORT NEWS had three 8” out-of-bore prematures, fired with reduced propellant charges. Dahlgren chamber pressure-vs-time tests revealed high and erratic 8” propellant pressure peaks with similar reduced charges, and established a high correlation between this propellant abnormality and premature functioning of the BDF (which included a .01 sec firing delay, compatible with the observed NEWPORT NEWS out-of-bore detonation distances). Subsequent NEWPORT NEWS 8” firings were full charge, pending correction of the reduced-charge abnormality.
12. Despite recent emphasis on the need for corrective actions for 5” shipboard prematures attributed to defective sealing of the projectile base, 5” projectiles with missing or gas check seals are still found in the ammunition supply system. As one example, reference (d) reported 21 defective and 2 missing seals found in a single group of some 4300 5”/38 projectiles. The circumstances involved for that caliber also have a direct bearing upon the quality of 5”/54, 6” and 8” projectiles with respect to user safety.
13. Instructions with regard to back-up inspections of projectile base sealing by fleet users appear to have been inadequate. Current practice appears to have been for ammunition depots to paint the entire projectile base, which makes it impossible for users to see certain documented base seal defects, because they are not permitted to remove the paint. Further, NAVORDSYSCOM instructions on the subject, reference (e), are not readily available, do not appear to have directive authority in the fleets, and are ambiguously worded. They are subject only for projectiles obviously damaged in handling. (The latter interpretation existed in NEWPORT NEWS.) The Type Commanders’ instructions, reference (f), call for user inspection only for 5” and 6”. Standard publications such as OP 4 and OP 3347, references (g) and (h), appear to be silent on this subject.
14. Reference (i) directed a comprehensive (and expensive) program of re-certification inspections of 8” projectiles, including full-length X-ray. This should assure that projectiles which have passed through this process are as safe as possible with present hardware, subject only to the reservation that those inspections are not able to provide added assurance against a possible BDF hazard, discussed later. The inspection also yields statistical data on existing discrepancies.
15. As of 16 November, reference (j), results of the above inspections confirm that there is indeed cause for substantial concern. With some 14,700 8” HC projectiles inspected, rejections have run nearly 60%. Many of the rejections are for discrepancies which may be determined, in due course, to be of no consequence with respect to bore safety. But by user safety standards, a grossly unsatisfactory number are for defects with respect to design safety features which were incorporated for good cause.
16. Of particular significance are the rejections for defects directly related to the causes (paragraph 9) to which previous in-bore prematures have been attributed. Specifically:
a. Eighteen are for defects in ADF safety features designed to prevent premature functioning. Six of these rejections are for ADF rotors found in the armed position.
b. 621 are for defective gas check seals in the projectile base.
c. 1686 are for excessive gaps between explosive filler and projectile base plug. The maximum allowable gap, .125”, was that found by firing tests to be safe for A3-loaded 5” (paragraph 12), but irregular gaps up to at least 1.5” maximum depth have been found during this re-certification program.
17. The fuze involved in this casualty was designed in 1944 by the Carnegie Institute of Technology: NOL White Oak is now the design agent for it. It was manufactured and delivered in 1968 or 1969 on a contract let by the Ship’s Parts Control Center (NAVSUPSYSCOM) with the Bermite Powder Company. It was accepted after inspection by the Defense Contract Administration Services (Defense Supply Agency), in accordance with a basic specification published in 1955 by the Bureau of Ordnance. COMNAVORDSYSCOM now controls that specification, but has command control over none of the other agencies here named; this contrasts notably with the direct command control formerly exercised by the Chief of the Bureau of Ordnance over all Naval gun ammunition technical matters.
NEWPORT NEWS OPERATIONS
18. In this deployment NEWPORT NEWS had been in Southeast Asia since May 1972, and before the casualty had fired 24,161 rounds of 8”. This had required two complete relinings of the 8” battery. Under the conditions of firing encountered she had experienced 1100-1200 equivalent service rounds (ESR) 8” barrel life, and had fired some 730 ESR from the gun here concerned since its most recent relining.
19. The ship was numerically well manned. Her 8” firing in this deployment alone was the equivalent of on the order of a half century of normal peacetime 8” experience for a single ship. The rank and rating structure was somewhat below the seniority levels which might be desired, and commissioned turret officers were not assigned. But in our judgment the personnel situation was not an adverse factor in this casualty.
20. NEWPORT NEWS routinely received replenishment 8” HC projectiles with steel nose plugs in place, to be replaced by ship’s force with separately-delivered PDF as required. (This has been the practice since the introduction of bombardment PD fuzes in World War II.) Removal of the steel plug could loosen the adapter into which it is threaded, and which in turn is threaded into the projectile body; and loosening the adapter would disturb the ADF in its cavity, which is regarded as undesirable. The ship’s force had been instructed and cautioned accordingly, and provided with a tool for holding the adapter against unscrewing as the nose plug is backed out. Some comment in the ship indicated dissatisfaction with that tool, so that if any difficulty was encountered in backing out the nose plug, it was left in place and the projectile was used in that configuration. Thus, two projectiles with steel nose plugs were found in the center gun hoist after the casualty, and it is apparent that the projectile involved in the casualty could have been fitted with either PDF or steel nose plug. We do not consider that this alternative is significant to the casualty.
21. During the night of 30 September-1 October (local time, zone minus eight) she was firing harassment and interdiction missions against pre-designated shore target areas in Quang Tri province, Republic of Vietnam. She was in condition THREE, which we consider an appropriate condition of readiness for the operational situation. She had fired 8” high capacity, full charge, rounds as follows:
Date - Time
Gun & Rounds
Start Stop Right-Center-Left
Time: 0923 - 0924
10- right gun
10- left gun
Time: 0926 - 0928
10- right gun
10- left gun
Time: 0930 - 0933
10- right gun
10- left gun
Time: 2350 - 2354
5- left gun
Time: 2355 - 2359
10- right gun
10- Center gun
Time: 0000 - 0005
5- center gun
Time: 0006 - 0007
2- center gun
3- left gun
0008 - 0010
5- right gun
Time: 0011 - 0012
5- right gun
Time: 0013 - 0016
5- left gun
Time: 0017 - 0020
5- right gun
Time: 0055 - 0100
4- center gun
22. The series of 4 rounds in the 0055-0100 time period were being fired from the after plotting room in slow salvo fire. In this procedure in NEWPORT NEWS, plot controls the firing circuit, but in all other respects the gun is in full automatic and, so long as the gun captain holds the rammer control in the ram position (which we understand was the procedure in effect), reloads itself automatically after each shot. After it is reloaded and ready, when plot wishes to fire they sound the standard two salvo warning buzzers, followed by a salvo buzzer simultaneously with closure of the firing key. For the last (casualty) shot, gun elevation was 4120 (35 deg) gun train 067 deg relative. The shot was fired by the procedure described, and the shock of firing was felt in plot in normal relationship to plot’s firing actions. Come observers thought they felt a second, more muffled, shock; this could have been from the additional cartridges which went off as described later.
23. All men in the turret and the upper and lower shell decks apparently died instantly at their stations. One lower shell deck man, apparently transiting the lower handling room at the moment of the casualty, escaped the turret but died outside, of chemical pneumonitis from fume inhalation, as did two others from the lower handling room. One man not part of the turret crew died from the same cause, bringing total deaths to 20. Many others, including officers and men from various below-deck stations outside the turret, were given medical treatment for various degrees of exposure to fume inhalation.
24. Principal material effects of the explosion included the following:
a. The gun liner bulged and ruptured longitudinally for a length of 6 to 7 feet, but was not severed. The principal rupture was on top (location of the liner keyway). The bulge, which opened a gap in this rupture to a maximum width of about 6”, was centered over the seated position of the projectile, about 10” forward of its base. Several additional longitudinal cracks and ruptures were found when the liner was sectioned at Dahlgren. Gases venting through the top rupture blew partly outside but primarily inside the turret face plate.
b. The tube was shattered throughout its circumference, with radial cracks both transverse and longitudinal, such that for a length of perhaps three feet the tube was blown away in large chunks. This area was approximately centered over the area of the liner bulge and rupture. The forward portion of the tube was blown some unknown distance up the liner (elevated to 35 deg.), then fell back, trapping some tube fragments below it; two large tube fragments were caught in the empty case net on the front on the turret. Ten large tube fragments were eventually recovered. Initially, the tube fracture surfaces did not show the “hoop spall” failures considered characteristic of high-order detonations. After damage clearance had progressed somewhat, parts of the tube were recovered which appeared to show hoop spall, and this was confirmed when the parts were sectioned at Dahlgren.
c. The main body of the projectile remained in the bore, with nose fittings blown away, the ogive blown out radially to bore diameter, the base blown off in the vicinity of the rotating band score, and the band blown forward and wedged around the projectile body. The latter’s location was 33” forward of the seated position, with about the rearward 8” of its remaining length being visible through the liner rupture.
d. The forward extension of the slide was broken off and blown 6-8 feet up the elevated tube and liner, before falling back.
e. Internal chamber pressure set the closed and locked breech back sufficiently for the cartridge case to rupture at the breech face, venting into the gun chamber.
f. The slide ruptured, and may have spread the trunnions: this may have also done some distortion of the girder structure which supports the deck lugs.
g. The entire breech mechanism, recoil and counter recoil system, and loading mechanism of the center gun appears to have been extensively damaged.
h. The explosion venting into the turret ignited a total of nine additional cartridges of the 21 in the powder hoists of all three guns, all of which were filled. The upper five charges in the left hoist burned, blowing wide open the hoist casings between deck and overhead in both the upper and lower shell decks; the two lowest charges did not burn. The upper two charges for the center gun burned, blowing open the hoist casing abreast them; the five cartridges below these two did not burn. The top charge for the right gun did not burn, though it was badly crushed and wedged in the loading cradle; the next two charges burned, with effects on the hoist similar to those previously described; and the four lowest in that hoist did not burn. Of the total of 1680 pounds of powder in the three hoist, 720 pounds were thus burned. We were able to recognize no specific mechanism by which any particular charge was ignited, or by which those which did not do so were spared.
i. There was considerable additional general damage in the turret and the two shell decks immediately below it. Much of the basic turret structure, including the roller path, training rack and pinion, main gun girder, right and left guns, and rotating structure in general, appeared likely to be either undamaged or to require only relatively minor repair. An initial, very preliminary, estimate of repair costs is $4 million.
25. The recovered tube fragments, liner, and projectile carcass were returned to Dahlgren for sectioning and analysis in comparison with corresponding material from experimental 8” HC detonations. Principal findings include the following:
a. Footprints in the bore from the projectile band score establish projectile location at the instant of detonation as approximately 0.1” forward of the estimated seating position.
b. Footprints in the bore from the expanded nose of the projectile conform to the above location, are similar in appearance to those of an experimental nose-initiated high-order detonation, and are recognizably different from those for other types of experimental detonations.
c. Location of the largest part of the liner bulge relative to the position of projectile detonation conforms to experimental nose initiation, and differs from that for other types of experimental initiations.
d. Metallurgy and fracture patterns seen in longitudinal sections of projectile and tube fragments correspond with conditions resulting from an experimental nose-initiated high order detonation, and differ from those for other types of experimental detonations.
26. The evidence noted in the preceding paragraph appears to establish that the casualty resulted from a nose-initiated high order detonation, which occurred within a fraction of a millisecond after propellant ignition. (This supersedes our initial judgment, and that of other experienced observers at Subic, that the detonation was low order, about .007 seconds after detonation, hence most probably caused by a defective projectile base seal.)
27. All experience with D-loaded projectiles seems to indicate that such a detonation, within that reaction time, could result only from initiation by the tetryl booster of the nose ADF.
28. In our judgment, the only realistic explanation as to how the latter could have occurred must include the requirement that the fuze rotor (which carries the primer) was in the armed position before the gun was fired. This is because we understand that tests have demonstrated that accidental firing of the ADF primer, by whatever cause, will not fire the booster unless the rotor is armed; and we visualize no mechanism for firing the ADF booster, within the reaction time of this casualty, except by firing the ADF primer. The possibility that the rotor was in fact armed as here postulated is supported by the evidence note in paragraph 16.
29. Various mechanisms for early premature firing of the primer in a pre-armed fuze rotor can be visualized. The first thing that happens after propellant ignition in this gun is that the heavy plastic front plug of the cartridge is blown forward and impacts the projectile base with a substantial shock. The shock wave from this impact, transmitted to the projectile nose and reflected, may subject the ADF to significant accelerations, positive followed by negative. A primer capsule improperly seated in the rotor, or improperly fixed in its seat, or a primer mixture not firmly held in its capsule, could be fired by the resultant set-forward against an unarmed firing pin. Or other assembly or dimensional errors could produce similar results from this initial cartridge plug shock; or if not from it, then from other transients in the initial projectile acceleration. In any case, we understand that tests have indeed confirmed that the primer in the Mk 54/Mk 55-type ADF can fire upon shock of the gun firing, if in the rotor-armed condition, in an otherwise-normal fuze. The timing of NEWPORT NEWS projectile detonation would be consistent with initial shock phenomena.
Further Fuze Implications
30. The discovery of armed fuze components in service projectiles (paragraph 16) emphasizes the question of how such a situation could be permitted to exist. Navy standards enforced for half a century required total, absolute, and unvarying adherence to the most rigorous standards of explosive safety, reflected among other things in ordnance safety precautions long ago described as having been written in blood. Somewhere in recent years those standards have been defaulted, and the lives of a turret crew have been added to the cost.
31. Where that malfeasance occurred in this case is not hard to find. The auxiliary detonating fuzes here concerned have been identified by manufacturer, lot number, and date. All thus far discovered with safety defects, as well as all those known to be in NEWPORT NEWS ammunition, were manufactured under two contracts with the Bermite Powder Company in 1968 and 1969, and were “inspected” by the Defense Contract Administration Services (DCAS), specifically DCASD, Van Nuys, in accordance with technical requirements which are the responsibility of COMNAVORDSYSCOM assigned by him to the Naval Ammunition Production Engineering Center (NAPEC), Crane, Indiana.
32. Reference (k) reviews the manufacturer’s quality control and the DCAS-administered inspection processes applicable to the two contracts. Basic government document controlling the quality of the product was reference (l), and the contractor’s was reference (m). These three references document an unsatisfactory situation with respect to safety (amongst other things), all too consistent with what’s being found relative to the ADF by the re-inspections directed by reference (i).
33. Specifically, those references appear to indicate a misplaced confidence in the degree of assurance afforded by a sampling-type approach, including the firing of a number of samples at Dahlgren. Reference (n) suggests that a similar misappreciation may exist at working levels within HQ NAVORDSYSCOM. We reemphasize that the sampling-type approach is normally adequate with respect to features such as those which concern functioning at the target; but it never was, is not now, and can never be acceptable with respect to defects which are potentially lethal to the fleet user – of which in-bore fuze prematures are among several notable examples.
34. We consider the design of the ADF Mk 55-0 (and of the similar Mk 54) to be adequate with respect both to safety and to intended performance at the target. The design could be improved somewhat by changes we have discussed with NAVORDSYSCOM staff. But no changes could protect against certain manufacture/inspection defects such as documented in reference (j). Thus nothing related to possible design improvements of the ADF Mks 54 or 55 can reduce the need for attention to the primary requirement for effective, responsible, and responsive manufacture and inspection.
35. Rather than undertaking design improvements for the ADF Mks 54 or 55 it appears preferable to adopt the Mks 379 (5”/38), 395 (5”/54), and a further variant (if required) for other fleet calibers. These fuzes incorporate an anti-malassembly feature, and also a mechanical arming delay intended to prevent primer firing until after projectile travel several hundred feet beyond the gun muzzle. Regardless, however, of whatever sanguine hopes might attach to the appearance of such fuzes in service (some years hence at the earliest), nothing would eliminate the requirement to return to former Navy standards of meticulous manufacture and assembly, and rigorous, accountable inspection, of all ammunition and components. This must include 100% inspection, at several successive echelons, of all features bearing on user safety.
36. Finally with respect to ADF bore safety, although the Mks 379 and 395 are in procurement, quantities of 5”/38 and 5”/54 projectiles remain in inventory which are fitted with ADF Mk 54 (Mods). The latter ADF is also standard in service 6”/47 HC projectiles. Design-wise this fuze is the same as the Mk 55 except for the difference in arming spin rates. Hence it must be recognized as essentially in the same category as the Mk 55 with respect to potential in-bore prematures resulting from manufacture/inspection derelictions. Thus far its actual safety record seems to have been statistically somewhat better than the Mk 55, possibly due to manufacture by different contractors (KDI and Maxson). But, as a minimum, it would appear that a substantial Mk 54 sampling program by X-ray should be initiated earliest.
37. Certain fuze considerations paralleling those for the ADF also require recognition for the BDF. By design these fuzes, of which the Mk 48-4 in NEWPORT NEWS 8” HC is an example, appear adequately safe against being armed and initiated in bore. (Of course, on the basis of its design, so does the ADF Mk 55). But the in-bore environment is not simple; it can generate substantial transients which momentarily modify or reverse the classic conditions. The NEWPORT NEWS out-of-bore prematures in 1971 (paragraph11) provide an example: if analysis of them is correct, rough propellant performance initiated the BDF while the projectile was somewhere within the bore, and only the .01-second delay pellet in that fuze prevented these initiations from causing one or more high order in-bore projectile detonations. Though the reduce-charge propellant anomalies associated with those events may have been corrected, other well-known factors encountered in service can also give the projectile a rough in-bore ride. BDF manufacture/inspection derelictions comparable to those tabulated in reference (j) for the ADF could result in omission of the delay pellet. Other errors could expose the fuze to more ready premature initiation. In this connection we note that the re-certification program of reference (i) is not capable of providing added assurance concerning the BDF. Hence the only BDF safeguard is positive inspection assurance prior to assembly into projectiles.
Other Potential Causes of Prematures
38. Although it seems clear that this particular casualty resulted from an ADF premature, there is no justification for focusing corrective actions solely on that, to an extent that could cause us to lose sight of the need for substantial improvement with respect to all elements of ammunition safety in all fleet calibers. Paragraph 7 depicts a glaring degradation of fleet ammunition safety since 1965. The rate of in-bore projectile prematures per shot fired since that date has increased by a factor of more than 25 over the rate experienced throughout the preceding nineteen years. That picture is substantiated for 8” ammunition by the appalling rate of deficiencies documented in reference (j). The 23 shipboard in-bore prematures since 1965 have cost 24 lives, apart from their costs in combat readiness and material.
39. Judging from the record noted in paragraph 9 and the findings recorded in reference (j), in addition to ADF malfunctions the other most probable causes of in-bore prematures appear to include BDF malfunction, defective projectile base sealing, and possibly gaps in the explosive filler (though this is uncertain for Explosives D). The BDF situation is discussed in paragraph 37. Our comments on these additional potential causes are as follows:
a. Defective projectile base seals: The record in reference (c) seems clear that these can cause prematures with either Explosive D or Composition A-3 projectile filler; and references (d) and (j) show that such defects continue to be found. They are attributable not to contractors or DCAS, but to manufacturing and inspection deficiencies in our own naval ammunition facilities. There are instructions intended to require back-up inspections of projectile base seals by fleet users, but these need clarification as previously noted, and in any case, they alone cannot assure against all forms of documented sealing defects. Further, provision of ready-to-use ammunition to the fleet should by no means depend upon inspection by fleet users, whose experience and competence in such matters may be uncertain, and who must often replenish ammunition at sea in darkness and under adverse conditions.
b. Gaps in the explosive filler: Principal present suspect is a gap between the filler and either the projectile base or the tip of the BDF. Paragraph 10 summarizes the basic problem, reference (j) shows the situation being found in current 8” HC, and paragraph 16 notes the extent of departures from the prescribed standard which are being identified. Here again, the defects are attributable only to manufacturing and inspection deficiencies in our own Navy facilities. Until some safe gap size for Explosive D is established (which would require extended testing, since negative information is all the results from each non-premature shot), the standards for Composition A-3 should be enforced for Explosive D as well, as is currently being done by reference (i).
40. Two more possible sources of a projectile premature have been suggested. These are noted as follows for the record, but no information yet available to us appears to justify their being regarded as significant to this casualty:
a. Copper azide problem: Reference (o) summarizes this situation. Briefly, copper azide is extremely sensitive to shock, and can be formed internally in any fuze containing lead azide (as does the ADF Mk 55) which also includes exposed copper or copper-alloy elements (which we understand the ADF Mk 55 is not supposed to do). Although the unsatisfactory production/inspection situation uncovered by reference (i) makes almost any defect believable, we judge that copper azide could not form in the Mk 55 in sufficient quantity in the necessary location to initiate the booster of the ADF Mk 55 prematurely, unless the fuze rotor were prearmed. In the latter case the presence of copper azide could become one more of the possible causes of firing a pre-armed rotor, with a resulting projectile premature.
b. Iron picrate problem: There has been some concern that contact within the projectile cavity between Explosive D (ammonium picrate) and unprotected steel could result in the formation of iron picrates in a form highly sensitive to shock, and in quantity sufficient to initiate projectile detonation. The possibility is intended to be protected against by specified cavity coatings. At the direction of COMNAVORDSYSCOM ex-NEWPORT NEWS 8” projectiles are being subjected to breakdown and inspection, including chemical analysis of explosives. We understand no abnormalities have been discovered thus far. We can make no further assessment.
41. There were other variously-postulated causes for this casualty, which we discount, as follows:
a. Improper performance by the turret crew, such as because of inadequate manning or training: We saw no indications of this whatsoever. Specifically, the turret was operating in a fully automatic mode, in which momentary crew performance was no factor; their longer-term performance, such as in maintenance and preparation for firing, had been part of an outstanding accomplishment by the ship; and the standard of material maintenance of the 8” battery as seen in the two undamaged turrets was commendable.
b. Double load or attempted double load of the gun: The breech was closed and locked, so the load had been completed (and had been so signaled to plot); the rammer could not achieve this if a preceding projectile were seated. Further, the special procedures required in order to permit a mistaken attempt to double load could not have been completed within the time available, in the firing sequence recorded. Finally, if a double load had occurred, an unfired cartridge would be at the rear of the case ejection tube; but the case recovered from there appeared to have been fired normally, as distinguished in appearance from the loaded cases which burned elsewhere in the turret as a result of the explosion.
c. Cook-off: The firing sequence for this gun (paragraph 21) shows 20 rounds fired in the preceding 65 minutes, which in turn had been preceded by several hours silence; the round in question had been loaded for not more than something on the order of one minute before the explosion. These conditions are at least an order of magnitude less severe than those considered likely to explode a projectile by cook-off as indicated in OP 1591, reference (p).
d. Obstruction in the bore, such as from a rotating band or primer extension left from the preceding round: Considerable experience has seen these blown out the muzzle with the shot from which they came, particularly the rotating band, which starts ahead of the entire volume of propellant gas. If a detached primer extension were so far to the rear that it didn’t blow out the muzzle, with the gun at 35 deg. elevation (as in this case) it should drop out the breech with case extraction; but this is immaterial to this casualty anyway, because the empty case from the preceding shot was recovered with the primer extension in place in it. Finally, before the days of decoppering agents in the ammunition, many wartime firing sequences went far beyond the numbers of shots known to build up copper obstructions in the bore, so that obviously in those firings the gun just blew such obstructions out the muzzle without damage. Comparable results are also reported from Dahlgren tests done with such objects as loose fuzes lying in the bore ahead of the projectile. In short, we are confident that an obstruction of this nature is the gun bore did not occur in this instance, but that if it had, it would have been unlikely to cause projectile detonation.
e. "Too tight liner”: This idea originated in the ship, apparently on the basis of bore erosion gage readings taken after some tens of shots fired since the most recent relining. The readings were less than the nominal for a new liner (actual readings not having been taken before firings with the new liners), and were thought possibly to indicate a smaller than normal bore diameter at the rifling origin. Such a decrease in gage readings (“negative erosion”) with initial firings has long been known, and denotes no abnormalities. It has been ascribed to a mandrelling action of the projectiles in the liner, which initially displaces the rifling origin longitudinally somewhat to the rear, rather than to any change in radial dimensions. Further, post-casualty star-gaging of the guns in the undamaged turrets revealed no dimensional abnormalities, and there is no reason to believe these differed from the casualty gun.
f. "Oversize projectiles” This thought also originated in the ship, apparently on the basis of a projectile received in a recent replenishment, judged abnormal by visual inspection (for reasons not recorded and no longer retrievable), and returned to the issuing unit. The diameter gaging included in the re-certification results reported in reference (j) lists 46 projectiles that would not pass the bourrelet “go” gage, a cylindrical gage ring of specified internal diameter. But all of these failures appear to be attributed to one or more local oversize areas at the bourrelet, including scars or burrs, which we do not believe would have discernable effects in either loading or firing. In any case, proving ground experience indicates that projectiles with some degree of excess diameter at the bourrelet (point of largest diameter) can be fired without apparent harmful effects: if the relatively low-powered rammer can load the projectile fully into the bore, the propellant charge will move it out the muzzle without further ado.
g. Steel constriction of the bore: This form of bore obstruction has been encountered in the past with some guns, caused by the mandrelling action of the projectiles tending to elongate the liner forward so it rides over a liner shoulder. The post-casualty star-gaging noted in paragraph e. above found no such constriction in the guns of the undamaged turrets, and there is no reason to believe there was one in the casualty gun. In any case, we now know the projectile detonated at a position in the bore such that its bourrelet was some 8” short of the only shoulder that could have caused such a condition.
Technical Control of Gun Ammunition
42. We are concerned over the control of technical features of Navy gun ammunition, particularly as regards ammunition safety for fleet users. Preceding paragraphs have noted the serous degradations which began to show up in the fleet, commencing in 1965. The correlation appears inescapable between them and the organizational changes which had begun not long before, and which bear upon command management and control of gun ammunition technical matters.
43. The chain of that command is now so diffuse that effective control, with clear and direct authority, responsibility, and accountability realistically vested in a single commander, does not appear to exist.
44. Within the Navy Department, the closest we come to such a commander is the Chief of Naval Material. Under his command are four of the key separate entities in ammunition design and procurement: NAVORDSYSCOM, NAVSUPSYSCOM, NOL White Oak, and NWL Dahlgren. But neither he nor anyone else within the Department of the Navy commands the inspectors who pass upon the technical quality of contract-procured components of Navy gun ammunition: they belong to the Defense Supply Agency. Nor does the CNM appear to have directive authority over ammunition technical matters at overseas Naval ammunition facilities, for example NAVMAG Subic: they are command-subordinated in the fleet.
45. But beyond all that, the span of CNM responsibilities is too broad to permit him to function effectively in person as the single commander for ammunition. It would be inevitable that most on-going decisions and actions nominally his as the commander would in fact be taken by staff functionaries – those necessary but faceless subordinates who can share none of the responsibility of command, and who, in the familiar bureaucratic system often know little about that kind of responsibility. It would be one more case of over-centralization to attempt to set up CNM as the single commander for naval gun ammunition technical matters.
46. Throughout World War II and much of the period after it the Chief of the Bureau of Ordnance, and later his successor the Chief of the Bureau of Naval Weapons, had full command and single-manager control of these ammunition matters, the authority and responsibility that went with them, and the staff assistance necessary for effective functioning. Organizational changes which need not be recounted here, including some imposed from the Department of the Navy, have progressively eroded that control (and indeed may be continuing the process at present) – notably, however, without fixing any of the responsibility.
47. COMNAVORDSYSCOM appears to be a suitable, and in our judgment the preferable, commander in whom to reestablish the closest feasible approach to the authority, responsibility, and accountability with respect to gun ammunition technical matters, which formerly reside in the Chief of the Bureau of Ordnance: and to whom to furnish the appropriate staff and money support. To accomplish these changes will present familiar administrative problems in predictable offices, including some outside the Department of the Navy. In our judgment, however, the unique technical problems inherent in gun ammunition, and the gravity and lethality of the results of anything less then the best control, require that those problems be faced and surmounted.
48. Then filled, NEWPORT NEWS 8” powder hoists constitute vulnerable and dangerous paths for the propagation of high-energy flame direct from the gun chamber to the powder handling room, at magazine level below the armor deck. Nearly a ton of powder could become involved in the hoists of each turret. The initiating flame could originate either with a casualty, as in this instance, or by enemy action. There is an automatically-actuated sprinkling system in these turrets which includes branches inside the powder hoists. Apparently the system actuated as intended in this casualty. But such systems do not have the capacity to extinguish powder fires once started, and because of both limited capacity and relatively slow response times, we would not judge that the hoist sprinklers were or could be factors in preventing flame propagation in the hoists under circumstances such as found in this case. The loading scuttles at the lower ends of the hoists would provide no protection from further flame propagation if the hoists themselves ruptured within the handling room, as they did in this casualty in the spaces above the handling room.
49. Thus one can only speculate on how close this casualty came to extending into a magazine. If it did, survival of the ship itself would be speculative. The powder hoists should be modified to incorporate at least one flame barrier in each hoist, perhaps at the pan plate level. As a minimum safety factor this barrier should open automatically for each hoist movement, but remain closed during the hoist’s stationary intervals.
50. Similar flame propagation paths also exist in the powder hoists of the 5”/54 rapid fire mounts and 6”/47 turrets. Corresponding safety features should be added to those hoists.
51. Paragraph 22 describes the slow salvo fire procedure in use in NEWPORT NEWS. That procedure was not a factor in this casualty, but nevertheless we consider it should be modified as here described. We consider that in slow salvo fire the gun controls should be so disposed that after each shot the gun would be left with breech open; successive loads for each shot would then be made only on control’s command, given separately for each shot, at an appropriate interval before each intention to fire a salvo. This would improve gun cooling between rounds because of the open bore; it would reduce the likelihood of cook-offs due to unexpected interruptions of firing; and it would enhance fire discipline in general.
52. Comparable procedure considerations concerning standing by with breech open between shots in slow salvo fire apply to all other naval guns for which this is not already the doctrine.
53. A commissioned turret officer was not in turret two at the time of the casualty, and apparently in NEWPORT NEWS they are in general not assigned. This too was not explicitly a factor in this casualty. But nevertheless we consider that any 6”/47 or larger turret should have a qualified commissioned turret officer regularly assigned, and that he should be in the turret during all occasions of expected firing. This was a Navy standard for years, and we see no justification for degrading it. His presence, judgment, and decision should surely ameliorate many a difficult situation, in many ways, some perhaps imponderable. Or else he’s not worthy of his uniform. The assignment would provide the officer an additional opportunity to develop and demonstrate naval leadership. And if his men are to die in a casualty, he dies with them – in the tradition of a Navy that fights and wins.
54. Throughout our inquiry we have been impressed by the thoroughness, vigor, and initiative of COMNAVORDSYSCOM in addressing every element of the ammunition safety problem which is within his authority; and by his effectiveness in gaining the support and assistance of other technical agencies also involved in the problem, but not under his command and control.
55. We have been equally impressed by the comprehensive approach of COMSERVGRU THREE to the formal investigation of the casualty, which was a principal factor in the recovery of substantial information and material evidence which otherwise would not have been available.
56. This casualty resulted from a high order projectile detonation which occurred within a fraction of a millisecond after propellant charge ignition in normal gun firing sequence. The detonation was caused by premature firing of the ADF Mk 55-0, which in turn resulted from the fuze rotor being in the armed position at the instant of gun firing.
57. The ADF concerned was manufactured by Bermite Powder Company in 1968 or 1969. Its design is adequate. Its premature functioning reflects faulty manufacture, including manufacturer’s quality control, and ineffective government inspection by DCAS, in response to inadequate criteria specified by NAVORDSYSCOM.
58. The ADF Mk 54 (Mods), in service in 5”/38, 5”/54 and 6”/47 projectiles, is similar in design to the Mk 55, and vulnerable to the same lethal defects which can result from manufacturing errors or inspection inadequacies.
59. Various additional manufacturing/inspection errors currently encountered are also potential causes of premature projectile detonations in bore, of which the most significant appear to be BDF malfunction, defective base sealing of projectiles, and gaps in projectile explosive filler. The NEWPORT NEWS 8” ADF casualty must not divert attention from the need for corrective actions with respect to all other safety hazards, in all fleet ammunition calibers.
60. Some improvement in projectile bore safety should result from adoption of the ADF Mks 379/395 design in all fleet calibers. But by itself this cannot provide adequate safeguard against manufacturing errors or inspection defects.
61. In early due course, the size of gap safety allowable between projectile base or BDF tip and the explosive filler should be verified by an experimental firing program for Explosive D.
62. An automatic flame barrier, as a minimum to be closed except during hoist motion, should be added to NEWPORT NEWS – class 8” powder hoists. Similar barriers should also be incorporated in the powder hoists of 6”/47 turret and 5”/54 rapid fire mounts.
63. Gun ammunition quality with respect to user safety has deteriorated markedly in recent years. The timing correlates with organizational changes which have degraded defective control over ammunition technical matters.
64. Excessive gun ammunition accident rates will continue until effective command control and management is restored.
65. At recent rates of Southeast Asia gun ammunition expenditure, the next in-bore projectile premature can be expected statistically at any moment.
66. Fleet gunnery doctrines for slow salvo fire should provide that gun remain unloaded, breech open, until loaded on command just before the intended time of firing each salvo.
67. Qualified commissioned turret officers should be assigned in all turrets 6”/47 and larger, to be on station during any expected firing.
68. We recommend that the Chief of Naval Material:
a. Take all actions within his authority to establish COMNAVORDSYSCOM as the single commander having responsibility, authority, and accountability for all technical matters concerning naval gun ammunition, including design, test, procurement, and life-cycle technical control. Among other actions these should include return of NOL White Oak and NWL Dahlgren to COMNAVORDSYSCOM command; return of ammunition procurement responsibility from COMNAVSUPSYSCOM to COMNAVORDSYSCOM; and so far as feasible pending other actions noted in subparagraph b. below, assignment to COMNAVORDSYSCOM of the inspection responsibility for contractor-procured gun ammunition components, with NAVORDSYSCOM inspection capabilities expanded accordingly.
b. Initiate actions in appropriate Navy Department and OSD channels to return from DCAS to COMNAVORDSYSCOM the inspection responsibility for contractor-procured naval gun ammunition components, and to adjust staff and money to correspond.
c. Initiate actions in appropriate Navy Department and fleet channels to establish technical control and direct liaison authority for COMNAVORDSYSCOM, with respect to all technical gun ammunition matters, for ammunition and ammunition facilities which are under fleet control.
69. We recommend that COMNAVORDSYSCOM take or continue the following actions (of which we understand those in subparagraphs a, b, d, e, f, g, h, and l are already in process):
a. Review, and modify as appropriate, his organization, to see that it gives him the direct staff support he needs to carry out his responsibilities for providing safe and effective gun ammunition to the fleet. This recommendation applies with respect both to his present responsibilities, and to the broader span reflected in paragraph 68 above.
b. Review, clarify, and strengthen as appropriate, all contracts and similar documents, assembly, and inspection of ammunition and components, with particular attention to all aspects involving safety for the fleet user. The latter aspects should include, in particular, fuze safety features, projectile base sealing, and explosive filler gaps. Ammunition-related agencies whose activities are to be covered by the documents here concerned include facilities under the command of COMNAVORDSYSCOM, as well as those now under other DOD agencies, and civilian contractors.
c. In connection with subparagraph b. above, establish requirements for 100% inspection, at several successive phases of manufacture and assembly, of all ammunition features bearing on user safety; and further establish a planned follow-up on defects found, to determine accountability.
d. Pending shift from DCAS to COMNAVORDSYSCOM of the responsibility for inspection of contractor-procured naval gun ammunition and components per paragraph 68, establish the interim requirement for 100% back-up inspection by X-ray, of all ADF Mks 54 and 55, and all BDF, to be done at naval ordnance facilities before the fuzes are assembled into projectiles.
e. Consider adding to prescribed inspection procedures, regardless of by whom done, X-ray inspection of all fully-assembled ADF Mks 54 and 55, and all BDF, for 3” and larger projectiles, with identification and appropriate retention of the X-ray photography for record. If established, this could supplant the back-up X-ray inspections recommended in paragraph d. above.
f. Continue the 8” projectile re-inspection program directed by reference (i) until the entire inventory has been certified. Priority of effort on this program could be readjusted when NEWPORT NEWS leaves Southeast Asia.
g. Initiate earliest an inspection program including X-ray, in extension of the provisions of subparagraph d. and e. above, for present stocks of ADF Mk 54 (Mods).
h. Accelerate all feasible the fitting of ADF Mks 379/395 in lieu of ADF Mk54 in service 5” ammunition.
i. Adapt and fit the ADF Mk 379/395 design to 6”/47 and larger service ammunition.
j. Adapt to any further fuze procurement, including the Mk 379/395 types and other detonating fuzes as appropriate, all applicable improvements with respect to safety which have emerged from investigations related to NEWPORT NEWS. An example is the complete encapsulation of the fuze primer mixture within a metal shell, shaped to permit assembly only in the correct direction.
k. Conduct a firing program to determine for Explosive D the size of gaps in projectile filler which are safely allowable.
l. Develop and install automatic flame barriers for the powder hoists in NEWPORT NEWS – class 8” turrets, all 6”/47 turrets, and all 5”/54 RF mounts.
70. And finally, we recommend that appropriate levels of fleet command:
a. Provide for the assignment of qualified commissioned turret officers to all turrets 6”/47 and larger, these officers normally to be present during any expected occasions of firing the turret.
b. Modify type gunnery doctrine as discussed herein with respect so slow salvo fire, for all classes of gun armaments, for use in situations in which a high rate of fire is not a consideration, and where periods of firing may be protracted.
c. Modify reference (f) with respect to fleet-user inspection of projectiles, to specify clearly the requirement for appropriate and feasible inspections.
(a) Conference 2 Oct 1972 in HQ NAVORDSYSCOM, among ADM Kidd (CNM), VADMs Masterson and Mustin, RADM Sappington (COMNAVORDSYSCOM), and others
(b) NAVORDSYSCOM Invitational Travel Orders Numbers T-3060/73 and T-3061/73, 2 Oct 1972
(c) NAVORD TR 71-1, 1 Jan 1971 “History of Navy Use of Comp A-3 and Explosive D in Projectiles”, R. L. Beauregard
(d) NAV MAG Subic msg 121130Z Aug 72
(e) COMNAVORDSYSCOM msg 222116Z Jan 71 as mod. by COMNAVORDSYSCOM msg 052304Z Feb 71
(f) COMCRUDESLANT/PAC Gunnery Notes, para 619
(g) OP 4 – Instructions for Ammunition Afloat
(h) Op 3347 – Ordnance Safety Precautions
(i) COMNAVORDSYSCOM msg 132347Z Oct 72, as superseded by msg 222343Z Oct 72
(j) NAVORDSYSCOM memo ORD-0442B/EJS of 16 Nov 1972
(k) NSMSES Por Hue (0700) Datatel transmission 001, 1 Nov 72
(l) Military Specification MIL-F-18671 (Nord) dated 25 May 1955
(m) Bermite Powder Company Quality Assurance Plan PQAP 67-10 dated 13 Oct 67
(n) NAVORD Point Paper ORD-044, Subject: Conventional Ammunition Quality, 16 Oct 72
(o) NOL White Oak ltr Ser 878 of 9 Feb 72
(p) Op1591 “Clearing of Live Ammunition from Guns”, 2nd Rev. 15 Jan 72
This is the end of the Detailed Report