Turret I fired its first round of split salvos, using 2,700-lb dummy (BL&P) shells, six silk powder bags filled with D845 propellant (reduced charge), and a primer from the right and center gun, in accordance with established procedures. The center and right gun fired normally, but the left gun misfired. A second round of shells and powder was loaded and fired again from the right and center guns. During this second round, the left gun misfired again. It was decided to treat this misfire as a hangfire-that is, the breech door was to be left closed for two hours, even though regulations required only a minimum of thirty minutes. When this breech door was opened, it was discovered that a large hole had been burned into the last powder bag from the seven primers that had been fired into it. The failure of the charges to I ignite was not particularly unusual for bag-loaded guns, especially when firing with reduced I charges. This incident had no bearing on the tragedy that was about to befall Turret II.
Orders were given to Turret II to commence firing. The crew of this turret prepared to load all three guns with dummy shells and D846 propellant, unlike the type D845 propellant used in Turret I. The D846 propellant, normally restricted to service 1,900-lb High-Capacity (HC) shells, was being used as part of a continuing developmental program to control and perfect the velocity and long-range accuracy of the 16-in guns. Official prohibitions against using the D846 propellant were based on the desire to limit high pressures in the guns, although reduced charges had been fired in these and other battleship guns during World War II. This special five-bag charge (the guns had a maximum capacity of six bags) resulted in a maximum pressure of about 38,700 psi, well below the normal service pressure of about 49,700 psi, and much less than half the theoretical strength limit of about 90,000 psi. The D846 charges would not create an inherently unsafe condition for firing this gun.
The breeches for the two outer guns were closed and locked within 17 and 44 seconds of the order to load and then elevated to their firing positions. Their powder hoists were lowered to the powder flat. GMG3 Robert Backherms was the rammerman for the center gun and was new to his position. GMG3 Richard Lawrence was the cradleman, and GMG2 Clayton M. Hartwig was acting as the gun captain for the center gun in place of Lawrence. The projectile for the center gun was rammed into the rifling, and three powder bags dropped into the loading tray. In accordance with normal practice, Hartwig should have inserted a small lead pouch behind powder bag number 1 and then shoved powder bags 1 and 2 into the breech, allowing room for two more powder bags to fall into the loading tray. Hartwig might have placed another small pouch behind powder bag number 2 before the last three bags were pushed into the gun by the ramming mechanism. These two lead pouches were to "de-copper" the guns. (Note 1) It is unlikely that Hartwig would have delayed the gun-firing process had he missed placing one of these pouches. Rammerman Backherms reached for the lever to activate the rammer that would push the five powder bags into the center gun. His actions sent the powder bags at a low ramming speed some 24 inches farther into the gun chamber than prescribed, so the last bag was not at the correct distance to the firing charge in the breech door. This spacing, well beyond the maximum of 4 inches, was undesirable because it can contribute to inconsistent ballistic performance or to a misfire. The last bag should have been positioned so that the mushroom of the plug was just touching the ignition bag of the last charge when the breech was closed.
The following conversation ensued in Turret II that morning:
GMCS(SW) Reginald O. Ziegler (turret captain)-"Left gun is loaded. Good job! Center gun is having a little trouble. We'll straighten that out."GMG3 Richard Lawrence Center gun cradleman -"I have a problem here; I am not ready yet."
Ziegler, now shouting to LTJG Robert M. Buch-"Tell plot we are not ready yet. There is a problem in the center gun!"
GMG3 Richard Lawrence (with annoyance) - "I'm not ready yet! I am not ready yet!"
Unidentified Seconds laterl- "Oh my . . ."
It should be noted that Lawrence had signed off as the captain of the center gun, and Hartwig had acted as the gun captain in his place. The gun captain had complete control of the loading and ramming of a shell and powder into a gun. Hartwig was believed to have been hunched over the breech door, presumably to investigate what happened. A few seconds later, a voice stated over the turret's intercom circuit, "Oh my . . ." The breech of the center gun had not been closed, and the rammer was still in the barrel when smoke, unspent powder grains, flame, and hot gases burst out of the open breech. Hot gases swept through the lower turret substructure and erupted through the three gun ports, the vent ducts, and the rangefinder hoods (see Figures 7-1 and 7-2). The gun bloomers (Note 2) were ripped from the turret's faceplate and blown away from all three gun ports. Thick, hot, gray smoke billowed forth, scorching the teak deck beneath Turret II, which was pointing to starboard. The antiflash seals in the scuttles to the powder magazines prevented a much greater calamity, a catastrophic explosion that likely would have destroyed the Iowa. Eleven crew members in the lower magazines, outside the rotating turret substructure, survived the holocaust, which killed forty-seven of their shipmates in the inferno above and within the turret substructure. One of the crew in the magazines of Turret II turned on the sprinkling systems in the magazines, and seven to eight minutes after the powder fire occurred in the turret, Captain Moosally ordered the magazines flooded. The powder and BL&P shells in the right and left guns were not affected and were later removed by fire-fighting teams. Powder for the next salvo had already been placed in the turret's lower revolving structure, according to procedure. Working its way into the lower substructure of the turret, the fire scorched some of these powder bags, and some caught fire. The flooding of the magazines extinguished these fires.
Secondary explosions briefly hampered the fire-fighting efforts, which succeeded in extinguishing the last of the fires within ninety minutes of the initial explosion. The fires and explosions consumed some twenty-five of the forty-five bags of propelling charges stored on the deck in the powder flat, in addition to two or three of the five bags that had been loaded into the center gun. (Note 3)
The force of the deflagration was so intense that it propelled the 2,700-lb BLIP shell some 44 inches into the rifling of the gun, where it stalled because the energy of the burning propellant had exited into the turret {see Figure 7-3). The center gun's rammer chain first began to move backward in its housing some 23 inches and then collapsed from the forces generated by the burning powder bags. The fifteen sections of the rammer were propelled backwards out of the gun and toward the turret officer's booth. The Mk 3 computer was destroyed, the optical rangefinder was wrecked, and much of the turret's interior structure and equipment was damaged or destroyed.
The Commander, Naval Surface Force, Atlantic at once appointed Rear Admiral Richard Milligan, who had been the New Jersey's first skipper upon her reactivation in 1980, to assemble an investigation team. Captain Joseph D. Miceli of NAVSEA formed a technical team that would collect the evidence, interview the survivors, and conduct special tests. These teams were unable to identify an accidental cause for an open-breech, cold-barrel explosion within the 16-in gun. Past turret disasters in U.S. battleships (see appendix C) had occurred in guns where previous rounds had been fired. The ength of rammer chain damaged by the deflagration indicated that an overram of 21-22 inches had taken place. The chain retracted into a semicircular housing that protects the chain from gun blast. It was assumed that the length of damaged chain outside the housing was the distance that the rammer had moved forward after the overram. Therefore, press reports, and Admiral Milligan in his briefing, alluded to a 21-inch overram The final analysis corrected the overram to 24 inches!. Extensive experimentation and analysis of Navy records from the battleship era revealed the remarkable insensitivity of bagged propelling charges to all forms of abuse. Despite exhaus-tive analysis of the physical evidence and post-accident experimentation through September 1989, the Navy was unable to exactly duplicate the accidental explosion that was believed to have occurred in the Iowa's turret. It is important to emphasize in these trials that the D846 propellant and black powder were also tested to determine their ignition properties. For example, a cigarette lighter required more than nine minutes to ignite the black powder through the quilted patch on the powder bags. The powder grains that make up the powder bags took 2.5-3.5 minutes to ignite, depending on whether they were inside a polyethylene wear-reducing jacket or the plain silk material. It proved impossible to ignite the powder bags by ramming or dropping them from heights of 40 and 100 ft. The powder grains were also insensitive to electromagnetic radiation. Post-incident analyses of the charges in the lowa's powder magazines confirmed that they were in a safe, stable condition.
A primer {similar to a 0.30-caliber blank rifle cartridge), fired from the breech block, normally ignited a bagged propelling charge. The gun was designed to prevent the primer from being fired until the breech block was closed and locked. Each bagged charge had an ignition pad containing black powder sewn onto its base and was quilted to spread the black powder evenly so that there would be virtually instantaneous Within milliseconds ignition. Each bag ignited the next in sequence, that is, it burned from the end much like a cigarette, creating a pressure that would propel the shell out of the gun. A test was done to determine if a preignition of the primer could have initiated the explosion, but preignition was proved to have been unlikely with an open breech block.
Tests were done at Dahlgren in 1989 on a 16-in/50 Mk 7 gun, using five bags of D846 propellant and the techniques employed in the lowas to load and fire these guns. Overrams of the type that had occurred in the lowa's turret were replicated without incident.(Note 4) Systematic tests were conducted: charges were initiated at each bag location, and measurements were taken of the pressures and of the movement of the projectile up the barrel. These tests determined that the point of ignition that caused the projectile travel in the center gun was most likely located between bags 1 and 2 or bags 2 and 3. Finally, the investigators decided to remove the rammer and its chain from the center gun of Turret II and take them to Dahlgren, where they were reassembled in a fixture similar to a 16-in gun. As the rammer and its mechanism were taken apart, each part was closely examined for the slightest evidence of failure; nothing indicating failure was found. In May 1989 a test was done with a timing device similar in size to the lead pouch and placed between bags 2 and 3, and the investigators found that this arrangement closely replicated what had occurred on board the Iowa.
The Navy Department and Congress exerted great pressure on the investigative teams to determine a cause for this disaster and the loss of forty-seven lives. After all possible accidental causes had been ruled out, the investigative officer concluded that sabotage was the most probable explanation for the tragedy. Minute traces of residue trapped in the projectile's rotating band (see Figure 7-4) when the shell had been driven into the gun tube, provided strong evidence of sabotage-steel wool, calcium hypochlorite, and glycol were found. Thorough experimentation confirmed that a plastic bag containing a glass tube filled with these everyday materials, positioned between the first and second bags, and then rammed into the breech, could have caused approximately what had occurred in the Iowa's gun. This type of sabotage would have explained the location of the projectile in the gun tube and the residue found trapped in the rotating band of the shell. Tests were conducted with a 16-in gun and a device that had the same chemical properties as those of the residue found in the rotating band of the shell. These tests indicated that it was possible for such a device to trigger an explosion such as the one that occurred in the Iowa.
The findings of Admiral Milligan's team, read before a press briefing in September 1989, were greeted with dismay and disbelief. The identification of one crewman IClayton Hartwigl as the perpetrator of the incident seemed to be the best explanation of the cause since he was acting gun captain. Admiral Milligan commented to the press, "There's an assumption you have to make. He was the gun captain and controlled the loading of that gun." Senator Alan Dixon called the Navy's findings "guesswork."
Chairman Sam Nunn of the Senate Armed Services Committee was also skeptical and requested Sandia National Laboratories at Kirtland Air Force Base in Albuquerque, New Mex-ico, to go over the evidence and provide a second opinion. Sandia conducted thorough tests of its own. The distance the rammer head had traveled when it was activated to shove the bags of D846 propellant into the gun was important to Sandia's investigation. In February 1990 Dr. Karl Schueler visited the damaged Turret II of the Iowa, which was undergoing repairs at the Norfolk Naval Shipyard. He made careful measurements of the spanning tray, which had gouges from the rammer chain imprinted in it. He determined that the overram had been 24 inches, a distance that indicated a severe compression of the powder charge during the fateful ram. Sandia's chemical analysis did not disprove the Navy's findings on the chemicals' presence, but did find the same materials in different concentrations on other battleships-the Wisconsin and Missouri. The investigative team, headed by Dr. Richard Schwoebel, decided to rig a half-scale drop test within a cylindrical chamber that would duplicate, in miniature, the fateful ram. The team determined that with sufficient velocity, the propellant could ignite-but this was only a half-scale charge under very high speed. Could the powder ignite under the high rammer speed of 14 fps?
The investigation shifted to the powder bags, with their eight layers of 225 pellets per layer, plus a tare (or trim) layer, which was used to compensate for the variation in the weights of the pellets. These are inserted into the trim bag during the manufacturing process (Note 6) to make sure that the powder bags weigh out to an average of 100 lb for the 16-in D846 bag and 110 lb for the 16-in D839 charges. There were on average some 15-65 grains per tare layer.
In the full-charge bag, the grains were carefully layered, whereas the reduced charges were simply dumped into the unbleached silk bags. The reduced charges weighed 52.5 lb, about half as much as full charges.
Sandia was suspicious of the trim-layer pellets, which were tubular nitrocellulose, about 2 inches long, with seven small holes in the pellets to ensure quick, even burning. If there had been fewer than twenty, could the powder be ignited under the pressure of the overram7 Sandia answered this question on 24 May 1990 at the Naval Weapons System Center, Dahlgren, Virginia. No ignition had occurred during seventeen drop tests of five powder bags of the 16-in/45 D846 type, replicating the rammer velocity of 14 fps. The last of these drop tests crushed five of the pellets in the trim layer. An eighteenth test was ordered with fewer pellets, which were specially arranged in a circle with one at the center in the vicinity of the black-powder pouch. Upon dropping the charge a bright light was observed coming from the space between powder bags 2 and 3. This was followed by smoke and yellow flames and an explosion that tore the experimental rig apart. On 25 May 1990 Schwoebel and Schueler appeared before the Senate Armed Services Committee with their findings and stated that they could neither prove nor disprove the theory that a chemical device had set off the explosion. These men did note that a reduced number of pellets in the trim layer increased the possibility of explosion, particularly if a high-speed overram did occur.(Note 6) This indeed was one of the conclusions reached in Sandia's final report. An actual count and statistical analysis of powder bags on board the Iowa found that there were powder bags with 1 to 65 pellets in the trim layer. Specialists from Sandia determined in subsequent mathematical analyses of the tests that the chances of an explosion greatly increased as the number of pellets decreased. Furthermore, they concluded, unlike the Navy's investigators, that despite the rammer lever's being in the low-speed rammer position, it was possible that the explosion could have moved the rammerman's seat and caused the rammer lever to show a rammer-speed position that was different from the actual speed used for ramming.(Note 7)There is yet another sensitive configuration of pellets that did not involve the trim layer. The charges were ignited when at high ramming speed a single pellet was misplaced at the rear of the bag adjacent to the black-powder patch. A subsequent examination of the powder bags showed that 3.39 percent of them had a misplaced pellet at the rear of the bag. Exhaustive testing at Dahlgren in 1990 with reduced pellets in the trim layer confirmed that a cold-barrel explosion was possible with a high-speed overram. The probability of this type of explosion was estimated to be 1 in 38,000, but almost zero {a probability of 1 in 1 x 10^-47) if a low speed overram occurred.
The need to properly train gun crews to operate the 16-in guns of the Iowas was a concern to the U.S. Navy before this explosion occurred. NAVSEA had established a 36-42-week course in main-battery gunnery, and companies like RCA had developed the course syllabus. Former battleship officers and crews were hired to teach these courses. One of the instructors wrote, ". . . competent turret crews and main battery fire-control personnel, continuous on-board training with hands on the equipment is the order of the day."(Note 8) According to Tom Meiners,(Note 9) who was a gun captain in Turret II and was discharged from the Navy in March 1989, there were frequent overrams because there was constant pressure from the bridge to decrease the loading time and fire rounds as quickly as possible. Meiners also said that Hartwig was really careful when he rammed the powder. "You could not get him to go fast enough. He was really careful around this stuff {powder)."
One other point turned up in our review of the various investigative reports; the morale of the Iowa's gunnery department had declined significantly from the day that Captain Seaquist had relinquished command to the day that the accident occurred. Under Captains Gnecknow and Seaquist, the gunnery department had had free access to the captain, and with the directive from Secretary Lehman and President Reagan, it was essential that such direct communication occur.(Note 10) Under the command of Captain Moosally, who had commanded the guided missile destroyer Kidd, the emphasis on gunnery changed, and the missile armament was given a higher priority; also, his major concern was the engineering plant, because he had a strong background in propulsion. In statements to the press, Captain Moosally maintained that he was not aware of the testing occurring on board the Iowa and that he was given a number of misfits for his crew. We find his statements difficult to believe because of the gunnery profi-ciency that the Iowa had attained by 1988, but we did determine that he was reluctant to allow civilians on board to conduct tests on powder or new projectiles. The gunnery department felt, based on his actions, that he would not make any of his time available for gunnery problems.
What does all this mean? It appears very probable that an overram did occur in Turret II. Even though it was known to gunnery personnel that overrams decreased ballistic performance and led to inconsistencies in salvo spreads, bridge personnel pressed the crews to fire rounds more quickly. The fire storm that swept through the gun house and turret substructure de-stroyed much of the evidence that might have explained the circumstances surrounding the overram. It was possible that there might have been a momentary mechanical problem with the lever for the rammer (i.e., it was stiff I and that the new rammerman overzealously used its high-speed setting momentarily. It must be emphasized here that Backherms was new to his position and that this was his first experience with live ammunition. Senior Chief Ziegler's "We have a problem here!" might have referred to the inexperienced rammerman. There is a possibility that one of the powder bags broke during this overram, but tests done at Dahlgren after the incident showed this to be an unlikely event with a slow ram. The efforts of the center-gun crew to bring the powder bags to the proper distance from the breech may also have been the source of Ziegler's statement. With this serious overram, the quilted patch containing the black powder might have been caught between the gun barrel and the silk liner of the bag itself. According to regulations, if a propellant bag or ignition pad is ripped, torn, or otherwise damaged, all operations in the turret should cease. Qualified disposal personnel should be contacted to dispose of the damaged bag and spilled black powder or propellant grains. There may have been some noises in the gun barrel as the damaged pellets were being compressed into the black-powder pouch that was quilted to the rear of each powder bag. Noises could also have indicated that the men were busily attempting to extract the rammer from the open breech. Frictional tests and shearing tests of the powder grains showed that this was an unlikely event as well. In the excitement of that moment, the powder-hoist man forgot to start the mechanism that would have returned the powder hoist to the powder flat in the lower turret substructure. There simply was not enough time to extract the rammer and close the breech. The deflagration was sudden and killed all forty-seven men who might have explained how and why this catastrophe occurred.
There was one other cold-barrel gun explosion involving a major-caliber gun of 12-in or larger caliber. In September 1940 the Vichy-French battleship Richelieu engaged a British and Free French invasion force off Dakar. The first salvo from the starboard gun number 7 of Turret II resulted in an in-bore explosion that destroyed the barrel (see appendix C). An official investigation revealed that either the powder or the shell may have been at fault in this incident because the powder bags were reconstituted from the reserves for the Dunkerclue and Strasbourg and the shells were of a new type. The official cause was attributed to an overpressure from the propellant being used in Turret II that caused a failure in the baseplate of the projectile in the gun barrel and caused the detonation of the explosive filler in the projectile cavity.
The allegations of unauthorized testing on board the Iowa have been overstated. The main concern of the U.S. Navy was that funds had been expended without proper authority. There was a mandate from Congress, the President, and the Secretary of the Navy-find ways to improve the accuracy of 16-in gunfire with a minimum expenditure of ammunition, or else the funding for the reactivation of the Missouri and Wisconsin would be withheld or canceled. Safety and reliability were also primary concerns of this mandate.
After the May 1990 tests by Sandia, the Naval Weapons System Center, Dahlgren continued its own research, trying to recreate in the gun barrel that which had been demonstrated in drop tests. Despite exhaustive testing of overrams with reduced pellets in the trim layer, it was possible to create a cold-barrel explosion only with a high-speed ram. These tests were accomplished with a 16-in/50 test fixture at Dahlgren with high-speed overrams (14 fpsl. In his testimony before the Senate Armed Services Committee, Dr. Schwoebel conceded that frictional forces in a 16-in gun might make a replication of the explosion difficult in a gun test rig. From May through November 1990 test firings using an overram did not produce the type of ignition at a low-speed rammer setting that had occurred on board the Iowa.
There are few absolutely certain things in life, but the authors are convinced that, according to the evidence, the turrets on the Iowa and her sisters were safe and reliable as long as safety precautions were followed. Not only were the turrets rugged, but the entire system of gun, projectile, and propellant was forgiving. Unfortunately, Sandia's tests showed with some cer-tainty that circumstances can lead to an open-breech tragedy, particularly when the number of pellets in the trim layer is reduced and high ramming speeds are used. Overrams had occurred in the past with powder bags with reduced pellets in the trim layer without producing explosions.
The likelihood of sabotage has been neither proved nor disproved conclusively, nor has an accidental ignition brought about by inexperience been confirmed. Forensic evidence col-lected in the turret from the remains of the deceased did show that there were lighters and other items not permitted by regulations in the gun house during this gun exercise. However, the cause for this turret deflagration will never be known conclusively, although personnel skilled in ordnance and ballistics have indicated that the accidental ignition of the propellant is a less likely explanation of the cause of the explosion. Warships, designed to be capable of destroying the enemy, are by their very nature susceptible to accident or tragedy-careful design, prudent equipment maintenance, adherence to regulations, and sound personnel training can lessen but can never completely prevent the likelihood of an accident. This unfortunate mishap might have been averted if Captain Moosally had authorized more gunnery drills, a situation that may have raised the crew s morale and lessened the potential for disaster. However, this would not have averted an intentional act of suicide or sabotage.
Based upon the intensive investigations into the turret deflagration, Sandia National Laboratories made the following recommendations for the future operation of 16-in guns on board Iowas:
Based upon these recommendations, the U.S. Navy did implement an earlier recommenda-tion that the powder bags for the 16-in/50 guns be redesigned to eliminate trim layers. All powder ammunition used in Operation Desert Storm, for example, was this type of bag. Before steaming back to Norfolk, the Iowa briefly stopped in Puerto Rico to land the remains of the crew members who had perished. A memorial service was conducted at the Norfolk Naval Station upon her return, with President and Mrs. George Bush attending. Afterward they spoke personally to each family of the forty-seven men killed in this disaster.
After her ammunition was unloaded, the Iowa underwent a limited ship repair. Turret II was trained in and its guns lowered using its own motors and gearing, testimony to the massive, rugged structure of the turret machinery. The damaged internal structure of the turret, the rangefinders, and equipment of Turret II were removed to be reconditioned or replaced. The Naval Ordnance Station, Louisville, refurbished some of this equipment. Included in the reconditioned equipment were the rammer assembly, gunfire-control computer, control panel, switches, periscopes, and rangefinder. These were later stowed in Turret II or at the Naval Ordnance Station, Louisville, where they can be accessed for future use. The turret was sealed.
On 7 June the Iowa departed from Norfolk for her scheduled six-month deployment to Northem Europe and the Mediterranean. She became the flagship of the Sixth Fleet, as flag facilities had been completed during one of her shipyard availabilities in 1988-89, and contin-ued in this capacity until relieved by the Belknap. She returned to Norfolk in December 1989 to commence final repairs to Turret II. Although these repairs were authorized and funded, they were never completed. All damaged equipment in Turret II was reconditioned and scheduled to be returned to the Norfolk Naval Shipyard for reinstallation on the Iowa. This included the optical rangefinder in the turret and the radar equipment.1 This is to prevent copper buildup on rifling from the rotating bands of the projectiles.
2 In a normal powder and shell ram, a slight over pressure is created in the gun house, and these gun bloomers will actually bulge out, indicating that the shells and powder have been successfully engaged in the barrel. This bulging was observed in videotapes of the incident in all three gun ports of Turret I that morning.
3 This theory is based on the fact that some of the powder grains from the powder bags in the center gun were scattered within the turrets gun house and substructure. An inventory, however, was never made of these grains, which were scattered throughout Turret Il. A similar scattering of unspent grains was discovered during drop tests at the Sandia laboratories in New Mexico.
4 During these tests some powder grains were spilled on a railroad track at Dahlgren. An empty flatcar rolled over one of these grains, and it did not ignite; but when loaded with a 1 6-in gun, this same flatcar caused the ignition of another grain on the rail.
5 The trim or tare layer was sewn on the opposite end of the bag than that to which the black powder quilted patch was attached
6 The reports indicate to us that the rammer lever was engaged in the low-speed position after the explosion
7 The rammerman~s seat was found to be lodged against this lever after the explosion. The Navy had discarded the seat, so Sandia was unable to formulate any further hypothesis regarding it. However, the U.S. Navy did some computer modeling and found that the seat was unlikely to have caused the lever to shift from a high-speed to a low-speed rammer position
8 Muir, The Iowa Class Battleships, p. 129
9 Quoted in "Ex-gun captain of Iowa questions Navy version of Blast," Washington Times, 17 November 1989.
10 Secretary John Lehman had directed that the gunnery of the lowas had to be improved such that they would be able to direct all their hits within an area the size of the grounds of the U.S. Capitol. Tests took place at the vieques gunnery range in the Caribbean. He did not want the Beirut experience of the New jersey to be repeated in the future use of the 16-in guns
11 This was implemented on the Missouri and Wisconsin, and paperwork was completed for the Iowa and New lersey should they ever be reactivated.
