The Torpedoes That Failed

We have all heard of the extraordinary daring of the American submarine commanders in their raids against Japanese shipping. But the actual box score was not disclosed, and not until recently has the public begun to realize that the performance of American torpedoes in the Pacific must stand as one of the great all-time failures of American ordinance. Why this was so and how it was remedied is now recounted by FLETCHER PRATT, widely known as a commentator and author of many military and naval histories dealing with the two world wars.


ON December 14, 1941, the submarine Sargo fired a torpedo at a cargo carrier of the Japanese invasion fleet bearing down on the Philippines. The torpedo missed. On December 24 Sargo got into a Japanese convoy and fired five torpedoes. They all missed. On December 27 she fired torpedoes at two freighters. They missed. On December 28 she fired at a transport and later in the same day at another cargo carrier. The torpedoes missed. On December 29 and January 1 she made three more attacks. All the torpedoes missed.

On October 19, 1942, nearly a year later, the submarine Guardfish made a five-hour approach on a large freighter under airplane escort in the East China Sea. Four torpedoes were fired, of which two were seen to run under the target. The submarine immediately fired her two remaining forward tubes. One torpedo ran under the target; the other exploded prematurely before reaching it, and the submarine was forced to dive to escape the counterattack of the planes.

On April 8, 1943, the submarine Tunny got into a Japanese carrier formation of Truk at night and, in a brilliant piece of maneuvering, worked into a position from which she almost simultaneously fired her six bow tubes at one carrier and her four stern tubes at another. All four stern torpedoes exploded and three of the salvo of six, but the explosions were all premature and the carriers escaped.

On November 9, 1943, when the war had been going on for nearly two years, the submarine Seawolf attacked the large Japanese transport Atatsuki Maru in the South China Sea, firing a total of eight torpedoes at her. Two of them ran under the ship; four struck her but failed to explode, and the vessel escaped.

These four cases happened to be rather spectacular, but they were by no means atypical. The commander of Submarine Division 102 endorsed the patrol report of one of his boats with the remark that 37.5 per cent of all the torpedoes fired were defective. There is no doubt whatever that the war was sensibly lengthened by the failure of American torpedoes to behave as expected. When the American submarines got really good torpedoes early in 1944, the Japanese cause began to go to hell in a hurry.

Was this condition inherent in the torpedo as a weapon, or were our torpedoes worse than those of our enemies? The answer to the second question is not simple or easy to recover, but if it has to be put in a single word — yes. The Germans had a lot of trouble with their torpedoes in the early months of the war, but it didn’t take them anything like two years to find out what was wrong and to apply the cure. The Japanese never did get an airplane torpedo as good as ours, and some of their submarine torpedoes malfunctioned. The one that sank the carrier Wasp, for instance, was a mechanical failure; it leaped out of the water, but it just happened to go off against a portion of the ship’s side behind which a gasoline line was running. Their destroyer torpedoes, however, remained right to the end of the war superior in range and hitting power to ours, and their submarine torpedoes never approached 37.5 per cent failures. In any case, the failures of other nations afford small excuse for those of our supposedly mechanically-minded and technically competent nation.

Why these failures occurred is a rather complicated story of official lassitude, official reticence, false economy, and inadequate research. It has long been covered by the necessities of the war, but it deserves rather close attention lest the same sort of thing happen again. The submariners, who were out on the firing line undergoing the heartbreaking experience of seeing perfectly planned attacks go wrong through the failure of their weapons, place the major blame on the incompetence of the Bureau of Ordnance, which has charge of torpedoes, and let it go at that. The Bureau of Ordnance, which, like all official procurement bodies, never has money enough to do what it wishes, places the major blame on lack of funds. The two agencies unite in allocating part of the fault to the Office of Naval Intelligence, which has charge of guarding secrets.


BEFORE examining the case it may be well to see what a torpedo is. The Mark 14 submarine torpedo, with which all but the oldest submarines were supplied at the beginning of the war, was a device costing something over $10,000, 21 inches in diameter, a little over 21 feet long, and 1¼ tons in weight. In effect it was a robot submarine. A spray of water forced by air pressure through an alcohol torch became steam, which operated turbines that drove the torpedo through the water at a speed of 46 knots to a limit distance of 4500 yards. It could also be set to travel nearly double this distance at a speed around 30 knots, but this set - ting was rarely used and was dropped out of most torpedoes late in the war. Gyroscopic controls could be set to give it any desired direction after leaving the tube. The depth at which it ran was controlled by a valve reacting to water pressure, and this could be set before firing to give any desired depth.

The business end of the Mark 14 was its war head, containing 507 pounds of TNT. It could be set off by banging into the side of a ship (contact exploder), or by entering the magnetic field of a ship’s steel hull (magnetic exploder). Neither type of exploder was operative until the torpedo had run 450 yards-an arrangement designed to protect the submarine against the premature explosion of her own torpedo, since a torpedo runs very erratically during the early part of its course, seeking the depth and direction for which it has been set. The shock of sudden changes might make it go off.

Much of the torpedo trouble traced ultimately to the magnetic exploder, which has a long history. Toward the end of World War I the Germans began using a magnetic exploder in mines, and an improved and developed form of it was the first and greatest of Hitler’s secret weapons in World War II, the magnetic mine which sank so many ships in the English Channel in the fall of 1939.

The heart of the German device was a compass needle which swung when approached by the mass of iron represented by a ship, the movement of the needle making an electrical contact which set off the explosive. During the twenty years between wars it occurred independently to Germans, Americans, and British that if this device could be applied to torpedoes, and the torpedoes then be fired to pass directly beneath a ship, the results would be sensational. The underwater side of a warship carries armor and compartmentation to deaden the shock of torpedoes. Not so the bottom; and in addition there was a better than even chance that an explosion beneath a vessel’s keel would break her back.

The official, stated view on the subject was that such a beneath-keel hit would be three times as effective as one against a ship’s side. This turned out to be an error; it was based on the idea that torpedoes would be used almost exclusively against armored warships-battleships, cruisers, carriers — and it contained a miscalculation, the first of several gaps to develop between theory and practice in the story of the torpedo. In 1943 the Bureau of Ships completed an investigation of how vessels went down under torpedo attack, and it was found that in a majority of cases they capsized; that is, a side hit which let water in to unbalance a ship was more effective than a bottom hit.

Also, as early as the Norwegian campaign of 1940, the Germans discovered that their magnetic exploders were not working at all well in sub-Arctic waters, because of the fact that the earth itself is a huge magnet, whose magnetism varies in kind and degree from place to place. Before the end of that year both they and the British decided that their magnetic exploders were so untrustworthy that they should be discarded. Before the end of 1942 Doenitz was prodding his scientists to produce one that would work.

The American magnetic exploder had been “perfected” as early as 1925, on lines quite different from the German. It gave satisfactory results in tests, and our people thought they had the only magnetic torpedo exploder in the world. The theory was that this torpedo, three times as efficient as any other, and knowledge of whose very existence would affect the design of potential enemy ships then under construction, should be held a deadly secret, to be launched at the enemy in mass and as a surprise.

When Admiral W. H. P. Blandy (later to command the Bikini experiment and the Atlantic Fleet) became head of the Bureau of Ordnance in February, 1941, he failed to agree with this theory. He pointed out that the general lines of scientific development are known on a world-wide basis, and that technical secrets keep about as well as eggs — that is, you can hold them under refrigeration for a time, hut not a very long time. In the meanwhile it is slightly important that the people who are going to use the device know something about what they’re handling. In April. 1941, after Admiral Stark had already written to his Fleet commanders that “the question as to our entry into the war now seems to be when not whether,”the new Mark 14 torpedoes were issued to the Fleet.

But even then only the captains and torpedo officers of submarines were told about them, with the warning that they must keep the matter very secret. Parenthetically, destroyers got torpedoes with magnetic exploders at the same time and it turned out to be a good idea that the issue was not delayed till the outbreak of hostilities; in at least one destroyer squadron the current tactical system called for firing torpedoes under our own ships to hit the enemy. The restrictions on knowledge of the Mark 14 wert later relaxed to the extent of allowing enlisted torpedo men to receive enough instruction in the mechanism to be able to handle torpedoes containing the magnetic exploder. But at least during the early part of the war no one in the Fleet knew the inner workings of the magnetic; there was no one who knew how to service it if it got out of whack, or what appearances would indicate a magnetic exploder was not in perfect condition. When the war began, there was no one in the Fleet who had any experience with the new torpedo under service conditions. The submariners went to war with a sword whose weight and temper they had never tested, and it would be surprising if they had not sometimes mishandled the weapon.


THE failures traceable to the magnetic exploder were primarily responsible for the difficulties of Tunny and Guardfish mentioned above. For a long time, however, they were overlaid by another set of performance failures, quite different and, on the whole, more serious in its effects. The faulty magnetic exploder could be deactivated, put out of business, and this was often done. The torpedo thereupon became an ordinary contact weapon. The other, the major, defect prevented torpedoes from hitting their targets at all. It was this: the Mark 14 consistently ran anywhere from 10 to 15 feet below the depth for which it was set, and it sometimes ran out of line as well. Yet every torpedo supplied to the Fleet was tested by the torpedo station at Newport and certified as sound.

It may be noted in passing that in the early part of the war the Germans experienced some difficulty with deep-running torpedoes, and from one of the same causes. On older torpedoes the hydrostatic valve which controlled depth running was well forward on the fat part of the torpedo. With the increase in range and speed achieved between wars, the space there was needed for other things. It was also felt to be an advantage to place the controlling mechanism nearer the rudders it was to control, so the valve was moved aft to where the torpedo thins down toward its tail. It occurred to no one, German or American, that thus setting the valve at a slight angle to the torpedo’s axis of progress would cause it to react to another depth than that for which it was set.

As a matter of fact, this difference was quite small, and as long as torpedoes were run at fairly shallow depths and in smooth waters, the normal testing conditions, it did not matter. Moreover, the depth recording instrument by which the performance of the valves on American torpedoes was checked turned out to be faulty. No one knew it for years, because the recorder erred in the same direction and by the same amount as the valve. This was straight hard luck; it was like using two adding machines to check each other and having both of them make the same mistake for quite different reasons — a peculiar coincidence.

To this was added another source of error in 1940, when reports from the war zone indicated the need of heavier explosive charges in torpedoes. Careful tinkering enabled the Ordnance people to pack another 115 pounds of TNT into the war head of the Mark 14. By mid-1942 the figure became still higher, but the exercise heads with which the torpedoes were tested before being sent to the Fleet were not changed. Partly this was because no one conceived that a mere 115 pounds would make much difference in an object weighing well over a ton and moving at well over 50 miles an hour. Partly it was because of a design difficulty. A normal exercise head was filled with water, which the torpedo automatically ejected at the end of its run so that it would float and could be picked up, water occupying about the same space for weight as TNT. But the extra explosive had been crowded into the Mark 14 war head by increasing density, so that the water-filled head, though occupying the same space as the TNT war head, no longer had the same weight.

Yet as the tests gave apparently satisfactory results, no inquiry was made into the question of whether the tests themselves had not fallen out of relation to the service conditions under which the torpedoes would be used — at least until the torpedoes began to fail in alarming numbers.

This was not entirely the fault of the Bureau of Ordnance. They had few or no submariners on hand to point out the differences between the test conditions and those encountered in service. It is the practice of the Navy, and not an unsound one, to allow officers to choose their line of duty so far as possible. A submarine officer’s duty-life in this specialty is very brief; he is old in the service at thirty-five and out of it by forty. None of the young men in that most active of branches wanted to spend two or three years fooling around with slide rules at Newport while their classmates were running up records of sea service. Nobody in the upper echelons thought of ordering submariners to Newport whether they wanted to go or not.

Even if the submariners had been on hand it is doubtful whether they could have done much about the real difficulty, which was the conditions under which the tests were conducted and the atmosphere with which they were surrounded. The only torpedo station was at Newport and the range there is too shoal to permit real deep-running tests. The torpedoes were tested by being fired from a “torpedo barge” instead of from the tubes of submarines.

On top of this the emphasis was always on saving money. It was important to recover those $10,000 torpedoes in good condition after a trial run; there were only so many of them. If they ran under a target, what the hell; they were to be used with the magnetic exploder in any case, and a near-miss underneath the ship was three times better than a direct hit. Above all, tests must be conducted against a target, that would permit recovery of the torpedo; the United States could not afford to expend ships by firing live torpedoes against them, and could not afford the torpedoes to fire. Behind this was the attitude of the Ordnance officers. They tended to look at a torpedo not as an expendable piece of ammunition but as something they could not bear to destroy.

The Germans and Japanese, to whom everything was expendable for military purposes, could and did afford actual torpedo explosion; that was one of the reasons why they got better torpedoes. There was also an administrative responsibility. The Roosevelt administration did not stint in providing funds for big, beautiful ships, but it was extremely niggardly about allotting funds for research. During the years of darkening war clouds the Bureau of Ordnance never once received as much as it asked, and it was constantly engaged in an effort to reopen the second torpedo station at Alexandria, Virginia, which had been set up during the First World War.

This effort succeeded in 1939, after Europe had already begun to burn, but too late to prevent another factor in the malfunctioning of the torpedoes that went to the Fleet in 1941. A torpedo is an extremely complex mechanism. Those produced at Newport were practically hand-tooled. Suppose one was fired for testing and after it was removed from the water some small part was replaced. It was no longer exactly the same torpedo. Even though the difference could be expressed in a fraction of a millimeter, it became important when multiplied by the distance a torpedo must travel in action.

The Bureau of Ordnance realized this from the beginning and, more importantly, realized that its stock of torpedoes was nowhere near sufficient to run a war with. In 1941 the United States owned fewer torpedoes than it was expending every month by 1944; and of the small number we did have, 233 — nearly half — were destroyed when the Japs bombed Cavite. Along with the effort to reopen Alexandria went one to get some machine tools installed and to put torpedoes into mass and standard production. Of course, everyone was yelling for machine tools in 1939, so the process was a slow one, and it was not until the war was several months old that International Harvester Company and American Can Company began work on machine-produced torpedoes, with identical spare parts that could be fitted to any other torpedo without special adjustment.


THE resulting early shortage of torpedoes, most acutely felt among the submarines operating out of Australia in the early days of the war, produced still another source of submarine attack failures in the form of an order not to fire spreads. A spread consists of several torpedoes fired along slightly different angles like the spokes of a fan in order to make sure of getting a hit on a vessel which may be maneuvering rapidly, and to allow for errors in computing the target’s course and speed. The normal torpedo attack for a large valuable vessel, as used later in the war, consisted of a spread of three or four. But when Sorgo made her famous and fruitless cruise at the beginning of the war, she was using singles.

Nevertheless, it seemed to that submarine’s commander, Lieut. Cdr. Tyrrell D. Jackobs, that with all allowance for human error and the normal run of mechanical imperfections, the mere law of averages should have given something better than thirteen misses out of thirteen shots. He suspected deep running and requested test firing of torpedoes to check the depth control of the Mark 14; and he was not the only complainant. Check firings were made at Newport; in February, 1942, the Bureau reported back to the Fleet that the Mark 14 showed a tendency to run deep, but only to the extent of 4 feet and only during the first 880 yards of its run. Their depth recording device was still giving the wrong answers.

Now this 4 feet would hardly have mattered if the magnetic exploder had functioned properly. But with most of the bugs in that device still unsuspected, attention became concentrated on depth control. \\ hen in the spring of 1942 the submarines Salmon, and Skipjack reported their belief that their torpedoes were running deep, Rear Admiral Charles A. Lockwood, Jr., newly appointed commander of Submarines Southwest Pacific, decided to make some tests for himself. He set up a big fishnet in deep water at Frenchman’s Bay, near Albany, Australia. A service torpedo from Skipjack’s stock was fitted with an exercise head devised on the spot, containing a calcium chloride solution considerably heavier than water and having the exact weight of a war head, then fired from a distance of 850 yards with a depth setting of 10 feet. It went through the net 25 feet down, and the next day two more torpedoes cut the net at 8 and 11 feet deeper than they were set.

The report on this performance reached the Bureau of Ordnance almost simultaneously with one from Admiral R. H. English, commander of Submarines Pacific at Pearl Harbor, that a test he had made showed torpedoes running at least 8 feet below their settings. In July the Bureau replied that the Albany test was not conclusive, since improper trim conditions had been introduced by using an exercise head that was shorter than the war head. This was not quite the brush-off it seems. When Admiral Blandy took office in the Bureau, almost the first subject that attracted his attention was the unsatisfactory state of the torpedo matter, both as to quality and quantity. He called out of retirement the best torpedo design man he knew, Commander James L. King, and made him head of the Subsurface Warfare Branch in the Research and Development Section. It was King who had succeeded in packing the extra power into the Mark 14’s war head, and earlier it had been King who devised the turbine engine that drove our torpedoes, a far better engine than anyone else in the world had.

Captain King (he had reached that rank by 1942) was a technician and design man who wanted scientific accuracy. He was already concerned with the difference in weights between the exercise heads and the war head of the Mark 14, and when the reports from Albany came in, he had developed a chemical liquid of the proper density and was building a new exercise head with it as a base. He secured the services of a submarine for firing, and shot his torpedoes through nets instead of trusting the depth recorder. At about the same time a second series of tests was run off at Albany with calcium chloride heads extended to war head length. Both sets of tests yielded the same results, and they were conclusive. On August 1, 1942, the Bureau of Ordnance advised all interested parties that the Mark 14 torpedo ran approximately 10 feet deeper than set; and after that date the curve of Japanese shipping sunk took an upturn.

The upturn was not really satisfactory for some time to come, since a differently calibrated depth control valve bad to be designed and installed. The Bureau continued to get reports about torpedoes that ran close under ships or even hit them directly without going off, as in the case of the Guardfish attack cited above.

The American magnetic exploder was not a compass device, like the German, but operated by induction coils whose electromotive force was changed by passing through a ship’s magnetic field, the change being boosted by vacuum tubes to actuate the firing pin. As the complaints continued and it became evident that depth setting could not be responsible, the Newport station realized the fact that had earlier struck the Germans: that a ship’s magnetic field is not the same in all parts of the earth. Captain King held more tests, and in April, 1943, the Bureau reported that the torpedo was apt to explode prematurely when running at 12 feet or less below the surface. It advised that the magnetic exploder ought to be inactivated unless the torpedo was being fired for the keel of the target in all waters south of 30 degrees, North magnetic latitude. It was also suggested that the magnetic exploder would work better if the torpedo could be allowed to run 700 yards instead of 450 before becoming alive.

These recommendations went to Admiral King, who himself had a background in submarines. He rejected the 700-yard period as unacceptable (submarines often have to fire torpedoes at closer range than this) and passed the rest to Admiral Nimitz, who promptly ordered that all submarines under his command should inactivate their magnetic exploders for keeps and fire for contact hits only. The Southwest Pacific submarines, which were under another command, kept the magnetic exploder for some time; that was why Seawolf had her unhappy experience in November. To complete this part of the story, Admiral Blandy farmed out the job of finding a really workable magnetic exploder to four separate research agencies, and one of them finally produced the goods — but only at the very end of the war.


THE curve of Japanese shipping sunk now turned upward again, but the torpedo troubles were not yet ended. On the very day that Admiral Nimitz issued his order to put the magnetic exploders out of business, there was a dramatic exhibition of another type of defect that had been in the Mark 14 all along, concealed by more obvious troubles. North of Truk the submarine Tinosa came upon the largest tanker in the Japanese service and perhaps in the world, Tonan Maru No. 3, 19,262 tons, a former whale factory ship. Two torpedoes were fired from a difficult angle; both hit, one exploded nicely and brought the giant tanker to a halt, settling slightly by the stern. It was insufficient to sink so big and well-compartmented a ship. Lieut. Cdr. L. R. Daspit of Tinosa maneuvered his submarine in to 875 yards broad off the tanker’s beam and, during a period of several hours, poured ten more torpedoes into her, firing for contact. Every torpedo hit — they could hardly miss at that range and angle—but not one went off. After the tenth failure Daspit took his one remaining torpedo back to Pearl Harbor to find out what was wrong with it, while the Japs towed their damaged tanker home and salvaged her.

Examination of that one remaining torpedo failed to disclose anything wrong with construction or adjustment. But it seemed to Admiral Lockwood, now commander of Submarines Pacific, that this could hardly have been the only sheep in a flock of goats. If it were not, then there must be something wrong with the contact exploder as well as the magnetic. Without waiting for Newport or scientific accuracy, he ordered two torpedoes fired against some submerged cliffs. One of them produced a perfectly lovely explosion; the other was a dud. Divers fished up the dud for examination. Like the Tinosa torpedo, it was found to be perfect in construction and adjustment; the exploder had released the firing pin correctly, but the pin had not struck the primer cap hard enough to cause an explosion.


The officers at the Pearl Harbor submarine base loaded several torpedo war heads with concrete, filled them with contact exploders and primer caps, and dropped them down a wire for 90 feet onto steel plates in a dry dock. When they hit the plates squarely, nose on, in the position Tinosa’s torpedoes had hit Tonan Maru No. 3 and Seawolf’s had hit Akatsuki Maru, seven out of ten were duds; but when the plate was tilted so that the blow was slightly glancing, the exploders worked every time. Just five days before, the Bureau of Ordnance had conducted similar tests by firing torpedoes directly at an armor plate suspended under water in Chesapeake Bay. Half of these also failed to go off.

This made the rest of the analysis relatively easy. The trouble lay in the firing pin itself, a device of several ounces in weight which, when released by a hit, was driven by a spring against the primer cap. The point was that when a torpedo running at high speed made a head-on hit, part of the tremendous force generated — something like 500 times gravity — appeared as friction between the firing pin and the guides along which it rode for accuracy. This friction was so great that the spring was unable to drive the pin home. But as soon as the difficulty was discovered and the pin lightened, the trouble disappeared; the sinkings of Japanese ships took another jump and never went down till the end of the war.

There had been a certain amount of suspicion about this firing pin early in the game. As far back as the late thirties the Newport station tested the contact exploder by hurling torpedoes from a catapult against a steel plate backed with sand, and found that in many cases the spring did not drive the firing pin home. At that time the spring was strengthened and appeared to give satisfactory resells, but torpedoes had not yet attained a water speed of 46 knots. The greater impact developed by high speeds produced more friction and set things back to where they were before. If Lieut. Cdr. Daspit had set his torpedoes for slow running or glancing shots, they would have functioned perfectly.

Now it is not altogether fair to blame the Bureau for not suspecting something was wrong, nor even to blame the Office of Naval Intelligence in its passion for secrecy. The data furnished to the researchers were both inaccurate and incomplete.

In the Fleet the man in command is held responsible and his complaints about imperfect instruments are taken at a heavy discount by headquarters. They have to be, or every failure would be blamed on the tools. Just before the Tinosa report the Bureau of Ordnance was in receipt of a report from Admiral Lockwood to the effect that out of 600 torpedoes evaluated as known hits only 1.5 per cent had failed to explode. So the responsibility is divided.

What the Bureau of Ordnance could have done was to get rid of its psychological block against the expenditure of valuable material and to conduct tests under conditions approaching those of service. It could have fired real torpedoes at steel plates and steel hulls. It could have wasted some torpedoes in discovering that the magnetic exploder would premature at depths of 12 feet or less and that it was affected by changes in magnetism. It could probably have insisted that some active submarine officers be detailed to Newport.

This is demonstrated by the story of the electric torpedo. The first one was a German job, captured in 1942, which was turned over to Westinghouse for duplication. It didn’t work; some of the parts were not suitable for American manufacture and there was no way it could be adapted to fit American tubes. It was 30 per cent slower than the turbine torpedo and it had more bugs in it than a Latin American jail. But it showed no wake (thus preventing an attacked ship from dodging an approaching torpedo and preventing escorts from running back along a torpedo track to depthcharge the submarine), it would run reliably at any depth without preliminary diving and hunting, and its exploder worked every time.

Admiral Blandy demanded real submarine officers at Newport to help work the thing out, headed by Lieut. Cdr. H. A. Pieczentkowski, who had been largely responsible for the contact exploder tests at Pearl Harbor. He was one of the best submarine commanders in the Fleet and Admiral Lockwood was reluctant to let him go, but eventually he turned up at Newport, working with Captain King. There was a constant interchange between scientific accuracy and practical experience under combat conditions. By the end of 1943, electric torpedoes were going out with the submarines, and in 1945, 65 per cent of all torpedoes fired were electrics.

This seems to be the answer. The research men aren’t practical enough, and the practical men tend to stick with what they have already. There has to be a constant interchange between the two groups, and we can only hope that the lesson has been learned.