A Wellington with 10-cm radar and a Leigh Light. The scanner was in the ‘chin’ under the aircraft’s nose. The Metox warning receiver was unable to detect this new radar, with the result that many U-boats were surprised on the surface, like U966 (below) which was bombed and sunk a few moments after this photograph was taken on 10 November 1943 off Cape Ortegal in the Bay of Biscay.
At the beginning of 1943 the Battle of the Atlantic was finely balanced: both sides in the grim struggle were in a potentially winning position. The surviving U-boat commanders and crews were skilled, highly disciplined and well led, as were the Escort captains and their crews. They had all gained their experience in a very hard school and were backed by scientists whose work, one way or another, was in the end to prove decisive.
One of the British scientists, Professor Blackett, led the unglamorous-sounding department of ‘Operational Research’, whose scientists brought some highly original thinking to the war at sea. It was of these men that the C-in-C Coastal Command, Air Chief Marshal Sir John Slessor, said:
‘A few years ago it would never have occurred to me, or I think to any officer of any fighting service, that what the RAF soon came to call a “Boffin”, a gentleman in grey flannel bags, whose occupation in life had previously been something markedly unmilitary such as biology or physiology, would be able to teach us a great deal about our business. Yet so it was.’
One of the most effective examples of operational research was in the matter of the setting of aircraft depth charges. These were set to explode at a depth of 100 feet, on the assumption that, when a submarine was under air attack, it would have seen the aircraft approaching and crash-dived to a depth of from fifty to 100 feet by the time the aircraft was overhead and had dropped its charges. One of Blackett’s ‘Boffins’, E. J. Williams, looked into what actually happened: he found that when aircraft sighted U-boats three times out of four they were on the surface or just diving and could therefore be attacked with accuracy; if, on the other hand, the U-boat was already submerged when sighted, then in all probability it would turn under water and be lost. Williams showed that if a submerged U-boat was regarded as a lost target and attacks were concentrated on surfaced or diving submarines, a better ‘kill’-rate could be achieved. All that was required was to alter the depth-setting on the charges from 100 to twenty-five feet. As soon as this was done, the success rate in U-boat sinkings went up by two to four times, a result so dramatic that survivors thought that the British depth charge had been given a double weight of explosive.
Another telling result of operational research was a brilliant analysis of the optimum size of convoys. It was found that the same number of ships were lost, roughly speaking, whether the convoy was large or small. The actual figures were an average loss of 2.6% for convoys of less than forty-five ships, and 1.7% for larger convoys. The number of escorts, about six, was the same in each case, since the area of a large convoy is only slightly larger than that of a small one. (The perimeter for a convoy of eighty ships would be only a seventh longer than that of a convoy of forty ships.) Even if a U-boat were to break through the screen of escorts, it would be unlikely to sink more ships in a large convoy than in a small one, since the limiting factor would not be the number of potential targets but the torpedo-reloading time and the number of torpedoes available.
By mid-1943, large-convoy techniques had reduced the number of close escorts needed by one third, allowing the formation of support escort groups, which could go to the aid of any convoy under attack and hunt for U-boats joining or leaving the area. In this way the convoy was not left unprotected as it had been previously when its escorts left to chase submarines.
However effective the escorts were they could not alone defeat the U-boats. Radar-equipped aircraft were essential, either working with the surface vessels or attacking U-boats directly, but there still remained the Gap, already referred to, which could not be effectively patrolled by shore-based aircraft, at least until sufficient numbers of long-range aircraft became available. The obvious answer was to use short-range anti-submarine aircraft operating from carriers, but the only aircraft carriers available were in desperately short supply: HMS Courageous and HMS Glorious had both been sunk in the first year of the war, and in November 1941 the most famous of them all, HMS Ark Royal, was torpedoed by U81 while transporting fighters to Malta. The few remaining Fleet carriers were far too valuable to be risked guarding merchant convoys.
The first British aircraft to fly over convoys in mid-Atlantic were Hurricane fighters which were catapulted from HMS Pegasus (formerly the first Ark Royal – a seaplane tender dating from the First World War), later supplemented by requisitioned and converted merchant ships which were fitted with a single catapult and a Hurricane or Fulmer fighter. They became operational in 1941, not to attack U-boats directly – the fighters carried no depth charges – but the long-range Focke Wulfe 200 Condors of KG40, which were being used as reconnaissance and spotter aircraft for the U-boat wolf packs as well as bombing isolated ships. Catapulting fighters was a successful tactic – up to a point: the aircraft were unable to land back on the ship, and, if out of range of the nearest land, had to be ditched after their single flight.
It was not until the spring of 1943, that largely American-built escort carriers (CVEs), nicknamed ‘Jeep’ or ‘Woolworth’, began sailing with the merchant convoys. Twenty-three CVEs were eventually used by the Royal Navy alone. They carried some twenty planes – Grumman Martlet fighters (now known by their American name, Wildcats) and Fairey Swordfish. In addition to these purpose-built ships there were several Merchant Aircraft Carriers – MAC ships – which were tankers or bulk grain-carriers with their superstructure removed and a short flight deck fitted. Like the CAM ships, they too flew the Red Ensign and carried a normal cargo in addition to their four Swordfish. From the time of their introduction in 1943 until the end of the war, not a single ship was sunk by a U-boat in any convoy with which these makeshift carriers sailed.
Although the escort carriers did much to close the Gap, they were augmented, on the personal orders of President Roosevelt, by some sixty Consolidated B24 Liberators to supplement the small number already in service with 120 Squadron. These big, four-engined bombers had much of their normal defensive armour and weapons removed to make way for the maximum number of fuel tanks but they could fly far over the Atlantic and were still able to remain with a convoy for up to three hours.
There had been a struggle to get these additional B24 Liberators, for the requirements of the bombing offensive against Germany had taken priority. Coastal Command had been starved of four-engined aircraft, having to make do firstly, as we have seen, with inadequate types such as the Avro Anson, which lacked range and practically everything else required for antisubmarine work. The Sunderland Flying Boat helped, but it too had a relatively short range. The best of the early aircraft was the American Catalina; twin-engined Whitleys and Wellingtons were used, but they still lacked the vital range, and were anyway obsolete bombers converted for Coastal Command. Eventually, at the end of 1942, some four-engined Halifax IIs were received, later to be followed by a maritime version – the Halifax V, but no Lancasters were supplied until after the war.
The reason was that these four-engined British aircraft simply could not be spared from their primary role, the bombing of Germany. In any case, the Air Staff believed that the best form of attack against U-boats was to bomb their bases and the shipbuilding yards where they were built. In fact, not one U-boat was destroyed by bombing the U-boat pens on the French Atlantic Coast. The pens were hit all right – some 15,000 tons of bombs were dropped on the bases – but the massive reinforced-concrete structures were all but indestructible. Over 100 heavy bombers were lost in attacks on the U-boat bases in the first five months of 1943 alone.
In January 1943 British, Canadian and United States escorts and aircraft were faced with an average of 116 U-boats at sea each day. This large fleet was to be aided by the work of the German cryptanalysts, the B-Dienst whose code-breakers had penetrated the Allied convoy radio code. The U-boat high command could therefore plot the route of many of the convoys across the Atlantic.
In March, acting on information supplied by the cryptanalysts, thirty-nine U-boats were concentrated to intercept two convoys: SC122, a ‘slow’ convoy of fifty-two ships, and HX229, a ‘fast’ one of twenty-five. HX229 was attacked first and eight ships were sunk in as many hours. For three days the running battle continued, the two convoys joining to help their combined escorts fight the submarines; but in all nine ships, totalling 140,000 tons, were sunk – four of them by U338 alone – for the loss of only three U-boats.
That German success was to prove the high point of the battle; had they been able to sustain it, Doenitz’s ambition of cutting the lifeline between Britain and America would have become a reality. It was not to be. The escort carriers and the long-range B24 aircraft now closed the Gap permanently, and – most significantly – the scientists were about to provide yet another weapon which would in the end prove decisive: 10-cm ASV radar.
It all began with Randall and Boot in their Birmingham University laboratory and the development of the cavity magnetron – work, it will be recalled, conducted under Admiralty patronage. The shipboard 271 radar had been the first operational 10-cm set and, following its success, work had begun on the 10-cm ASV as early as the winter of 1941; but pressure from the RAF had secured priority for H2S and that radar had been developed first. However, the H2S team under Dee at Malvern had designed an ASV capability into H2S. Indeed it was known at TRE as H2S/ASV.
The Metox countermeasure enabling the U-boat crews to be warned of the approach of the l½-metre ASV-equipped aircraft had made a drastic change of wavelength essential. Since it was known that the Germans, lacking the magnetron, considered 10-cm radar impractical, it was thought unlikely that they would be expecting British aircraft to be using it. Ten centimetres was therefore the obvious answer.
There was some opposition: in the first place, Bomber Command were claiming total priority for H2S radar in the bombing of Berlin and, even at TRE itself, many people felt that 10-cm ASV was insufficiently developed, and that its introduction was premature. The adaptation of H2S to ASV was, as Sir Bernard Lovell remembers:
‘… most bitterly opposed and we were not allowed to divert any H2S equipment from Bomber Command. In the end, we made the [ASV] equipment ourselves in TRE.’
The answer to the objections to ASV was provided by the tonnage of ships being sunk in 1942: 5,970,679 tons – 1354 ships – by September. The early success of airborne ASV and Leigh Lights had been nullified by Metox, so in the autumn of 1942 the decision was made to divert some H2S sets from Bomber Command for fitting as ASV Mk III in Leigh Light Wellingtons.
The main difference between ASV Mk III and H2S was in the scanner position: it was simply not possible, due to ground clearance and other structural problems, to mount the H2S cupola under a Wellington. The only possible alternative was in a ‘chin’ under the nose; this entailed a considerable redesign of the scanner, which in any case now had to work at an altitude of 2000 feet instead of 20,000, with a 40° blind spot behind the aircraft. Even when these problems had been solved, further delay was caused – at a time when shipping losses were running at around 600,000 tons a month – by the RAF insisting on unnecessary refinements such as blind-landing and homing-beacon facilities being incorporated in the sets.
In spite of all the difficulties and delays, two prototype ASV Mk III sets were hand-built at TRE and fitted to two Wellington VIIIs (LB129 and LB135) at Defford during December 1942. By the end of February 1943 twelve Wellingtons based at Chivenor, backed up by almost as many TRE scientists as airmen, had ASVs installed. On the evening of 1 March two Wellingtons took off from Chivenor for the first patrol with the new radar over the Bay of Biscay (it was just a month after the first H2S raid over Germany by Bomber Command). No contacts were reported but, to the relief of the scientists, the crews had no difficulties with the new set.
During the night of 17 March, the first U-boat contact was made by 10-cm ASV at a range of nine miles; unfortunately the Leigh Light jammed, and no attack was possible. The same aircraft, a Wellington XIII (HZ538), obtained another sighting the next night at seven miles; this time all went well and the U-boat was attacked with six depth charges, the crew reporting that ‘the submarine was fully surfaced and under way, showing no signs of suspecting attacks’. This was, of course, the whole point of the 10-cm ASV; with Metox unable to detect the new radar, the success rate grew, particularly in the Bay of Biscay. In March thirteen submarines were caught at night, and there were twenty-four attacks in April. It was the same success that Coastal Command had enjoyed in June 1942 when the Leigh Light had first been introduced: the U-boats could no longer risk crossing the Bay at night, and had to run the gauntlet on the surface in daylight.
The success of the new ASV and the larger number of long-range aircraft were such that in May two convoys, ON 184 and HX239, arrived in British ports without a single ship lost: the German Navy, on the other hand, lost six U-boats vainly trying to attack them. In all, no fewer than forty-one U-boats were sunk by escorts and land- and carrier-based aircraft that month. Faced with these increasing losses, Doenitz ordered his submarines to fight it out on the surface when attacked by aircraft in daylight. Additional flak weapons were added to the existing armament, and initially the heavily-armed U-boats had some success against unwary aircraft. But the RAF soon developed a very simple counter; aircraft finding a U-boat on the surface simply flew in circles, just out of range of the submarine’s guns, and called up the nearest surface units. The U-boat was constantly watched and, as soon as the tell-tale plumes of spray indicated the submarine was blowing her tanks to dive, the plane would close for an attack. This developed into a grim race – the ‘battle of the seconds’. With a highly trained crew it was possible to clear a U-boat’s deck and dive in about thirty seconds. If they got down in that time they had a chance; many did not and were sunk by the aircraft’s depth charges or rocket projectiles.
As more and more U-boats were sunk and the lucky ones that had limped back to base reported that the Metox sets were not giving any warning of aircraft attacks, German scientists were perplexed. They discounted 10-cm radar for the not very sound reason that they themselves had been unable to produce a practical one.
Then a captured RAF airman mentioned under interrogation that the attacking aircraft were homing on the U-boats by signals radiated by the Metox set itself. To this day, the identity of this man remains a mystery; also a mystery is his motive. Whatever the reason, the effect on the Germans of this bogus intelligence was dramatic.
Laboratory tests showed that the Metox receiver did radiate a small signal (most radio sets do). U-boat HQ immediately ordered the entire fleet to turn off their receivers at once. The sets were then extensively redesigned and completely screened, so that not the smallest signal was radiated. It made no difference, of course. The U-boats were still attacked out of black night skies; one moment sailing on the surface, secure and seemingly invisible, the next a blinding light heading straight for them, then the roar of four 1200-hp engines and the blast of the shallow-set depth charges as a B24 Liberator or a Sunderland flew over at fifty feet.
How were the aircraft finding their small targets now? The Metox sets were no longer radiating: infra-red was a possibility, but German scientists came to the conclusion that a new and unknown radar was far more likely. Its identity was not long in coming: an RAF Stirling night bomber, equipped with H2S, was shot down over Rotterdam. As British experts had feared, the magnetron was recovered intact and the Germans established its working frequency and wavelength: 10 centimetres. It was to be reconstructed as the ‘Rotterdam Geräte’, but for the moment the answer to the U-boats’ problem was clear and simple: a new Metox-type warning receiver, but working on 10 centimetres.
Now to receive 10-cm radar pulses is a very much easier proposition than to transmit them, and a search receiver, the Telefunken FuMB7, ‘Naxos’, was quickly installed in Atlantic U-boats. A small dipole aerial was held up by a crewman on the conning tower whenever a U-boat was surfaced; any 10-cm pulses would be picked up and the receiver would give a warning on a cathode-ray tube below in the U-boat. The Naxos set covered the ‘S’ band, from 2500 to 3700 mHz (12 to 9 cms).
Naxos was not as effective as Metox for two reasons: it was omnidirectional – all it would do was to warn that an ASV-equipped aircraft was in the vicinity; and that vicinity would be pretty near, since the simple dipole aerial had no gain, and the set itself was insensitive. Early Naxos had an additional disadvantage in that the aerial, small as it was, still had to be brought down the conning tower so that its cable cleared the pressure-tight hatch. The handling of a long length of co-axial cable often damaged it, with the result that the receiver gave no warning at all.
The big worry on the Allied side was that the Germans would find an effective warning receiver as good on 10 centimetres as Metox had been on 1½ metres in September 1942. This problem had been much in the minds of the H2S/ASV team at TRE, where there had been rather pessimistic estimates of how long the ASV Mk III would be ‘safe’. It was an axiom of those days that any new device was considered to be safe from countermeasures for only a matter of weeks and therefore, whenever possible, a second-generation model was ready to take over. The obvious immediate counter, should Naxos prove efficient, was a jump to another frequency; this meant even shorter wavelengths – to ‘X’ band, 3 centimetres. An X-band H2S had been developed, and an ASV version was also produced. However, simply changing the wavelength was only a short-term solution: sooner or later the Germans would discover it and provide a suitable receiver. But there were other ways of defeating the listening sets.
ASV Mk VI, which operated on 3 centimetres, was issued to Coastal Command in January 1944. It had two new features. Its range was increased by stepping up the output from 50 kw to 200 kw – a staggering figure for an airborne radar in those days. The anti-listening measure was the provision of an attenuator control. The operator, once he had gained a contact, could then reduce the output of the radar to a point where he could just maintain the target on his screen as they flew towards the U-boat. The effect of this on the enemy listeners was that the pulses did not appear to get any stronger and they therefore assumed that the aircraft was not homing on them until it was too late to dive to safety.
A Schnorkel tube covered with ‘Sumpf’. This reduced radar returns to some extent but it soon became detached from the U-boat’s structure by wave action, and salt deposits reduced its electrical effectiveness. The small dipole is the aerial for Tunis; it gave a good echo to 3-cm radars.
The Germans went to extraordinary lengths to counter the 3-cm ASV: there was a new search receiver – FuMB36, the ‘Tunis’ – which covered from 15 to 3 centimetres; and the periscope standards, schnorkel trunking and even, in some cases, the entire conning tower were covered in a special material called ‘Sumpf’, a sandwich of rubber with carbon granules impregnated in it. The sandwich was composed of two types of Sumpf: one had the property of variable resistance over its area, the other variable dielectric or electrical density. The object of the coating, code-named ‘Schornsteinfeger’ (chimneysweep), was to make the structures above the water absorb the radar pulses, reducing the strength of the echo and making them less ‘visible’ to radar. In tests under laboratory conditions, Sumpf appeared to offer some promise, but it was not really practical. The sea tended to remove the covering, salt deposits reduced its electrical properties, and finally, the permanent aerials now fitted to detect the 3-cm radars returned excellent echoes. So in spite of these measures the attacks on the U-boats continued. The Luftwaffe sent Ju88 fighters into the Bay of Biscay to intercept Coastal Command’s aircraft; the RAF countered by escorting the anti-submarine patrols with Mosquitoes and Beaufighters, and the Royal Navy increased its surface ‘Hunter Killer’ groups.
A new phase of warfare then appeared: the radio-controlled glider bomb, the Henschel 293. This bomb achieved some success but was soon countered by jamming its simple radio-command link. Throughout 1943 the Battle of the Bay of Biscay was fought out. It resulted in the sinking of forty U-boats. Referring to the use of 10-cm ASV, Hitler conceded:
‘The temporary setback to our U-boats is due to one single technical invention of our enemy.’
The Germans too had new technical inventions, for example ‘Pillenwerfer’, or ‘Bolde’ – an abbreviation of the German word ‘Lügenbold’, meaning habitual liar. This was a canister of chemicals which, when released from a submarine, caused a cloud of fine bubbles which blocked the ASDIC: a marine version of ‘Window’.
There was also an acoustic torpedo, the T5, Zaunkönig, which could ‘home’ on to the sound produced by the propellers of a ship travelling at between five and twenty-five knots. These torpedoes were known to the Allies as GNATs, German Naval Acoustic Torpedoes; they were first used operationally in September 1943 when the frigate Lagen and two merchant ships in convoy ONS18/ON202 were sunk. During the next six days three more merchant ships and two escorts were sunk by the new torpedoes, though the Germans lost three U-boats during the attack. Countermeasures were soon devised. First of all it was very slow for a torpedo – about twenty-five knots – and could therefore be avoided with alert lookouts; and then it was found that, at certain engine revolutions, the sound produced by the ship’s propellers made the homing of the torpedoes ineffective. Disturbed water and the explosions produced by dropping depth charges also attracted the torpedoes; and finally a ‘Foxer’ was devised: a noise-producing decoy which consisted of two lengths of steel piping which banged together when towed some distance astern of the escorts and on to which the GNAT homed and harmlessly exploded. The countermeasures were so effective that the GNAT torpedoes were soon withdrawn.
The next German device was a half-forgotten Dutch invention, the Schnorkel – now familiar to skin divers. An air trunk on the conning tower remained above the surface when the U-boat was fully submerged, enabling it to cruise under water powered by its diesel engines. There was a float-operated valve at the top of the schnorkel just above the water line, which automatically closed the trunking when the submarine dived; unfortunately sea swell, waves or a badly trimmed boat could also close the valve, which had a most unpleasant effect on the crew since the big diesels immediately sucked the air out of the hull, creating a partial vacuum which caused severe discomfort as the boat depressurised. Another unpleasant aspect of a schnorkelling submarine was the gale of freezing salt-laden air howling through the boat. There was also the uncomfortable feeling, while submerged, that the top of the schnorkel tube, with its tell-tale wake, was visible above the sea, inviting attack from aircraft whose 3-cm radar could detect even that small target. But the schnorkel did enable a submerged boat to travel more or less indefinitely under water at a far higher speed than possible on its electric motors. All newly constructed U-boats and many of the existing ones were fitted with schnorkels and they were used extensively when crossing the Bay of Biscay.
Useful though the schnorkel was, it was still a physical link with the surface: the boats that were fitted with it were still submersibles. But it was a step towards the true submarines which were now under construction in German shipyards. These were the revolutionary ‘Walter’ Type XVII boats. The main feature of this class was its very high underwater speed of twenty-five knots. This was achieved by a ‘closed-circuit’ system – that is, it was independent of external oxygen and therefore needed no schnorkel. The main propulsive motor was a Walter turbine, driven by gases created by the decomposition of a concentrated fuel of hydrogen peroxide called ‘Ingolin’ or ‘Perhydrol’. Unfortunately, not only was this fuel difficult and very costly to make, it was also highly unstable and dangerous, and the Walter engine required such a vast amount of it that the Type XVII U-boat would have been restricted to a range of about eighty miles at full speed.
None of the XVIIs became operational, though one of them, U1407, which was scuttled at Cuxhaven at the end of the war, was salvaged and went to the Royal Navy as HMS Meteorite. It was used to evaluate the Walter system for four years but was scrapped in 1950, the British considering it to be ‘highly dangerous’. Had the German Navy had time to develop the Walter boats, they could have proved highly successful; independent of outside air, they could have remained submerged for an indefinite period.
It is improbable that the Walter boats could have been developed in time to affect the outcome of the Atlantic battle, but another design, the Type XXI, first seriously considered in 1943, could well have done so, had it been put into production earlier. The Type XXI was known as the ‘electro submarine’ and had a much increased battery capacity, giving it a high underwater speed of sixteen knots. These large, very well-designed, 1800-ton submarines had a range of 11,000 miles and carried twenty-three torpedoes as standard. There was also a smaller coastal version, the XXIII, which displaced 256 tons and had a small 14-man crew.
The XXI U-boats were designed for mass production, being entirely welded and extensively prefabricated in eight major assemblies, much of the construction being done by semi-skilled labour. Had it proved possible to produce the Type XXI in sufficient numbers, they might well have prolonged the U-boat offensive; but it was not to be. Materials, manpower and supply communication were increasingly affected by Allied ‘round the clock’ bombing; many Type XXI U-boats were bombed under construction in the shipyards. Even if the planned construction programmes of 634 U-boats could have been completed, the 62,000 trained crew members they would have required were non-existent. From late 1944 onwards the German Army took priority over all else. Another factor which has rarely been mentioned was the extensive RAF mining of the Baltic, which seriously disrupted U-boat training.
