Chapter 9. Production: Building the Ships to Win the Battle

The British shipbuilding industry suffered badly during the series of depressions in the twenties and thirties, with the closure of about thirty shipyards. The smaller yards suffered most; Beardmore and Palmer were the only major yards to close completely. A drip-feed of orders had kept the thirteen specialist warship builders in business. With the start of rearmament, these specialists had become overloaded and it was clear that escorts would have to be built by yards unaccustomed to warship-building. The marine engineering industry had suffered even more badly and the capacity for key items such as reduction gearing, turbine blades and water-tube boilers was much smaller than in 1918. The Vickers-Armstrongs group was the only firm capable of producing major weapons, while armour production was of serious concern. Modern machine tools were scarce, as was welding equipment.

These cuts in the shipbuilding area had a number of knock-on effects. Perhaps the most serious was on recruitment. Shipbuilding was no longer the sort of job that attracted bright young men either at tradesman level or as graduates. This led to a very old-fashioned approach to new technology. There was a belief that the collapse of Beardmore was due to over-investment in new equipment followed by inability to keep up interest payments on the loans needed to finance this work. In consequence, there had been very little investment in any new equipment. The frontispiece of Buxton’s Big Gun Monitors shows the main slipway and fitting-out basin of John Brown’s yard in 1916 and 1941. The only change is the addition of a single tower crane in the later view.

Unions, concerned to protect the remaining jobs for their members, rigidly enforced many restrictive practices such as demarcation and tended to oppose new technology, particularly if it was intended to save labour. These practices led to strikes, further reducing any profits that might have fed back into investment in capital equipment. However, not all was doom and gloom; there were still sixty shipyards with double the output of their nearest overseas competitors and output per man was high.

Following the outbreak of war, the demand was overwhelmingly for numbers and, as explained in chapter 2, the only design that could be produced in numbers by non-specialist ship-and engine-builders was the Flower class; ninety ships were on order in the UK by September 1939, and a further sixty-four in Canada. The thirteen regular warship builders were fully occupied with major warships and were only able to contribute a few Black Swans and the prototype Loch. Swan Hunter built a few Castles and Lochs. There was a remarkable variation in the building times, both from ship to ship in the same yard and from yard to yard.

Loch Fada was built in conventional style by John Brown. Drawings were taken off so that engineering companies unused to shipbuilding could pre-fabricate sections for later ships.

Variation within a yard was sometimes due to bombing, which would also account for some of the variation between yards, while late delivery of equipment was a continuing problem It is interesting that later ships usually took longer than earlier ones, contrary to the expected benefits of the ‘learning curve’. The longer time reflects the combined effects of bombing, shortages of both labour and materials and of war-weariness. It has been suggested that malnutrition during the depression may have contributed. The more extensive electronic equipment in later ships was also a factor.

As described above, management, labour relations, trade union practices and capital equipment were all outdated, while the blackout and air raid alarms created further difficulties. At one meeting, a prominent shipbuilder, Sir James Lithgow (then Controller of Merchant Shipbuilding), said that shipbuilding practices were out of date. Stanley Goodall wryly noted in his diary, ‘Satan rebuking Sin.’¹ Despite all these problems, Buxton has shown that the remarkable building times achieved in the USA were at the cost of much greater man-hours per ship and possible only because of the vast resources of manpower and material available.²

A crude comparison of building times for various classes of escorts shows British yards in a bad light (see table 9.2). In particular, US dates are for commissioning, usually a few weeks before completion. These figures tell only part of the story, and that the unfavourable part. Man-hour figures are scarce but early US DEs required about one million man-hours, reducing to 600,000 with experience. In contrast, British-built Rivers needed 350,000–400,000 man-hours for a somewhat simpler ship.³ It is also noteworthy that US-built Liberty ships cost $1.78 million (equivalent then to £450,000), while a similar British-built Empire cost about £180,000.

Strikes were a serious problem, with about one million working days lost in 1944, over three times the 1939 figure. There seems to have been an idea that the war was being fought to preserve British liberties, including the right to strike in defence of trade union practices. There was probably some communist agitation prior to the German attack on the Soviet Union. Between 1938 and 1945, average wages in engineering roughly doubled from £3.50 to £7 per week. This did not go down well with the naval crews of vessels fitting out. A petty officer’s pay had gone up from £2.50 to £3.50 per week, with much longer hours (and no overtime pay), discomfort and considerable danger.

Official figures in thousands for employment in shipbuilding and repair for naval work are shown in table 9.3. Similar figures for marine engineering show 58,000 before the war, 88,900 in June 1943 and 80,500 in June 1945. Many engines were diverted from Lochs and Castles to the landing ship programme. The problem lay more in the balance between trades than in overall numbers. For example, the number of welders on naval work increased 80 per cent during the war. There was considerable pressure from the DNC department for ever more welding, which was lighter and stronger (at least if properly carried out). However, welding required new equipment that was not readily available in wartime Britain and, since few yards had the space or facilities for welded construction, in the short term it made sense to use the skilled riveters already available.

The shortage of electrical fitters got worse, as discussed under the Loch class programme. Installed electrical power is an indication of the complexity of the ship and, as table 9.4 indicates, is reflected in the overall cost. Degaussing, radar, heating and ventilation all made demands on electrical power generation and cabling.

There can be a considerable difference between the price paid to the shipbuilder (and engine manufacturer) and the total cost to the Admiralty, as many items of equipment were ordered centrally in bulk and supplied to the yard – even machinery in the later years. For a destroyer, the shipyard cost was about 80 per cent of the total. Ships were still ordered in accordance with an approved annual programme but changing priorities and availability of slips and of equipment caused many variations.

Some 27,000 women were recruited and though few were employed on production, they released men for heavier duties.

Production in the USA

Production in the USA had some, but not all, of the UK problems and these were less severe. There were no air raids – not even blackout – and labour was plentiful. There was an overriding intent that escort production should not interfere with the production of major warships down to, and including, destroyers. This meant using shipyards unaccustomed to warship work, and even many new yards on ‘green-field’ sites. In turn, this obliged the US Navy (including the US Maritime Commission) to fund a significant amount of capital equipment, even in the existing yards. There were severe bottlenecks in the supply of major components such as engines, gearboxes and 5in guns, etc. The results were outstanding.⁴

The first order came in November 1941 for fifty escorts for the RN under Lend-Lease. America’s entry into the war meant that the RN received only six of the original fifty, but this was more than compensated for by the eventual delivery of thirty-two diesel-electric (GMT) and forty-six turbo-electric ships (TE). Within a year, 1,005 DEs were on order, of which 563 from six different classes were completed (including fifty-six fast transports). The first were laid down in February 1942.

Initially, there was delay in starting as landing craft were given priority, but production built up rapidly from mid-1942. RN ships came mainly from Boston Navy Yard (GMT) and Bethlehem-Hingham (TE), though there were exceptions. Most RN Captains completed in the second half of 1943. There is a story that at Bethlehem-Hingham (which cost the US Navy $35 million), each ship taking shape on the long line of sixteen slipways was one week behind its neighbour and would be launched when its turn came, regardless of any outstanding work. At peak this yard employed 23,000 workers. There were a total of seventeen yards involved in the whole DE construction programme, of which the six biggest accounted for 73 per cent of the ships built. The new yard at Orange, Texas built ninety-three DEs, using 20,000 workers. Despite the slow start, the DE production output was invaluable, in the Pacific as well as the Atlantic.

Just in case 1,000 DEs were not enough, in June 1942 the US Maritime Commission suggested that there was some spare capacity in small merchant shipyards, including some on the Great Lakes. The Canadian-built River class was selected as a prototype (PF), with the structure modified by Gibbs and Cox for welded, prefabricated construction – and habitability to USN standards. Altogether, 100 were ordered, of which twenty-one joined the RN, all built by Walsh-Kaiser (Providence, RI). That yard averaged two and a half months from laying down to launch, with a further five months to completion. The ships were commissioned from the very end of 1943, with most joining in early 1944. As discussed in chapter 6, most had machinery teething problems, which delayed acceptance considerably.

US sources suggest that there was a structural weakness in the class. Since the Rivers had no such difficulties, it must be assumed that the PF problems were due to weld details.

The Canadian Contribution

Before the war, the RCN had only thirteen ships, mostly built in the UK, which grew to about 375 by the end of the war. The Canadian shipbuilding and supporting industries were tiny but they expanded very rapidly. For the Battle of the Atlantic, they built 60 Rivers for the RCN, with 2 for the USN and 8 for the RN. Then there were 97 RCN Flowers, 17 for the RN and 8 for the USN. This was not all, as Canadian yards built 4 Tribal class destroyers, 122 ocean minesweepers, 16 trawlers, 172 MLs, MTBs, etc., 26 Landing Ships Tank and 18 maintenance ships, not to mention a large number of Fort class merchant ships – a wonderful achievement.⁵ Numerous problems arose, mainly due to inexperienced staff, but these were all overcome thanks to the willingness and enthusiasm shown by all concerned. Pre-war experience in Canada was largely with machine work and it was difficult to develop the hand-working skills needed for ships’ machinery.

An aerial view of the Bethlehem-Hingham shipyard. The sixteen slipways built turboelectric Captains for the RN.

Initially, it was intended that Canada would build hulls and main machinery but rely on the UK for auxiliaries and armament. The first fourteen Flowers ‘completed’ without guns, and at least two crossed the Atlantic with wooden dummies. The earliest ships required some reworking in the UK. After the fall of France, this became increasingly difficult and more and more items were home-grown. It was an all-Canadian effort, though valuable guidance was given by the British technical team headed by the constructors Stanton and, later, Harrison. Canada even produced her own radar (SW-1 and SW-2), which was no worse than the contemporary RN Type 268. Overseeing was in the hands of local staff of the Classification Societies, Lloyd’s and British Corporation. These were not accustomed to Admiralty Standards and had to be convinced. Trials were a particular problem, as they could not see the need for lengthy and detailed records. In the UK, the specialist naval schools – gunnery, torpedo, etc. – played an important part in setting systems to work. The tiny RCN lacked comparable schools but played an increasing role as experience developed.

Canada had her own specific problems, such as the size of the country; it is 3,500 miles from Halifax to Victoria. Imagine the waste of time if an officer had to travel this distance by train. Moreover, winter temperatures caused many difficulties, including ice blocking the Saint Lawrence. Uncompleted ships were often towed downriver as winter approached. There was little experience in Canada of working from outline drawings, with each shipyard producing its own working drawings. Drawing offices were small, even non-existent, and their staff lacked the education and experience of UK draughtsmen. Detailed drawings had to be produced in the UK, which meant that at least one ship of the class had to have been built, delaying the start of the Canadian programme. Some of the drawings supplied were not fully corrected up to date. Once complete, the drawings had to come by sea, as there was no regular airmail until 1941. The locks on the canal from the Great Lakes to the Saint Lawrence were only 270 feet long, so that frigates could not be built on the lakes.

Main engines were built by five firms. Initially, one was not satisfactory – it was found that only one man in the company had worked on marine engines before – but they learned quickly. Details could cause serious problems. In the early days there were problems with brittle bolts, due, it was found, to the use of steel with high sulphur and phosphorus content. Everything was rushed and much of the steel used in the hulls was still covered in mill scale. Normally, plates and sections would have been stored in the open for some time and the scale would have fallen off, but in war there was no time and ships completed still covered in scale. This would be painted over, and both scale and paint would fall off at sea, giving the scruffy look wrongly attributed by the RN to ill discipline.

It was hoped that many auxiliaries would come from the UK but they were in short supply there, too, and some were lost in transit. Valves were a particular problem. Gyrocompasses were unobtainable and most Canadian Flowers had to rely on a single magnetic compass (without repeaters) until late in the war. This caused navigational problems, particularly after exposure to shock, but worse, the gyrocompass formed part of the asdic control system and the magnetic unit was less reliable. Eventually the problems were overcome and Canada became self-supporting.

Canada was even less well positioned to carry out refits. For various reasons it was mid-1942 before it was decided to update the Canadian Flowers and work did not commence till the beginning of 1943. The work involved lengthening the forecastle, a new bridge, Hedgehog and associated Type 144 asdic. The latter involved big changes to the electrical supply and distribution systems. The package took about twenty-two weeks in Canada, compared with fifteen weeks in the UK and ten in the USA.

All these ships had to be manned with trained crews, a problem discussed in chapter 3. Initially, Canadian ships went to sea built by inexperienced yards, lacking important equipments and with inexperienced crews. It is no wonder that there were problems at first but, more notable is the speed with which such problems were overcome to make Canada’s contribution to victory.

Australian ships played no significant part in the Battle of the Atlantic but, starting in 1939, the RAN promised to supply twelve A/S officers and twelve SD (special duties) ratings every two months, and numbers subsequently increased.⁶

Availability

For various reasons, a considerable number of escort vessels were not available at any given time. There were material problems, such as boiler cleaning, general maintenance, repairs, etc., and the crew needed rest and, if possible, leave. More and more time was devoted to training, both individually and for the complete group. Table 9.5 shows the results of a post-war analysis of the non-availability of escorts. The impact of damage during the evacuations of 1940 is apparent, while one may also note the effect of weather damage in the winter months. The early Flowers often completed with poorly aligned crankshafts, leading to early bearing failure. The older British destroyers had no intrinsic problems but old age led to continual difficulties with leaking rivets, making life unpleasant for the crew and causing contamination in both fuel and feed-water. The Towns suffered from incurable ‘con-denseritis’ as the tube plates were weak and not parallel, from bearing problems due to corrosion of the cast-iron housings and from leaky rivets and a bridge structure too weak to withstand the impact of heavy seas.

Neither the diesel nor the turbine Captains had any serious maintenance problems, showing that a complex system is not necessarily unreliable. As discussed, the Colony class had initial machinery problems but seem to have given few troubles in service.

Note: The overall average was 22.0 per cent

The Canadian River class Swansea in heavy seas. It is probably Sea State 7 with waves up to 25ft high (estimating wave heights from photos is not easy). She seems to be riding them well.