Timber Truss Technology

One of the first timber constructions of the British colony of New South Wales (NSW) was the bridge across the Tank Stream in 1788, most likely using Sydney blue gum. This was soon found to be the most satisfactory timber for building and after two centuries remains a respectable choice for structural design in timber. However, it was roof trusses that were the most needed and common type of timber construction in the early years at Sydney Cove – one of the most successful examples can still be seen in the king post roof trusses of Sydney’s 1819 Hyde Park Barracks, with spans of over thirteen metres.

Engineering textbooks available in the Colony’s first decades were mostly about road and masonry construction and of little help where buildings were the priority and timber the main material available. By the 1850s the first technical guidance on calculating stresses and proportions of timber trusses had been published in the United States by Squire Whipple and in 1862 A Manual of Civil Engineering was produced by the University of Glasgow’s Professor WJM Rankine. The latter included details of a type of truss ‘first introduced in America by Mr Howe’ with iron rods as the vertical tension members.1

Long before manuals like these helped disseminate and democratise civil engineering knowledge, this need was identified as the chief objective of Britain’s Institution of Civil Engineers (ICE). Founded in London in 1818, the new body soon also turned its attention to the qualifications and training needed to practise civil engineering. Endorsed with a Royal Charter, the ICE set up the grades of ‘associate’ and ‘member’, the latter for those with seniority in the profession. Like other professions, the first step was a training period spent articled to an approved employer, followed by some years gaining suitable experience on civil engineering works and finally an admission interview. In 1857, 33-year-old William Bennett gained the Associate Membership of the Institution of Civil Engineers (AMICE) this way, and the same year was appointed to the Public Works Department in NSW. Five years later he was the colony’s Chief Engineer for Roads and Bridges (see Chapter 3, ‘The designers’).

In 1827, a decade after the founding of the ICE, the University of London introduced subjects for an engineering degree and established the world’s first engineering teaching laboratory. However it was not until 1883 that the University of Sydney, founded in 1850, started teaching in engineering, following the appointment of William Warren. Before then candidates took advantage of instruction at the colonial mechanics’ institutes.

Once universities offered courses in engineering, these could be completed in preparation for examinations other than those leading to a degree. Towards the end of Bennett’s long career, colonial universities first offered courses in engineering that included study for non-degree examinations. Among those recorded in the 1885 University of Sydney Calendar as ‘not passing through the regular course’ were Bennett’s first cadet engineer, Percy Allan, who had joined the Department in 1878 at seventeen, and his Public Works colleague Leslie (LAB) Wade who was appointed in 1880. Bennett’s long career ended before Warren’s first students led the new graduate path into the engineering profession, among them bridge engineering recruits Henry Harvey Dare and John Job Crew Bradfield.

By then the engineering profession in NSW reflected the development of the profession in Britain, but with the strong influence of local needs. Two already qualified bridge engineers who came to NSW were John McDonald, a graduate of King’s College at the University of London, and Ernest Macartney de Burgh, of Dublin’s Royal College of Science for Ireland. Like Bennett, Allan, and Dare, their names are also now part of Australia’s timber truss bridge history.

Colonial engineers had founded the Engineering Association of NSW in 1870, with Minutes of Proceedings published from 1885. The University of Sydney’s new Engineering Society Journal and Abstract of Proceedings first appeared in 1896 and engineering papers were also published in the Journal and Proceedings of the Royal Society of New South Wales. This enabled an invaluable exchange of knowledge on structural timber and timber bridges, readily available to us today through the digitisation of these publications.

The colonial engineers also drew on far earlier works, standing on the shoulders of giants like 17th century Isaac Newton in England and 16th century Venetian Andrea Palladio. As noted in Chapter 2, ‘Developing the truss’, though deservedly known as one of the world’s greatest architects, Palladio’s contribution to the field of structural engineering was just as impressive. The Italian peninsula has an even earlier example of the use of timber trusses though, in Apollodorus, technical advisor to Trajan, the Roman emperor who pushed the Empire to its greatest extent at the beginning of the second century AD.

Trajan’s two campaigns to conquer Dacia (approximately, modern Romania) are commemorated in an extraordinary masonry illustration, the chronologically ascending helix of sculptures in remarkably fine detail on Trajan’s Column in Rome. Near the base of the column is depicted the first invasion across the Danube River in AD101, achieved by a bridge of boats, using prefabricated timber trusses (see Figure 4.1).2 The northernmost span is visible, revealing what is recognisable to us as the truss design patented by Britain’s James Warren in 1848.

In the basic Warren design, the inclined members of the truss slope in alternate, opposite directions, with the structural design options of vertical members at every possible panel point, or only in alternate panels. The Danube boat bridge shows the first option, while further up the column there is a small single-span bridge, with vertical members only in alternate panels. That the conscientious sculptors depicted the difference is evidence that they, like Apollodorus and the military builders, knew what they were doing.

The timber truss bridge is another legacy from the Romans that, like concrete and central heating, disappeared in the Dark Ages only to be rediscovered after a millennium or more.

The Warren Truss was never taken up in Australia for timber bridges, though it was for steel. Connoisseurs of timber truss bridges will be provoked as to why the Warren Truss was ‘the one that got away’, particularly as it is now much favoured in countries building 21st century timber truss bridges.

Before Warren completed his design, across the Atlantic Massachusetts bridge builders William Howe and the Pratt brothers patented separate truss designs in 1840 and 1844 respectively. As Chapter 2 notes, it was from those truss arrangements that Percy Allan and Ernest de Burgh derived their Australian trusses.

At this time, scientific experimentation informed engineering practice, an influence that was not readily welcomed. In his contempt for mathematics in engineering, legendary Scottish civil engineer Thomas Telford (1757 – 1834) was expressing a contemporary professional view that the experience and judgement of practical engineers was pre- eminent and the mathematician inconsequential. The conflict was with theoretical calculation, exemplified for instance in the work in Europe of Charles- Augustin Coulomb and Leonhard Euler, 18th century pioneers of the science and engineering of structural mechanics.

It was a losing argument, not least because the increasing availability of cast iron and wrought iron at competitive prices meant a pressing need for precise cost, quantity and performance comparisons. Questions that had to be answered included: how does a wrought iron or a cast iron beam measure up to the carrying capacity of a timber beam? By how much was cast iron so obviously weaker in tension than in compression? By the mid-19th century, the need to know had triggered a series of testing programs using newly devised machinery. The arrangements of levers and weights that comprised the first machines used by two Scottish engineers in England, William Fairbairn (1789 – 1874) and David Kirkaldy (1820 – 1897), though primitive, generated useful numbers nonetheless.

Even these primitive devices were not available in the Australian colonies, and improvisations developed using materials to hand. In particular, the extremely strong native hardwoods were soon in use beyond roof trusses. For instance in simple beam bridges. From the 1850s various testing was done in NSW Public Works’ Roads and Railways branches to discover bending strengths.

In Van Diemen’s Land (now Tasmania) in 1851, forester James Mitchell had devised tests for three local species. Except possibly for the Stringybark, these were less applicable in NSW, but Mitchell’s procedure is worth noting. He tested a simply supported beam of seven feet (2.13 m) in span and measured central deflections to a taut string between the ends. Instead of simply increasing the central weight to failure, the total load was removed after every increment and the return of the centre of the beam towards its starting point was noted. Mitchell thus demonstrated an appreciation of the hysteresis, or cyclic loading, behaviour of the material. His approach might sound primitive, but more than a hundred years later, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) guidelines had only four strength levels for timber design, which Mitchell would have been able to target with acceptable accuracy.3

The work of inventors and industrialists Joseph Bramah and William Armstrong in Britain had demonstrated the applications of hydraulic power, and David Kirkaldy reasoned that a hydraulic mechanism would be a superior means of supplying, controlling and measuring force in a materials testing machine. With his motto ‘facts not opinions’, he was not likely to leave anything unproved; he designed a monster testing machine 14.5 metres long which was later used for applying loads in excess of 300 tonnes. Kirkaldy commissioned Leeds firm Greenwood and Batley to build his machine but, impatient at the slow progress, he completed the manufacturing himself and the machine started operating in 1866. Hydraulic power was later supplied by the London Hydraulic Power Company using pressure intensifiers. The testing laboratory in London’s Southwark based on this machine earned a worldwide reputation over its 99 years of operation; the machine is now in an industrial museum.

With the success of Kirkaldy’s operation, and under the terms of their contract with him, Greenwood and Batley produced smaller and more affordable replicas of the original machine.

In 1884, The Engineer, a leading journal with many subscribers in the Australian colonies, published details of a Greenwood and Batley machine recently purchased by the University of London, with dimensions about two-thirds those of the Southwark original.4

In 1882, while the University of Melbourne was considering appointing a professor of engineering, the University of Sydney ended a debate of many years by appointing as lecturer in engineering William Warren. Then a 30-year-old on the staff of the Roads and Bridges Branch of the NSW Department of Public Works, Warren had been born in Bristol and brought up in Wolverton, Buckinghamshire before securing an engineering apprenticeship with the London and North Western Railway Company. Outstanding results in the examinations of the local technical institute won him a one-year scholarship to the Royal College of Science for Ireland, where a decade later Ernest de Burgh became an undergraduate. The prestigious Whitworth Scholarship enabled Warren to attend Owens College, Manchester and he began his engineering career in that busy industrial area. Warren migrated to NSW in 1881 and was appointed as a draftsman in William Bennett’s Roads and Bridges branch of the Department of Public Works.

Taking up his prestigious challenge at the University of Sydney just two years later, Warren was faced not only with devising a curriculum, but also with sourcing equipment for the engineering building under construction on the northern side of the campus. For him, understanding the behaviour of materials was basic to engineering design and a machine for testing the properties of materials was his priority.

In 1884, the same year The Engineer reported the University of London’s purchase of the Greenwood and Batley testing machine, Warren had one on order. The 100,000 pound (445 kN) capacity universal (implying a range of testing capabilities) testing machine arrived at the still unfinished engineering building in 1885 (see Figures 4.2 and 4.3).5

Warren’s short time on Bennett’s staff showed him that one material in particular had priority for investigation to yield data of immediate practical application. The generally high quality of NSW timbers concealed considerable variability in strength and other properties across species. There was no question about the first program and the new machine started timber-testing at the beginning of 1886. As well as acknowledging the research assistance of students like Harvey Dare, Warren credited Public Works bridge engineer John McDonald as co-inventor of an autographic device for recording deflections (see Figure 4.4). It was a fruitful collaboration, McDonald’s help in the laboratory affording him test results hot off the bench and the first McDonald Truss timber bridges appeared that same year.

Testing progress during 1886 also enabled Warren to present the preliminary properties of Ironbark to the Royal Society of NSW at the end of that first year. The full results of tension, compression and bending tests were published by the NSW Government in 1887 (see Figure 4.5).7 In all, 22 species were tested covering a wide range of applications specific to the properties of each. As timber column and timber joist construction in Sydney continued over the next thirty years, the data would have been as invaluable to the building industry as it was for bridge design. Warren’s results confirmed early colonists’ rejection of some building timbers, for instance Red gum, with its poor and variable performance in his results eliminating it as a candidate for further testing.

Another significant publication was a 1911 NSW Government report that covered a greatly expanded testing program where Warren had access to additional machines. These results provided data for Australian structural designers until the testing and design guidelines of the CSIRO were published in an accessible form in the 1950s. The value of Warren’s efforts is indicated by the number of his reports and articles now being published online, more than a century later.

In recent years, the University of Sydney obtained from Roads and Maritime Services NSW some samples of Ironbark taken from bridges that had completed their service after several decades. Once the sometimes decayed surface layer was shaved away, the main material had usually survived in excellent condition. Strength testing confirmed information in textbooks that the range of strength within a particular species could spread over a ratio of two to one from the highest strength to the lowest, as might be expected from a natural product. However, we can take comfort from the fact that a very high margin of safety is used in conventional timber design: much more than in, say, steel structures or aircraft design where source materials are more consistent.

The Greenwood and Batley machine continued in operation in three different locations within the University of Sydney for over eighty years; it is now part of the Museum of Applied Arts and Sciences collection in Sydney.

Engineering is the art of modelling materials we do not wholly understand, into shapes we cannot precisely analyse so as to withstand forces we cannot properly assess, in such a way that the public has no reason to suspect the extent of our ignorance.8

This quote, familiar to engineers, seems truest when talking about timber. Timber is so different from most other materials that structural engineers use simply because it is organic. Unlike steel and concrete, where there is a high level of control in the composition of the material (for instance, the percentage of carbon in steel, and specification of admixtures in concrete), timber comes as it grows, and every piece of timber is different. Even along its length, a single piece of timber has significant variations in properties. Timber is highly anisotropic (directionally dependent), being significantly weaker when loaded across the grain rather than along it.

Timber also has a tendency to creep under sustained load, and to shrink under changing moisture conditions.

However, the hardwood timbers of NSW are very strong. In 1896, young Public Works engineer JJC Bradfield reported on the comparative strength of Ironbark and iron. He found that, for the same weight, Ironbark is more than three times stronger than iron in tension, and almost twice as strong as iron in compression.

The first timber truss bridges in NSW were generally constructed of Ironbark for truss members and Tallowwood for the deck. As Ironbark has become more difficult to obtain, many truss members have since been replaced with other species such as Tallowwood, Blackbutt, Grey gum, White mahogany and Spotted gum. Turpentine is used for piers, especially those in water.

Timber is generally categorised into hardwoods and softwoods. Some hardwoods are actually quite soft, while some softwoods are comparatively hard, because the categorisation is based on the cell structure and the presence of ‘vessels’ or ‘pores’, rather than the hardness of the timber. Due to its cell structure, timber differs in its strength, stiffness and shrinkage properties in the three directions corresponding to the radial, tangential and longitudinal directions (see Figure 4.6).

The wood of the tree is usually differentiated into two distinct zones, outer sapwood and inner heartwood (see Figure 4.7). Although the sapwood has the same strength as the heartwood, it does not have the same durability. Toxic treatments that can be given to sapwood to improve its durability are increasingly being limited by legislation.

The primary agents of wood deterioration are rot, termites and fire. Rot is caused by fungi which live within wood. Termites are insects that generally live in colonies in wood. Both termites and fungi are a major hazard to timber in bridges. Fire is also a significant hazard for bridges, although the large section timbers from which bridges are constructed tend to char slowly in a fire, and some timber bridges have therefore remained serviceable after fires.

When Europeans first explored Australian coasts, they were less than impressed by the timbers. James Cook said the trees observed on the coast in 1770 were so ‘hard and ponderous’ that they were pretty much useless. Surgeon John White reported in 1790:

I do not know any one purpose for which it will answer except for firewood; and for that it is excellent; but in other respects it is the worst wood that any country or climate ever produced.11

Reports from the first decades of the colony describe the difficulties the convicts had in dealing with the timbers due to their ‘monstrous bulk’, hardness and incredible weight. The trees in the immediate vicinity of the settlement at Sydney were too crooked, too hard to work, and too damaged by fire to be used as a structural material.12

However, it wasn’t long before timbers were discovered which would rival any in the world. When Red cedar was discovered on the flats of the Hawkesbury River, gangs of convicts were immediately sent to cut them down. As early as 1795, sixty logs from the Hawkesbury were exported to India, followed by loads to England, China, South Africa and New Zealand. In the colony the timber was used to build houses, barns, rough bush furniture as well as very fine furniture. It built pigsties, cow sheds, paling fences, railway sleepers and, of course, bridges.13

The duty of inspecting exported timber fell to the Department of Public Works; whatever the advisability of sending off large quantities of our best timbers, such exports had to be properly inspected and classed. By 1904, the rapid disappearance of hardwoods accelerated due to the recognition of its value by the commercial world of Europe, South Africa, and the East. In 1907 it was reported that excessive exports had greatly increased the price of timber, and unless the trade were slowed, national works would be seriously handicapped. Little was done to replenish the enormous losses to meet the constant demand for railway sleepers, bridges, wharves, and building timbers.16

A great range of processes is available to convert logs into marketable timber and timber products. These processes can vary significantly in their form and sophistication. For most operations, processing usually starts in the forest, with the output and end products from the forest being dictated by demand, species and log quality. Once felled, trees will normally be assessed, graded and docked into logs for conversion using one of the available processes.

Today, Forests NSW manages the two million hectares of state forests sustainably to balance timber production, recreational activities and the conservation of wildlife. The large section, old growth, native hardwood timber used in bridge construction is very difficult to obtain from these sustainably managed state forests, due to the rarity of trees of the correct species of sufficient height and age. Moreover, even when suitable logs exist, they are often cut into small marketable lengths for buildings, rather than set aside for use in timber bridges.

Timber, like any other structural material, has a tendency to deteriorate with time. If timber is protected from water, UV exposure and termites, it can last for centuries, but the timber in bridges in NSW is very exposed to all these agents of deterioration. The longitudinal sheeting on which vehicles drive lasts on average seven years before having to be replaced with new timbers. Round timber girders frequently found on approach spans last on average thirty years. Truss members such as diagonals on an Allan or Dare Truss can last up to fifty years. Timber piles are also very susceptible to rot and termite attacks in the region just below the ground surface, so installation of new piles is a critical and regular aspect of bridge maintenance.

The early forms of timber truss bridges were designed to be replaced rather than maintained, and this is generally what happened. Often, before a bridge was replaced, it would be under- trussed with steel cables as a temporary measure in order to allow traffic to cross while a new (generally concrete) bridge was being designed and constructed (see Figure 4.9).

When there were insufficient funds for a new bridge, temporary support systems were designed and used so that a piece of timber could be removed and replaced without the bridge collapsing. The Allan Truss as well as de Burgh and Dare trusses were specifically designed so that members could be replaced, thereby considerably lengthening the expected life of the bridge.

One example of this was at Wee Jasper in 2005, where the 1896 Allan trusses had deteriorated. Instead of simply replacing the deteriorated members as would have been done in the past, a Bailey bridge was installed, the whole trusses were removed, and two brand new, prefabricated trusses were placed into position by a crane (see Figure 4.11).

The opportunity was also taken to change some of the design details and materials in order to strengthen the bridge for modern vehicles. Similarly, at the Junction Bridge over the Tumut River the 1893 McDonald trusses had severely deteriorated, and rather than replacing each member in-situ, a new temporary bridge was constructed beside the original, the whole spans were removed with a crane, and new, strengthened spans were fabricated next to the bridge away from traffic and lifted into position with a crane (see Figure 4.12). Again, the opportunity was taken to modify details for additional strength.

Figure 4.16: A register of foundry marks on Dare Truss bridges. Source: Amie Nicholas

As discussed, Professor Warren ensured that engineering students at the University of Sydney studied timber in detail as well as cast iron, wrought iron, steel, masonry and concrete. However, as steel and concrete became more readily available, and the timber resource diminished, timber was less a priority for both the University and the Public Works Department. By the 1930s, bridge design engineers’ great enthusiasm for concrete bridges meant that very few new timber truss bridges were built. In addition to strength and durability, bridge designers of the 1930s thought concrete provided ‘an impressive series of plane surfaces of pleasing and uniform texture, bounded by sharp and true lines’, while ‘the greatest drawback in the use of timber is the impossibility of incorporating any graceful curves. Any curves appearing are all of the wrong kind’.18

Timber still had some support, however, with William Attwood warning in 1932:

This generation of engineers may well ask themselves whether in their frequent disregard of the oldest and one of the best of the available structural materials, they are worthily ‘carrying the torch of progress’.19

Although timber design was still taught in the universities, its scope was limited to roof trusses and smaller constructions, while design of larger engineering works such as dams and bridges focused on concrete and steel. The focus of the 1939 Handbook of Structural Timber Design on the structural design of timber buildings was followed in the CSIRO’s 1958 Timber Engineering Design Handbook.20 Bridge design codes like the 1976 National Association of Australian State Road Authorities (NAASRA) Bridge Design Specification simply referred to these other documents. With the introduction of the AUSTROADS bridge design code in 1992, even the reference to timber bridges was dropped, so very rare was their occurrence by then.

By 2000, there was no current design code for timber bridges, with the old codes so excessively conservative in some parts, and so much the opposite in others, they led to inefficiency and uncertainty in design. As well, most university engineering courses focused on steel and concrete with little to no coverage of timber as a structural material, robbing designers of an authoritative basis for their work. Designs became largely experimental, depending upon ‘rules of thumb’, ‘engineering judgement’, ‘experience’ and ‘precedent’, with more trials than results.

This is now changing, with knowledge of timber as a structural material undergoing a renaissance. Bridges built in timber are enjoying a significant revival around the world, both for pedestrian and vehicular traffic. Among the influences are a growing interest in the environmental problem of reducing carbon emissions, new and innovative use of timber such as stress laminated timber decks, as well as better connections and engineered materials. The fact that reinforced concrete did not turn out to be as durable as first thought is another factor, as many countries experience serious issues with concrete bridges less than 50 years old.

Timber’s high strength-to-weight ratio, its environmental sustainability, its ability to capture and store carbon, and its aesthetic appeal, combined with the ease and speed of construction in the off-site prefabrication methods now used, makes timber again the material of choice for many situations. Centuries of experience in the use of timber for bridges coupled with extensive research around the world over the past 25 years has provided the knowledge required to design and construct safe, strong, durable and beautiful modern timber bridges.

In 2010, the Australian Standard for design of timber structures (AS1720.1) was updated, bringing it into line with the modern design techniques. In 2011, research on timber truss bridges provided a reliable method for estimating the capacity of compression members in heritage timber bridges, giving designers a theoretically sound basis for design and assessment of such bridges.

The words of botanist JH Maiden, who pioneered the study of Australian timbers, have a new resonance 120 years later:

Ironbark stands alone as the embodiment of the combination of a number of qualities valued in timber, viz., hardness, strength, and durability…one of the main reasons why colonial timbers are not more used is because users are nervous through ignorance…I plead for a wider interest to be taken in our trees and our timbers.22