Bridges are among the most ancient and honourable members of society with a background rich in tradition and culture. For countless generations they have borne the burdens of the world and many of them have been great works of art.
Each of the five different types of bridge: beam, cantilever, arch, suspended structure and truss, can be related to forms occurring in nature. A fallen tree trunk across a waterway makes a beam bridge; horizontal tree branches in living trees are cantilever bridges; arch bridges are shaped by the natural weathering of rocks; and masses of vines in forests form ‘pathways in the sky’ for tree-dwelling wildlife.
The truss however, is the most sophisticated type, needing ingenuity to ensure adequate reliable connections. In nature, we find the essential members and tied joints of the truss in the ligaments and bones within the bodies of animals (see Figure 2.1).2
Figure 2.1: The truss structure in nature. Source: Guise (2006)3
Figure 2.2: Bridge types. Source: Jack Pulczynski (Unless otherwise indicated, all diagrams are the work of Jack Pulczynski)
Evolution Of The Truss
Technically, a truss is a special class of structure in which members are connected at joints allowing rotation, so each member carries either tension (pulling force) or compression (pushing force). A truss is a structure made up of triangles and just as there are many different ways to arrange triangles, so there are many different types of truss.
Truss principles were first used in the simple pitched roofs of ancient times, where the thrust of the rafters was provided for by the provision of a lower horizontal tie-beam. With the addition of inclined braces and a central king- post, such truss frames were used in Roman timber roofs.4
Development of the truss dates from the 2nd century trusses of Apollodorus (see Chapter 4 ‘Timbers and trusses’). In the 15th century, Leonardo da Vinci analysed the forces in triangulated structures and produced a design for a timber truss bridge.5 A century later, Palladio published illustrations of two timber truss bridges in his four books of architecture.6 Until the 19th century design was purely intuitive, based on experience. Even American engineers William Howe and Thomas and Caleb Pratt, who introduced the most significant truss developments, could not make accurate calculations of their systems.7 The first scientific discussion of the determination of stresses and proportioning truss members was Squire Whipple’s 1847 A Work on Bridge Building, with William Rankine’s Applied Mechanics published in 1858, a classic work on the theory of structures.8
The timber truss bridge is sometimes regarded as an American development because during the thirty years to 1840, a great deal of inventiveness, adaptation and refinement to the truss form took place in the United States of America. There were individual builders and bridge- building companies marketing bridges around the country, just like any other commercial product. Ethiel Town, for example, extolled the adaptability of his lattice truss by claiming to ‘make it by the mile and cut it off by the foot’.9
British practice was controlled, cautious and studious, producing expensive, conservative structures in contrast to the more entrepreneurial and experimental American approach, selling low cost bridges for the purpose at hand. Consequently, American bridges were perceived (and not entirely without reason) as cheap and nasty.10
Timber Truss Railway Bridges
Bridge engineering for the railways of New South Wales (NSW) from 1850 to 1915 had two eras of dominant technologies, British until 1890 and American after 1890.11 The British era covered the long term in office of the Yorkshire- trained ‘father of NSW railways’ John Whitton, who was ill-disposed both to American bridges, and to the use of timber for bridges.12 As a British colony, all NSW’ early engineers were educated in Britain and the skills they used in the Colony were direct copies or adaptations of British or European technology. Whitton in particular wanted to build railways to the grand scale of Britain, with his first two major bridges at Menangle (see Figure 2.3) and Penrith, both over the Nepean River, almost bankrupting the Colony.13
The very different approach of his successor Henry Deane, engineer- in-chief of NSW railways from 1890 to 1906, is suggested by Deane’s official visit to the USA in 1894 and his 1900 paper for the Institution of Civil Engineers in Britain, ‘Economical railway construction in New South Wales’.14
While the government’s response to Whitton was to decree that local materials be used as much as possible, he still succeeded in building some very expensive and remarkable railway bridges, such as those at Menangle and Picton (see Figures 2.3 and 2.4). But even Whitton had to face economic reality and his railway bridges include many kilometres of timber beam bridges, some laminated timber arch bridges (see Figure 2.5) and some timber truss bridges.
When Whitton sought funds for an iron bridge to carry the railway across the Hunter River at Singleton, he protested that “the House refused to vote money for it until I had laid the plans on the table to show that I intended to use timber not iron”. The laminated timber arch bridge built instead proved so costly to maintain it had to be replaced in 1902 by steel trusses.15 No laminated timber arches remain today, though some timber truss railway bridges remain.
Figure 2.3: The wrought iron cellular box girders of the 1863 railway bridge at Menangle. Source: DJ Fraser
Figure 2.4: Picton’s 1867 stone arch railway viaduct. Source: Wollondilly Heritage Centre & Museum
Figure 2.5: The laminated timber arch rail bridge at Parramatta, c1871. Source: ML, SLNSW
Figure 2.6: The Palladio-style timber truss spans of the 1889 railway bridge over Umaralla Creek at Chakola on the Goulburn to Cooma line in winter 1976, eight years before the line closed. Source: Jim Lunt photo, Australian Railways Historical Society NSW
Figure 2.7: Design for Tenterfield Creek railway bridge, with the train tracks laid above the queen post trusses. Source: Australian Railways Historical Society NSW
A set of timber trusses had been constructed over the Fish River at Gunning in 1875 for the Great Southern Railway, with striking similarities to Palladio’s design. Whitton successfully resisted their wider use for some years until three more sets were built in the 1880s, over the Bogan River at Nyngan, the Cudgegong River near Rylstone on the Mudgee line and over Umaralla Creek (see Figure 2.6). Now the only remaining of these bridges, Umaralla Creek bridge was described at the time as:
Five 42 feet [12.8 m] spans with timber over-trusses on squared pile trestles, and forty 26 feet [7.9 m] timber openings [ie beam spans]; also twelve 26 feet [7.9 m] timber openings detached from the main viaduct to give additional space for flood waters.16 These early railway timber trusses were in service for between 40 and 50 years and thus reasonably cost-effective, contrary to Whitton’s strong opinion. However, they soon became inadequate for the increasing weight and speed of locomotives and speed restrictions in turn reduced the efficient operation of the rail traffic.
A far more successful and durable set of railway trusses was built in the last three years of the railway boom, to 1889, the design adapted directly from British railway practice.17 Four sets of these queen post trusses were built between Glen Innes and Wallangarra, and two sets on the Cooma line south of Michelago. All six bridges remain today, though not in service. The original design for this truss type was described in the Railways’ 1888 annual report (see Figure 2.7):
The principal work of importance is the bridge over the Tenterfield Creek, at 288 miles 50 chains, which consists of six 42-inch [actually 42 feet, 12.8 m] timber under-trusses, resting on brick piers and abutments.18
In the earlier Palladio-style design used at Umaralla Creek the train ran between the trusses, a design more usual with road bridges. This meant only two trusses could be used, one on each side of the railway tracks, while the queen post design of the Sunnyside rail bridge over Tenterfield Creek allows an additional central truss. Today this rail bridge with three trusses per span is the only easily accessible queen post rail bridge, with most railway bridges obscured from view on private land distant from public roads. (see Figure 2.8).
Another beautiful example of the queen post rail bridge, but considerably more difficult to access, is the Yarraford rail bridge over Beardy River, also on this northern section of the line (see Figure 2.9).
Figure 2.8: Overtaken by the highway – a glimpse of Sunnyside railway bridge over Tenterfield Creek, 2013. Source: Amie Nicholas
With no timber truss rail bridges in service anywhere in NSW, these bridges can now been seen at work only in photographs, like those of historical re-enactments commemorating Sir Henry Parkes’ 1889 Tenterfield Oration (see Figure 2.11).
Exceptions to the use of British technology in railway bridge engineering before 1890 were the United States Whipple Truss used for Nowra’s 1886 metal railway bridge (for a railway line that never crossed the Shoalhaven River but ended at Bomaderry on the northern side) and the first Hawkesbury River metal railway bridge in 1889, constructed outside Whitton’s authority.19
Figure 2.9: The 1886 Yarraford rail bridge over the Beardy River, 2013. Source: Amie Nicholas
Figure 2.10: The large timbers of the three lines of trusses beneath Yarraford rail bridge that support the overhead rails. Source: Amie Nicholas
Figure 2.11: An historic locomotive with carriages crossing the railway bridge over the Severn River for memorial celebrations at Tenterfield. Source: Megan Alt photo, Glen Innes & District Historical Society
In the 1840s carpenters in the United States had experimented with and patented many forms of timber truss bridge. With scientific improvements and in-service experience, the two most successful were the Howe and Pratt trusses, United States types that became the dominant truss forms used in NSW. The feature distinguishing them is the arrangement of the sloping and vertical web members and the forces in them (see Figure 2.12). The diagonals in compression in the Howe Truss are well suited to stocky timber members, with vertical iron rods in tension. While the Pratt Truss is theoretically superior, the longer diagonals in tension and shorter verticals in compression made the connections in timber difficult and the Pratt Truss is more often metal rather than timber.20
Figure 2.12: The diagonals of the Howe Truss are in compression and the vertical iron rods in tension, while the Pratt Truss arrangement has longer diagonals in tension and shorter verticals in compression.
Figure 2.13: The 1908 railway bridge over Two Mile Creek at Walgett in 2013. Source: Amie Nicholas
In 1892, with the NSW main trunk railway lines completed, Whitton’s successor Henry Deane advocated the United States’ practice of constructing cheap developmental railway lines. Known locally as ‘Pioneer Lines’, the policy was for no major bridges to be built. Where rivers still had to be crossed, the design engineers used the Howe Truss, incorporating some local innovations from the successful timber truss road bridges. Two types were constructed, one with the trains again travelling between the trusses like the Colony’s first timber truss rail bridges, the other with the trains travelling above the trusses.21 The best example of the first type is at Walgett, and of the second, at Gundagai, the longest timber truss bridge in Australia (see Figures 2.13 & 2.14).
Figure 2.14: The 1903 railway viaduct over the Murrumbidgee floodplain at Gundagai, pictured here in 2013, is Australia’s longest timber truss bridge. Source: Amie Nicholas
Timber Truss Road Bridges
While the designs of railway bridges were influenced by the strong personalities and political connections involved in the development of the NSW railways, this was not the case with road bridge design in the Colony, where funding was always harder to secure. The story of the design of timber truss road bridges tells of an innovative application of skills and knowledge sharpened by the constant drive for economy. These designers could not afford to ignore the NSW hardwood timbers and decades of experience were distilled into the report of Public Works head Robert Hickson in 1897:
We must not lose sight of the fact that in New South Wales we possess the best timber in the world for bridge- building… There are cases, no doubt, where it would be more economical in the long run to erect iron or, better still, stone bridges; but I feel satisfied that for many years to come it will be found more advantageous to use timber in the construction of a large proportion of our bridges.22
The earlier timber truss bridges, designed between the 1850s and the 1880s, made use of vast forests of large, long, strong and durable NSW hardwoods. As discussed in Chapter 4 ‘Timbers and trusses’, once the comparative strength and durability of these hardwoods became known around the world, so much timber was exported that these earlier types of timber truss bridges could no longer be built. Bridges designed in the 1890s and 1900s still made use of the strength and durability of the local hardwoods, but timber sizes were limited to the smaller and shorter sections still readily available.
Between 1856 and 1936 over 400 timber truss road bridges were built in NSW, all of them designed by engineers of the Department of Public Works. For more on these men, see Chapter 3 ‘The designers’.
There was a clear evolution in design as later designers learned from the earlier, as knowledge of NSW hardwoods grew, and as the availability and economy of materials changed. Today, the timber truss road bridges are classified in five types according to their engineering design, each with the name of the designer (see Figure 2.15).
We must not lose sight of the fact that in NSW we possess the best timber in the world for bridge-building… There are cases, no doubt, where it would be more economical in the long run to erect iron or, better still, stone bridges; but I feel satisfied that for many years to come it will be found more advantageous to use timber in the construction of a large proportion of our bridges.
Figure 2.15: The five types of NSW timber truss bridge.
The Bennett Truss
Of approximately 150 Bennett Truss timber bridges built in NSW between 1858 and 1886, two remain in 2018. Plentiful high quality hardwood drove the design of these bridges, an example of innovative and practical engineering when large and long section timbers were readily available. Vast numbers of bridges were built, though budgets were tight and skilled workmen were few.
As Chief Engineer and Commissioner for Roads and Bridges from 1862 until 1889, William Bennett oversaw the construction of 6,000 miles (9,600 km) of main roads, 4,000 miles (6,400 km) of other roads and 40 miles (64 km) of bridges, mostly timber bridges, and also some quite spectacular metal bridges, some of which remain today.
But it was Bennett’s design for timber truss bridges that most heavily influenced the road bridge building practices into the next century. The Bennett Truss was designed for spans ranging from 55 to 100 feet (about 17 to 30 m). Based on Bennett’s 1865 analysis, the cost per square metre was half that of laminated timber arches and one third of the cost of an iron bridge. The cost advantage recommended it, but it was ingenuity that ensured the timber truss remained so long and so widely used for road bridges in NSW.
Irish-trained, Bennett was well- travelled and had no hesitation in applying technological advances from anywhere in the world to the needs of NSW.
At first glance, some of Bennett’s earlier truss designs might be considered a standard timber truss following the European design of Palladio (see Figures 2.16 & 2.17). A closer look reveals a number of differences. In the Howe Truss the most critical innovation was not the shape of the truss, but replacing the timber verticals with wrought iron rods with threaded ends. Howe was not the first American engineer to use iron components in a bridge, but he was the first to incorporate primary members made of iron in a predominantly timber truss, enabling adjustments if timber deformed over time.23 Bennett used this American technology to good effect in his Bennett Truss, while keeping the advantages of the Palladio-style geometry (see Figures 2.18 & 2.19).
Figure 2.16: Bennett’s design for the timber truss bridge built over the Pages River at Murrurundi in 1861. Source: Roads and Maritime
Figure 2.17: Palladio’s design for the 16th century Cismon Bridge in the mountains of north-eastern Italy. Source: Palladio 24
Figure 2.18: The Bennett Truss Monkerai Bridge over the Karuah River, completed in 1882 and still extant. Source: Roads and Maritime
The primary structural advantage of the shape of the Bennett truss is that it works not only as a pure truss, but also partly as an arch. When Bennett first started using local hardwoods, he noticed that they suffered from shrinkage and warping more than the American or European timbers.25
The use of larger timbers meant that the trees were older and stronger, and so the timber was considerably less subject to warping and excessive shrinkage. The arched shape of the Bennett truss allowed maximum benefit to be gained from the substantial capacity of the large section timbers forming the arch.
Bennett did not design his timber truss bridges for permanence like stone or iron structures, but for replacement, probably by iron bridges, on later road alignments as funds allowed.26 That the Bennett Truss had an average life of 54 years, with 26 bridges remaining in service beyond 80 years, would have well exceeded his expectations.27
While all the timber truss road bridges make use of cast iron shoes, Bennett showed the most care in making these an aesthetic feature. He had the least number of cast iron shoes of any truss type, with the lower ones always tear-drop shaped and the upper ones also distinctive.
Bennett’s shoes were specifically designed to frame the truss, both aesthetically and structurally, with the top chords, bottom chords and principals all being of the same dimensions. With these the most critical structural members, the shoes ensured strong connections as well as aesthetic highlights. Unfortunately, as later attempts to strengthen Bennett trusses have often changed the sizes of the timbers, the shoes don’t fit so well, as has happened at Monkerai (see Figure 2.20).
The excellence in design of the Bennett Truss proved and popularised the timber truss as the preferred form of bridge construction for medium span road bridges in NSW for more than seventy years. Two factors prompted a revision of the Bennett Truss design in 1886, which led to the McDonald Truss. The steadily increasing vehicle weights, from one tonne when Bennett first started to five tonnes on a pair of wheels, then to six and a half tonnes on four wheels by 1865, then tripling in the next twenty years, was the first, while new access to accurate information on the properties of timber was the second.28
Figure 2.19: Monkerai Bridge, shown here in 1968, is the second oldest Bennett Truss timber bridge remaining 50 years later. Source: Roads and Maritime
Figure 2.20: Bennett Truss Monkerai Bridge, in 2017 showing upper shoe and lower tear-drop shaped shoe
The McDonald Truss
Of approximately ninety McDonald Truss bridges built in NSW between 1886 and 1894, four remain in 2018. McDonald joined the Public Works Department in 1879, and although he specialised in metal bridges, he updated the standard timber truss design for heavier design loads (a distributed live load of 4kPa [kilopascals] and a traction engine weighing 16 tonnes).
Large, long, quality hardwoods were still plentiful and permanent stone or iron bridges were not considered economical. McDonald’s design changes were informed by the growing knowledge of timber as a structural material, with extensive testing at the University of Sydney in 1886.
As discussed more fully in Chapter 4, McDonald’s work with William Warren at his new timber testing laboratory gave him invaluable knowledge of the properties of Australian timber. McDonald did not greatly change the overall shape of the truss (see Figure 2.21), and retained Bennett’s innovation of combining the Palladio shape suited to large section timbers, with the Howe Truss tension rods.
Figure 2.21: Comparing the Bennett and McDonald trusses.
Figure 2.22: The bottom chords of the Bennett, McDonald and Allan trusses, sketched by Percy Allan for comparison. Source: Allan (1895)
The first part of the Bennett Truss that required strengthening for the heavier loads was the bottom chord. His design consisted of three rows of laminates and a small metal fish plate at each joint location, designed not to carry any loading but to provide a template to ensure correct bolting.
Designed for heavier vehicles, the bottom chord of the McDonald Truss was considerably wider than Bennett’s. It comprised four rows of laminates and two large central metal splice plates, designed to share the load. As discussed later, the double timber bottom chord of the Allan Truss had a gap between each element and a direct tension connection made up of four wrought iron plates with metal shear keys riveted to them, and then these plates bolted to the timber (see Figure 2.22).
Figure 2.23: The splayed principals of the 1893 McDonald Truss bridge over the Coolumbooka River at Crankies Plain near Bombala, 2013. Source: Amie Nicholas
Working closely with Bennett for ten years, McDonald understood that a change in bottom chord also required a change in principals and shoes. He returned to something which Bennett had tried much earlier, splaying the principals at the base to provide extra stability (see Figure 2.23). With this change, McDonald also did away with the tear-drop shaped shoe at the base (see Figure 2.24).
Because McDonald introduced the splayed principals, he no longer required the timber sway braces which were a feature of the Bennett Truss (see Figure 2.19) and replaced them with far more slender metal sway braces. Unlike Bennett’s timber sway braces, these were not designed to provide lateral support to the top chord, but only to resist excessive sway or vibration when heavy vehicles crossed the bridge.
Bennett had maximised the use of large section timbers in order to manage shrinkage and warping of timbers by applying the structural system of an arch within an arch. With better information on the capabilities of the NSW hardwoods, McDonald managed the shrinkage and warping by using double timbers bowed around timber spacers, with metal wedges at the bases of diagonals to take up the slack (see Figure 2.24).
In 1893 two McDonald trusses were constructed down the bottom of the long and windy Galston Gorge road to the north of Sydney (see Figure 2.25). A concrete bridge replaced the longer bridge in the 1930s, with Pearces Creek Bridge remaining in 2018 an example of the shortest of the standard McDonald Truss designs.
Figure 2.24: The Bennett (left) and McDonald (right) trusses for a typical medium span, with the arch shown in yellow and the Bennett Truss inner arch marked green. The cast iron shoes of each truss type are highlighted in red, and the metal wedges of the McDonald Truss are shown in green.
Figure 2.25: Two McDonald Truss bridges at Galston Gorge provided examples of the longest and shortest of the standard McDonald Truss designs, the former Berowra Creek Bridge at 90 feet (bottom) and the remaining Tunks/Pearces Creek at 65 feet (27.4 m and 19.8m). Source: WF Hall photo, Australian National Maritime Museum
The Allan Truss
Of over 100 Allan Truss timber bridges built between 1894 and 1929, twenty remain in 2018. The increasing difficulty in obtaining large section long timbers, as discussed in Chapter 4 ‘Timbers and trusses’, and the need for durable and maintainable bridges drove the design of these bridges.
Percy Allan joined the Department of Public Works as a cadet in 1878. As Chief Draftsman in 1893 when McDonald was retrenched, Allan made the most of the opportunity to revise completely the standard timber truss bridge design. He had worked with Bennett for over ten years and with McDonald for almost 15 years. He had worked briefly with Professor Warren, and also had had the benefit of a team of exceptional young engineers to assist in design development. With this considerable asset as well as access to information on the structural properties of hardwoods and to some 35 years of historical data on the performance of the previous timber truss designs, Allan made significant progress in the design of timber truss bridges.
Financial and resource constraints drove Allan’s bridge design. Funds were scarce and retrenchments imposed after the bank crashes in London in 1890 and then in the Australian colonies in 1892-3.29 As well, long and large section timbers were increasingly scarce and expensive.
Allan’s design focus was on using shorter, smaller section timbers designed for maintainability. To achieve this, he made fuller use of the Howe Truss geometry, a much simpler arrangement of triangulations compared with earlier trusses, and well suited to shorter, smaller timbers. Allan’s two important timber truss design innovations were replaceability of timber and of the bottom chord splice, as discussed below. This meant that the timber bridge could be more economical than an equivalent ‘permanent’ metal bridge not only for initial construction, but over the whole life of the bridge (see Figure 2.26).
Figure 2.26: Annual costs for iron and for timber bridges. Source: Allan (1895)
Allan’s design philosophy echoed 19th century architect, landscape gardener, and poet, Thomas Pope:
When Time, with hungry teeth, has wrought decay, Then what will sceptics be dispos’d to say? Why, “down the Bridge must fall, without repair, And all the author’s pleadings will be air.” Not so, he’s better arm’d than you’d expect, For nought can bring to ruin but neglect; A means provided, which can never fail, To keep up strength whate’er the Bridge may ail: Each log of wood, where’er its station be, Is safely shifted for a sounder tree…30
Replacement of timbers in the Allan Truss is considerably simpler, with Allan describing the process with earlier designs as something of a Chinese puzzle.31 However, the easier replacement was not achieved in quite the way Allan anticipated when he reported:
One of the features of the new type of truss is that any member can be renewed without staging from below, a matter of importance when deep gorges or fast running streams have to be crossed. Briefly stated, the top and bottom chords being in two pieces, the suspension rods are removed and re-arranged so as to throw the whole weight on one flitch; there being no strain on the remaining flitch, any member of the top and bottom chord can be replaced with sound timber; and by slacking the suspension rods and inserting temporary struts, any of the braces can be renewed, whilst the renewal of the cross girders is obviously a simple matter.32
As early as 1904 Allan’s colleague Harvey Dare reported that, “in almost every case the timber lower-chord has been the first member of the truss to fail, and the flitches, being in tension, are very difficult to replace”.33 The flitches of the top chords are also more difficult to replace than Allan’s enthusiastic report implied. The top chords are bowed to prevent warping and twisting, so a single flitch is not able to keep its shape and take the load without the other. Fortunately for Allan, other engineers invented methods of temporary support to replace the timbers.
Allan’s second key innovation was his splice connection in the bottom chord. Stronger than previous bottom chord connections and not needing such long lengths of timber, Allan’s splice connection was subsequently used by the railways (for example, for the Walgett bridge shown in Figure 2.13) and as far away as the United States.34 With the bottom chord splices so critical, Allan tested the full-size joint in a machine specially designed for the purpose (see Figure 2.28).
Figure 2.27: Typical 90’ (27.4 m) Allan Truss, showing elevation (centre), with plan of bowed top chord (above) and straight bottom chord (below).
Figure 2.28: Bottom chord testing at the Biloela Dockyard (Cockatoo Island). Source: Allan (1917)36
Allan’s bottom chord tension splice was a very clever innovation, achieving a stronger and more economical timber bottom chord than any previous timber truss design. This did not come without cost. The earlier laminated timber bottom chords did not fail suddenly without warning, but tended to stretch and sag over time until they were replaced. Allan’s new bottom chord tended to break suddenly under a heavy load, sometimes causing the entire bridge to collapse. This problem is hinted at by chief bridge engineer Ernest de Burgh as early as 1899, and made explicit by his successor Harvey Dare in 1904, with the first Allan Truss bridge less than ten years old.35 Although Allan had sought to improve durability over the previous designs, his bottom chord splices were subject to accelerated deterioration. The lighter weight and efficient design of the Allan Truss came at the cost of robustness.
In apparent contrast to both Bennett and McDonald, Allan aimed for the widest awareness of his achievements. As a result, there is far more recorded detail of the thinking behind his designs, including his own comparison of the design loads of previous timber truss bridges with his own, published in the Public Works Department annual report for 1893-94:
The design for truss bridges in use since 1884 [sic.] has been superseded by a truss of more modern design, the principal features of which are: the use of marketable lengths of timber, the adoption of open chords and braces always accessible to the brush, and the ease with which any defective timber can be replaced. In each of the new 90-feet spans there is a saving of 450 cubic feet of timber, while the trusses are capable of carrying 10 feet more roadway than in the old type of truss, thus affording greater travelling facilities at reduced cost. Not only is there a saving in materials in the new type of truss, but a considerable saving is effected owing to the shorter lengths of timber employed and the greater ease in framing together. Altogether the saving effected by the adoption of the new type of truss bridge is on the average about 20 per cent.37
There was no record remaining of the design load for the Bennett Truss, only that it was less than for the McDonald Truss. McDonald had designed his 90 foot truss for a 16 tonne traction engine plus a distributed live load of approximately 50 tonne, while Allan’s 90 foot truss accommodated a 16 tonne traction engine, but with a distributed live load of approximately 75 tonne.
There is an interesting reason for the additional design load in the Allan Truss, which Allan designed to accommodate a central roadway and two footways (see Figure 2.29), although this configuration was never actually used. Allan attempted to do away with the sway braces – which inevitably make the footways difficult – replacing them with what he called ‘wind stays’.
Since the ‘wind stays’ could not provide any lateral support to the top chord, Allan designed his top chord so lateral support would not be needed:
Figure 2.29: Allan’s sketch of his proposed layout for a 90 foot (27.4 m) Allan Truss with two footways. Source: Percy Allan’s Calculation Book, held by Roads and Maritime
Figure 2.30: Typical sections of a 90 foot (27.4 m) Allan Truss showing the early design without sway bracing (L) and the modified design (R).
The side braces adopted in previous trusses being a source of inconvenience when footways had to be provided, the author decided to design this chord as a column with a varying load, unsupported in a lateral direction; none of the text books consulted [gave an example] however … The author therefore has dealt with the case in what he submits, is a practical way of looking at the question…
Figure 2.31: Tharwa Bridge soon after completion in 1895, with short cross girders and no sway braces. Source: SARANSW
Some of the very earliest Allan Truss designs showed bridges with only short cross girders and no sway braces (see Figure 2.30), but those actually built to that design generally had to be modified shortly after to provide the necessary lateral stability to the truss.
The only bridge remaining today which was originally constructed with ‘wind stays’ instead of sway bracing is Tharwa Bridge over the Murrumbidgee River, opened in March 1895 (see Figures 2.31 & 2.32).
Figure 2.32: Tharwa Bridge in 1964 with ‘wind stays’, a unique and short-lived feature of the earliest Allan Truss. Source: Richard Collins photo, courtesy Colin Mackellar
Figure 2.33: Tharwa Bridge with chunky sway braces prior to reconstruction work in 2007. Source: Groz Klaric.
Figure 2.34: Tharwa Bridge with slender sway braces after completion of reconstruction work in 2011. Source: Roads and Maritime.
In 1965 a 25 tonne load limit was placed on Tharwa Bridge because of signs of deterioration, and it was probably around this time that the ‘wind stays’ were removed and rather chunky metal sway braces added. These remained until the bridge was rehabilitated in 2008-11, with a larger number of more slender sway braces installed on each of the trusses.
The most beautiful example of a modification to make up for lack of sway braces was the fourth Allan Truss bridge, completed in 1894 over the Dry River at Quaama on the south coast (see Figure 2.35).
The second Allan Truss bridge, over Stoney Creek near Bega, also opened in 1894; three years later Allan had to provide a design for rather awkward cross girder extensions and metal sway braces (see Figure 2.36).
Figure 2.35: The 1894 Dry River Bridge, with tastefully added overhead metal bracing to limit excessive lateral movement of the trusses. Source: NMA
As well as quickly reverting to sway braces like Bennett and McDonald, Allan found over time that he also had to strengthen his original top chord splice connection. For the earlier Allan Truss bridges, these splice connections consisted of simple metal plates and top chord splices were staggered (see Figure 2.32). By 1898 this design had been modified to two metal angles on each side of a central splice with a timber packer and a greatly increased number of bolts, giving a considerably stiffer connection (see Figure 2.37).
The other difficulty faced in the Allan Truss was the cast iron shoes. Bennett and McDonald had been very selective in their use of cast iron, limiting its use to carrying compression stresses, as cast iron is strong in compression but relatively weak in tension or bending. Allan greatly increased the number of cast iron shoes and relied on them for shear transfer. These items fairly commonly fracture, with Allan himself reporting in 1924 that at the Kempsey Bridge:
Figure 2.36: Plan for sway braces added to Stoney Creek Bridge in 1897. Source: Roads and Maritime
During the erection of trusses, the lugs of a defective cast iron shoe carrying the heel of the batter brace of truss was sheared off, and to avoid the risk of further defective shoes, built- up shoes of wrought steel were substituted for the cast iron shoes… and have proved quite efficient in service.38
The evidence challenges various claims of the superiority of the Allan Truss over all others. Allan was indeed an eminent and influential engineer who designed a very efficient and beautiful truss. With close evaluation of the Allan Truss and of Allan’s plentiful writings, it is clear that Allan succeeded because he was standing on the shoulders of giants, particularly Bennett but also McDonald, as well as working with a cohort of astute engineers.
Figure 2.37: The Allan Truss top chord splice of the 1927 Beryl Bridge over Wyaldra Creek, near Gulgong, in 2013. Source: Amie Nicholas
The de Burgh Truss
Eight of the approximately 20 de Burgh Truss timber bridges built between 1900 and 1905 remain in 2018. Difficulties with the Allan Truss bottom chords and the fact that materials other than timber had become increasingly available, as well as economy, drove the design of these bridges.
Ernest de Burgh joined the Public Works Department on survey and construction work in 1885 and went on to design and superintend the construction of bridges throughout NSW. After ten years as supervising bridge engineer, de Burgh was chief engineer for bridges from 1901-1903.
A primary reason for the new design was that, despite Allan’s attention to detail and significant innovations, in almost every case the timber bottom chord had been the first member of the Allan Truss to fail, and being in tension it was difficult to replace. As well, the lower chord timbers had to be of the best quality Ironbark but the extensive timber export trade made this difficult to obtain and increasingly expensive.39
The geometry of the new de Burgh Truss was based on the Pratt Truss, not the Howe Truss that Percy Allan had used (see Figure 2.12). Then chief engineer for bridge design, Allan signed a number of drawings on the same day – 2 April 1896 – showing early forms of Pratt trusses applied to NSW bridges (for an example see Figure 2.38). These designs were alternatives for a bridge over the Hunter River at Morpeth and one over the Turon River at Wallaby Rocks, but the Allan Truss was found to be cheaper to build at both places. The depth of the truss was the main structural difference between the early form of Pratt Truss signed by Allan and the later de Burgh Truss. The 1896 design was 10 feet (3 m) deep for the 90 feet (27.4 m) span, while the standard de Burgh Truss is 13 feet (4 m) deep for the 91 feet (27.7 m) span, a considerable improvement in efficiency. The 1896 designs also reverted to diagonal decking, and had the cross girder hung beneath the bottom chord rather than above, as well as other details that would probably have proved problematic had they been built. Nevertheless, these designs formed the basis of a much improved de Burgh Truss.
Within three years though, the shortage of bottom chord timbers and de Burgh’s design improvements made his de Burgh Truss more economical than the Allan Truss. Reporting on the first of the new bridges being built over the Queanbeyan River in 1899, de Burgh explained:
This bridge is rapidly nearing completion, and is of considerable interest, as, with the exception of the bridge over the Lachlan River at Cowra, it is the first in which the composite form of truss has been used in New South Wales, and also because the Pratt style of truss, with vertical posts and inclined tension members, has been adopted, in lieu of the Howe type, in order to obtain a stiff cross-section. The superiority of steel over timber in tension, and the great cost of replacing the timber chords, which, from their position, are the first portion of the truss to decay, points to a great economy in maintenance being effected by the use of this type for important bridges.40
While all five types of the NSW timber truss bridges are composite trusses, as they all contain timber primary compression members and metal primary tension members, the de Burgh Truss and the later Dare Truss extend this principle further than the first composite truss, the American Howe Truss. All of their primary tension members are metal, an idea first introduced by John McDonald.
Figure 2.38: The unbuilt 1896 Pratt Truss design for the bridge at Morpeth. Source: Roads and Maritime
The de Burgh Truss includes the greatest variety of materials found in any of the NSW timber truss bridges, with mass concrete and reinforced concrete piers, rolled steel bottom chords, cast steel washer blocks, wrought iron cross girders, cast iron anchor blocks, brass in bearings and, of course timber, for top chords, verticals, stringers and decks. Using each material to its best advantage in this way demonstrates excellence in design and in understanding.
For his third bridge de Burgh made a minor but obvious technical change, replacing the sloping end members with a conventional rectangular Pratt panel, giving the de Burgh Truss its characteristic squared ends (see Figures 2.39 & 2.40).
Figure 2.39: Sloping ends of the second de Burgh Truss bridge, in operation over the Macintyre River at Inverell from 1901 until 1983. Source: DJ Fraser
Harvey Dare was design engineer for many of the de Burgh Truss bridges including one built over the Wyong River on the NSW Central Coast in 1902 (see Figures 2.41 & 2.42). He explained his design intent:
The lower chords in Wyong Bridge are of the standard type adopted in this form of truss, viz., two steel plates 12 inches in depth, spaced 12 inches apart, laced together in the end bays, and connected at each apex with diaphragms and saddle-plates carrying the timber cross-girders. The vertical struts are of timber, each formed of two sawn pieces seated on the saddle-plates, and securely connected to the lower chords by extending the angle-bars of one of the diaphragms upwards, and bolting right through the verticals and cross-girder. The top chords consist of two sawn timbers free of heart, with a space of 4 inches between [actually 8 inches]. They are connected at each apex by a casting recessed 11⁄4 inch into the inner side of each flitch, for the full depth, and bolted through. The notching takes the horizontal component of the stress in the diagonal-rods, and the castings, acting as rigid distance-pieces connected to the vertical struts, prevent any tendency to twist on the part of the timber flitches, and ensure that the chords shall keep a good line. Wind-bracing, consisting of diagonal-rods with turn- buckles, is provided between the lower chords, and the top chords are stiffened against vibration by side stiffeners of T-section, connecting the chord with the cross-girders, which are extended outwards for that purpose. The diagonal-rods are of wrought iron, screwed at the upper end, and having an eye forged on the lower end, which is connected to the lower chord by a pin at each apex. The ends of the chords are seated on cast-iron bed-plates, a gun-metal or rolled-brass plate, 1⁄4 inch in thickness, being interposed loosely between the wrought-iron bearing-plate on the underside of the chord and the bed-plate at the expansion-end of each span.
Figure 2.40: Squared ends of the 1902 de Burgh Truss Gillies Bridge built over Black Creek in the Hunter Valley, in 2013. Source: Amie Nicholas
Figure 2.41: The de Burgh Truss Wyong Bridge in 1902. Source: PWDAR, 1902
Figure 2.42: The new bridge over the Wyong River at Wyong in 1902, demolished in 1967. Source: PWDAR, 1902
The Dare Truss
In 2018, 17 bridges remain of about 40 Dare Truss timber bridges built between 1905 and 1936. The design intent was to combine the best aspects of the de Burgh and Allan trusses, while avoiding their primary problems. With the Allan Truss, this was the tendency for the timber bottom chords to fail, while the expensive metal fabrication of the de Burgh Truss was a key problem arising from the Pratt Truss configuration and pinned connections. Of the five NSW truss types, the Dare Truss has the simplest geometry both overall and in its members, and also allows the easiest replacement of timbers.
Harvey Dare had joined the Public Works Department in 1889. In 1903, when he was in charge of highway bridge design, he took the opportunity to change the composite truss. The new Dare Truss returned to a Howe Truss configuration, with diagonal compression members and vertical tension members, but substituting a pair of steel channels for the timber bottom chord, and redesigning the connections to eliminate the pins of the de Burgh Truss. Dare did away with the bowed flitches, simplifying the geometry by designing with straight timbers only.
Figure 2.43: Comparing the 90’ Allan Truss (top) and the 91’ Dare Truss.
Dare did not return to the Allan Truss geometry, but had his own unique and very simple geometry of square panels and 45 degree angles (see Figure 2.43). As Allan’s bowed top chord did not perform the structural intention intended, Dare did away with the bow, simplifying fabrication not only of the timber but also of the cast iron shoes.
The bowed timbers in the diagonals and top chords of the Allan Truss as well as the larger number of diagonals can be seen in Figure 2.46, with the straight timbers in the diagonals and top chords of the Dare Truss evident in Figure 2.45.
Dare designed his Dare truss up to a maximum span of 104 feet (31.7 m), and there are four bridges remaining today with this span, two of which have been constructed on cast iron piers which originally carried Bennett trusses. These are Sportsmans Creek Bridge at Lawrence (see Figure 2.44) and Colemans Bridge over Leycester Creek at Lismore.
As discussed in Chapter 3 ‘The designers’, Dare was also heavily involved with the design of the longest span timber truss bridge built, the de Burgh Truss bridge over the Lane Cove River in Sydney.
In covering timber truss bridge design engineering in NSW from the 1850s until the 1930s, this chapter identifies an era categorised by the five distinctive truss designs created by Public Works engineers William Bennett, John McDonald, Percy Allan, Ernest de Burgh and Harvey Dare. There is however an unexpected postscript to the historical record.
After the new Main Roads Board took over most bridge construction in NSW from 1925, its first design engineer and later Commissioner for Main Roads, Howard Sherrard, produced an interesting and indeed unique timber truss bridge design.41 A Victorian appointed design engineer for the newly established NSW Main Roads Board in 1926, Sherrard became chief engineer and was the State’s Commissioner for Roads on his retirement in 1962.
Sherrard’s design for the Mill Creek Bridge near Wiseman’s Ferry is unique, blending elements of the timber truss bridge designs of both NSW and Victoria (see Figure 2.47). Completed in 1929, this bridge is still extant though now bypassed. Highlighting the design elements Sherrard combined, Mill Creek Bridge might be read as a tribute to the ingenuity represented in the development of colonial timber truss bridge design.
Figure 2.44: The 1911 Dare Truss Sportsmans Creek Bridge at Lawrence, in 2013. Source: Amie Nicholas
Figure 2.45: The 1921 Dare Truss New Buildings Bridge over the Towamba River, in 2013. Source: Amie Nicholas
Figure 2.46: The 1897 Allan Truss Wallaby Rocks Bridge over the Turon River at Sofala, 2013. Source: Amie Nicholas still in operation at 120 years of age. Source: Amie Nicholas
Figure 2.47: The 1929 Mill Creek Bridge near Wiseman’s Ferry. Source: Amie Nicholas