Chapter 2.

2. Why Do Timber Frames Fail? 1/5.

The fundamental question as to why are ironwork repairs important enough to warrant study can be answered reasonably simply. Iron-work repairs represent the pinnacle of low-tech, minimum intervention repairs to failed timber joints and thus they highlight the inherent failings of the timber, its fabric, construction and properties and the efforts that the craftsmen went to in overcoming its failings.

There are several reasons why timber frames fail and these are summarised following.

2.1 Inadequate design - lack of integrity - lack of bearing faces - lack of pegs.

Given the diversity of joints in timber framing and their steady evolution as highlighted by Cecil Hewett in English Historic Carpentry it is evident that some joints served their purpose better than others. Unsatisfactory jointing techniques were abandoned or rapidly reconfigured to function satisfactorily. There is a specific difficulty in finding early unsatisfactory work still in-situ as its very nature means that the building has failed and been remodelled. Vestiges may, however, be visible.

In the mediaeval period, from the 11th century in this country, virtually all these joints were held fast with wooden pegs. The exceptions are those that relied on the weight of the timber and the structure above to keep them in place. Thus we find that mortice and tenon joints, dovetails, scarves, laps, and birdsmouths are all pegged whereas joists and flat rafters (such as those in Tudor churches for example) rely on their own weight (and to some extent an interference fit) for their stability.

The pegged joints are more vulnerable to failure for a number of reasons. Firstly there is less material to stress and decay before failure. This is self-evident: interlocking components must be reduced in section to produce the matrices and at the same time present greater surface area open to parasitic attack and ingress of moisture. Also holes are bored to accommodate the pegs which provide access directly into the core of the timber.

It is found that designs that were adequate when the wood was cut in its green state sometimes became inadequate and prone to failure as the timber dried out. Jowled post heads (or those with upstands) and dovetailed tiebeams (Fig 2.1) were especially prone to this lack of foresight by the carpenter.

Upstand being slowly ripped out of its joint.

Fig.2.1. Truss 3 of the Wheat Barn at Cressing Temple illustrating the weaknesses associated with a post-head. The dovetail on the underside of the tie beam has broken its peg and jumped out of its matrix. The upstand has been forced to shear off along its grain and was slowly twisting out. The notched lap joint on the diagonal passing brace has also broken its peg and has been nailed. The entire movement has been arrested by a single, slender, wrought iron tie. Almost certainly an early Victorian intervention as part of a large scale repair phase. This truss was the subject of a movement study for this dissertation.

It can often be noted that a post-head will have the most comprehensive strapping in a failing building. This is because the joints all shrink away from each other and the upstanding tenon is prone to split down the grain (Fig. 2.2). This leads to movement in the joint in all directions and the springing out of the dovetail. L-ties and U-straps are common remedies to arrest further movement.

Jowled posts in the Barley Barn

Fig.2.2. Truss 3 of the Barley Barn at Cressing Temple. The remodelled barn has walls dating to c.1400Ad. The forelock bolt is one of a series inserted to prevent the jowled posts from further splitting out. An additional L-tie has been added to prevent the tie-beam from lifting out of its dovetailed joint.

The other joint to suffer from this type of failure is the scarf joint (Fig. 2.3) and in particular the earlier bladed types identified by Cecil Hewett (Hewett. CA, 1980). The blade tends to snap off along the grain and, being buried in the centre of the timber, it is inaccessible. Often the scarf is in the middle of a wall plate and this leads to the wall bowing out under the weight of the structure above.

Strapped scarf joint in Harlowbury

Fig 2.3. A photograph taken by John McCann for Cecil Hewitt in the roof of Harlowbury in the 1980's. The scarf joint is shown heavily strapped (just above the scale) and a large L-tie reconnects the tie beam to the top plate. This photo was found in a sheaf of loose papers and has never been published.

Early repairs clamped a second bracing timber across the joint, held fast with forelock bolts or T-pins (Fig. 2.4). Later repairs could incorporate threaded bolts (Figs. 2.5 & 2,6), iron-strapwork (Fig. 2.7), fish-plates (Fig. 2.8) or even a flitch plate. (See Chapter 5 for a full description of these components).

Clamp repair in Granary

Fig 2.4. A clamp placed parallel to a failing top plate in the Granary at Cressing Temple. The plate is bowing outwards under the pressure of the roof. The clamp is secured in place with forelock bolts whose round heads set them apart from threaded bolts.

Clamp repair in Wheat Barn

Fig 2.5. A clamp reinforcing a failed scarf joint in the Wheat Barn at Cressing Temple held in place with square headed bolts and nuts. Compare with Fig.2.6 below.

Scarf joint and threaded bolts

Fig 2.6. The failed scarf hidden under the eaves. This is a splayed and tabled scarf with under-squinted butts and a wedge key. An extremely complicated joint and not prone to early internal failure this joint dates to c.1260AD. The threaded bolts indicate a post 17th century repair.

Strap repair across scarf joint

Fig 2.7. A simple strap repair nailed and stapled across an edge halved and bridled scarf joint in the top plate of the north wall of the Barley Barn at Cressing Temple. The pegs have rotted away.

Fish plate across scarf joint

Fig. 2.8. A modern fish-plate reinforcing a failing top-plate in the Barley Barn at Cressing Temple. Inserted in 1989 as part of the post 1987/ 89 gale damage repairs, this ugly intervention was supposed to be temporary. Plates of this dimension with hex headed bolts all date to the late 20th century.

2.2 Materials failure

Click for Index

src=/bin/counter.pl?h=1 width=1 height=1>