Identifying the Location and Condition of Fasteners and Connections
Once a radiograph is generated to optimize contrast, the internal components of walls, individual timbers, or other structures can be examined in more detail. Identifying the location and condition of metal connections in a wall made predominately of wood is relatively straightforward because of the high contrast between the two materials. Wood connections, such as mortise and tenon joints, dovetail joints or wooden pegs, are more subtle features that may require image enhancement techniques to visualize on the radiograph.
In many of the examples given above, metal fasteners were evident in the radiographs. Because of the high contrast between the metal and the surrounding wood, inclusion of metal fasteners allowed for comparison of radiographs using different pulses, distance and wall construction. In fact, much of the research to examine fasteners coincided with the work discussed above. For purposes of illustrating the effectiveness of using radioscopy to locate and examine fasteners, a brief discussion follows.
Split ring connectors were shown in Figure 8, iron rods in Figures 12 and 17, nails and bolts in Figure 13, and screws in Figure 18. Nail types can be determined from the radiographs, which can be useful in identifying historic fabric or at least dating the period of construction. Hand-forged nails, cut nails, rosehead nails, and common wire nails are all easily distinguished in a proper radiograph. Thread patterns can be determined for screws, bolts and reinforcing rod.
Of equal significance to the historian or architect and perhaps greater significance to the structural engineer is the condition of the fasteners. Corrosion is often hidden within the timber and can lead to catastrophic failure. As is visible in the iron rod embedded in timber in the radiograph shown in Figure 17, the loss of diameter of the rod due to corrosion can be estimated.
Also useful for structural considerations, the connector spacing can be established when none of the fasteners are exposed. Figure 30 shows the nailing pattern in a granary crib wall. Using stacked lumber as the structural walls, no nails, bolts, or rods were visible that would indicate how the walls were connected. A simple radiograph quickly revealed the spacing of nails. Based on the known thickness of the wall, it was also possible to determine the length and diameter of the nails.
Examples of how radioscopy was used to locate and identify metal fasteners are given in the case studies section of the report.
Easily locating and determining the condition of metal fasteners embedded in wood is possible, in part, due to the different densities of wood and metal. For wood connections, such as mortise and tenon joints, the interpretation of the radiographs is more challenging.
Laboratory and field research conducted on mortise and tenon joints showed that differences in the grain orientation of the mortise and tenon can be seen on a radiograph. A cross-hatched pattern is typically visible where the two pieces of wood overlap as the tenon penetrates the mortise.
Figure 31 shows the x-ray source in position to provide an image of a mortise and tenon joint in the field. Visible moisture stains raised concerns about the condition of this joint. Since this joint was between timbers with good access on both sides of the joint and, thus, it was easy to shoot perpendicular to the face of the timber, it was an excellent candidate for examining an all-wood joint. The radiograph of the joint is shown in Figure 32.
Note the light spot near the center of the radiograph, corresponding to a improperly drilled hole through the tenon. Figure 33 is an enhanced version of the radiograph that better illustrates the cross-hatch pattern indicative of two pieces of wood with the grain perpendicular to one another. Further image enhancement was able to show the wood grain in the beam with the mortise, verifying that the mis-drilled hole was in the tenon. There is more discussion of image enhancement techniques in a later section of the report.
The images shown above demonstrate the ability to “visualize” a mortise and tenon joint when it is possible to place the source and imager perpendicular to the face of the joint. Since many field situations do not provide such easy access, a mortise and tenon joint was evaluated in the lab to determine the effect of an angled alignment with the face of the joint. The timbers making up the joint were 6 inches by 5 3/8 inches and secured with two timber pegs.
The initial setup was similar to that described above with the source and imager perpendicular to the face of the joint (Figure 34). With the imager placed directly behind the joint, the radiograph shown in Figure 35 is generated. The size of the tenon, relative to the mortise, is readily visible, as are the knots in the beam and the pegs securing the joint.
So how does the image change if it is not possible to be aligned directly with the joint, as would be the case if it was necessary to place the setup so as to examine a joint on top of a timber-frame wall? Visually, the joint would have the appearance shown in Figure 36. The corresponding radiograph is shown in Figure 37.
Although the images in Figures 35 and 37 are of the same joint, the affect of orientation of the setup is immediately apparent. Laid side-by-side, it would be impossible to know that the two radiographs were of the same joint. This reinforces the findings discussed in the geometry section regarding the significance of the setup. It also conveys the importance of documenting the position of the setup relative to a point of reference on the object (in this case, the joint). Photographs and a sketch showing distances and orientation of the setup ￼￼￼￼￼￼￼￼￼￼￼￼are critically important for accurately interpreting the radiographs after the field work is completed.
Identifying Wood Deterioration
The effect of decay, termites and mechanical damage in timber is often difficult to assess without extensive destructive tests. This phase of the research focused on the suitability of digital radioscopy for identifying internal deterioration in wood. Decayed timbers and logs and termite-damaged studs served as the research materials.
To assess whether the relative extent of deterioration could be determined visually, the wooden stepped block that was described in the Producing Contrast ￼￼￼￼￼￼￼￼￼￼￼￼￼￼￼￼￼￼section was used. The block, shown in Figure 38, represents a uniform increase in material at 25, 50, 75 and 100 percent of the thickness. The radiograph in Figure 39 does show increasing brightness from left to right as the thickness of the block decreases. Because the length of the block is greater than the length of the imager, only the edges of the thickest and thinnest sections of the block are visible in the radiograph.
Since the relative brightness is a function of the setup (distance from source to object and number of pulses), it is difficult to quantify the loss of thickness based simply on the grayscale image. Colorizing the image results in similar distinction, ranging from yellow for the section that is 100 percent of the block thickness, to green, then blue. For this image, the two thinnest sections of the block cannot readily be distinguished. Quantifying the loss of material in this ￼￼￼￼￼￼￼￼￼￼non-deteriorated block is not possible visually from either of the radiographs. With that knowledge in mind, examining decayed wood, which does not have the benefit of uniform boundaries between the different thicknesses of wood, is even more difficult.
A decayed timber removed near the base of an old ice house served as an excellent object for radioscopy because one side was intact with no visible signs of deterioration (Figure 41). However, the back face was completely deteriorated and the thickness at any location could be measured (Figure 42).
The radiograph as viewed from the face of the timber is shown in Figure 43. Note that in addition to the visible knot there are knots above and below the visible knot that are not visible on the photograph of the timber profile. Of more importance is that the radiograph is darkest at the bottom of the image where the timber is solid and lightest at the top of the image where the void is the most extensive. We can conclude from this image that there is less material at the top of the timber than at the bottom; however, we cannot yet determine how much material is missing.
Colorizing the radiograph in Figure 43 makes it more obvious that the amount (thickness) of wood is not uniform from top to bottom. As discussed regarding the stepped block, the enhanced image still does not give us the means to quantify the loss of material, although the presence of decay is obvious.
Similar research was conducted on termite-damaged wood. A stud with termite galleries (Figure 45) was placed into the test wall with wood paneling on the interior of the wall and OSB sheathing on the exterior of the wall. The radiograph of the termite-damaged stud is shown in Figure 46. Note the similarity between the non-uniform color pattern in the radiograph of the decayed timber and the stud. Since termite galleries tend to be very irregular along the length of an affected piece of lumber, the color pattern is more ragged for the stud than the decayed timber. This allows us to differentiate between hidden damage caused by decay fungi versus termite or other insects. With the enhanced image in Figure 46 it is interesting that the vertical cut lines in the bead board paneling are visible once colorized.
Many historic structures are of log construction. A solid log section is shown in Figure 47. Because the log is round, the most effective system to use was the Logos Imaging system with its flexible imaging plates. The plate was wrapped around the circumference of the log and secured with duct tape. Other than the plates being curved to conform to the shape of the log (to maximize contrast), the setup was similar to those used for the rest of the research.
The radiograph of the solid log shows uniformity in the grain of the log throughout most of the image (Figure 48). As with the decayed timber discussed above, the dark areas are knots. Figure 49 shows a section of the log with an internal void. The radiograph of the section with the void has an absence of grain, as characterized by the light areas in center of the radiograph where no wood remains (Figure 50). The lighter area without visible grain is characteristic of internal voids in any wood members. As with the solid section, the dark areas of the image are knots. Decayed wood can also sometimes be seen as a mottled pattern on the radiograph (see Figures 43 and 50).
In summary, decayed wood can often be identified in a radiograph when it
displays the following features:
- Lighter color(s) due to reduced cross section
- A transition from non-decayed wood with intact grain pattern visible on
the radiograph to the lack of a visible grain pattern in the decayed area
- A generally mottled appearance in the lighter decayed area
Additionally, wood that has been attacked by insects will also show lighter areas corresponding to a loss of material. The boundaries between undamaged and damaged wood are more abrupt than for decayed wood and tend to display a ragged or tunneled pattern.
Clearly, viewing raw radiographs can be useful for identifying the presence of deteriorated wood. As is discussed next, image enhancement and post- processing can help to quantify the extent of damage from a digital radiograph.