420 
Blake 
• Variation of pattern quality. There is a noticeable grouping of components by quality of finish; the uppermost radials separating 
the ribs are of much poorer quality than those lower down. It is suggested that some components have been cast at speed from 
crude patterns to resolve critical erection problems. The wide spread use of open castings rather than closed also suggests an 
urgency to the component manufacture. 
• Use of retro fitted parts. It is clear that the use of retro fitted parts was part of the construction sequence; the erection of a single 
frame requires the one off casting of at least 10 major parts out of approximately 40, to a fitting tolerances of +/- 25mm. The 
casting process from the fitting up of patterns, casting to fit at the tolerance required, oversize mould making and volatile 
material handling was probably unique to the Coalbrookdale foundries at the time. 
5.2 Evidence for the development of fixing technologies 
It has been widely asserted (J.Dupre 1997, McGuire+Matthews 1958) that the structure relies on developments of the wood fixing 
technology of the day for the basis of its fitting and fixing of parts. This is not entirely so, there a number of fixings used that are 
unique to metal work. 
• Expanding soft iron wedge or plug. This is unique to iron fixing and shows a highly developed understanding of friction 
behaviours in cast iron of different hardness. 
• Engineering technology. Nut and bolts (square and hex) are used extensively for joints subject to twist (there are at least 229 
Cl8th nut & bolt fixings used in the structure). 
• Blacksmith work. Strap and band ties are shrunk to fit the 6 braces between the frames. 
Some fixing methods are clearly derived from woodwork practice. Hidden halved dovetail, dovetail, mortice and tennon are all used, 
but using lugs, housings and the web closing mortises shaped to make best use of the superior strength to weight properties of cast- 
iron. If wooden components were of similar proportions the lug and web shapes would be much thicker. 
5.3 Erection sequence 
A recent archive discovery (DeHaan /Skandia 1997) has indicated the erection sequence did not rely on massive timber false-work 
for the first ribs. Inspection of the crown joint for the survey (Blake 1999/2000) reveals a fixing system of some considerable 
complexity that supports the sequence suggested by Selby in 1999. The handling of the bridge materials is worthy of consideration, 
as the main castings were some of the largest cast objects on earth at the time. 
Inspection of the bridge reveals that frame C [the central frame] differs from the others in 3 ways: 
-. No shoulder is cast into the outer vertical, the other frames all have this. 
-. The upper rib re-uses the 2 sided mortice mouldings from the mid-rib: this is a mistake as the mortices for the upper rib are only 
one sided. The error is not repeated on the other frames. 
-. The mid rib rests on a shoe fitted to the sole-plate all the other frames use a surface fixing detail. 
The design modifications to the frames on either side of FrameC suggest some refinement of the parts took place during the 
construction phase and that Frame C was the first erected. 
The CAD model can be used to test erection theories. By fitting views of the model to the 1779 sketch by Elias Martin it is possible 
to re-create the sequence of assembly starting with frame C. 
Testing the fit of the parts by virtual assembly shows many of the displacements in the bridge may date from its erection. For 
example the cast size of the circles on the south side of the span are oversize to accommodate the cant on the inner verticals. This 
can be proved by bringing the verticals upright in CAD and testing the fit of the circles. 
5.4 Assessment of pre photographic image/ drawing regression study 
Comparison with a matched perspective view by Rooker (cl782) reveals frame C details omitted adding to the evidence for this view 
being prepared from design drawings rather than from the structure itself. 
6. CONCLUSION 
6.1 Survey Methods 
Photogrammetry is the most reliable method for mass 3D data capture. However it can be compromised by the image quality of the 
stereo pair. The production of the wireframe used in the survey required detailed briefing of the contractor who was chosen on the 
basis of previous work of this kind. The selection of edges to produce a good 3D wire-frame for complex historic structures needs to 
be based on experience informed by a well founded architectural knowledge. 
The infill by REDM matched the wire-frame integrity from photogrammetry. Real-time CAD data collection allowed data selection 
and checking of a high order of precision and observational accuracy. The selection of targets was informed by real-time checking of 
the wire-frame on site.