Two Paths

This is a picture from that same church I was showing last week, but the topic really isn’t about the church:

This is a connection of two heavy-timber pieces of a roof truss. There’s no visible bird’s mouth cut where they meet, although there may be a mortise and tenon that we simply can’t see from this angle. More importantly, we have a large cast-iron bracket (the brown thing) and a this steel strap (the black thing), both bolted to the wood. Most likely the bracket’s bolts are carriage bolts or similar, and the strap’s bolts are through bolts running to another strap on the far side.

The cast-iron bracket is older, possibly original to the building, and the straps look to be a more recent retrofit. More importantly, they represent different philosophies of structural detailing. The bracket is a factory-manufactured object that has little flexibility of use: the angle of the timbers it helps connect is set in the factory, long before the timbers are erected on site. The strap was cut to length and had holes punched for the bolts specifically to match the existing conditions, and its use is not all that precise: you can see shims between the first and second bolts from the left.

Massive simplification incoming: As engineering and engineered construction has developed in the US, factory pre-fabrication has been used mostly for machines and cut-to-fit fabrication has been used for those items assembled on site: structure and MEP mechanical risers and branches. The various efforts at pre-fabrication of buildings (including those happening now) have changed this some, but less than you might think: a prefabricated building chunk still has to be built, and the question becomes for that chunk whether the pieces that get assembled for it in the factory were premade or cut to fit.

This is not simply a philosophical question. Pre-manufacture requires knowing ahead of time the parameters that effect use. In theory that’s always possible for any building element, but in practice it’s not: the combination of fast-tracking (beginning construction before design is 100 percent complete), cumulative error from geometric tolerance over the long distances within a building, and imperfect coordination of the 3D assembly that makes up the entire building make machine-like prefabrication difficult. Keep in mind that shop assembly is typically less expensive than field assembly, so there’s a huge incentive to follow the prefab path, and the fact that the industry has not, for the most part, suggests that the problems are not yet solvable with current technology. Finally, if construction could be reduced to the assembly of prefabricated objects, than robotic assembly would be readily at hand. I’ve been reading about robots in construction for ore than 30 years and have yet to see one on site.