The PhD applied: a structural diary

People ask what the PhD actually does for a carpentry job. It's a fair question. Most of the time, the answer is "nothing visible". Sometimes the answer is "saves the job". Here are three moments from a single deck build last spring in Kellyville.

Moment one: the footings. Spec called for 12 concrete piers, 600 deep, 300 across. Standard. Walked the site, found one corner where the soil read was wrong. The slope hadn't been graded properly by the previous landscaper and the topsoil was sitting over a layer of fill from when the pool went in. Fill is not load-bearing. If we'd put a pier into that without thinking about it, the deck would have settled at one corner within eighteen months. We added one extra pier, dug another 400mm to find native ground, and used a tied-cage reinforcement instead of plain concrete. Costs $180 extra in materials and one hour of labour. Saves a callback in 2027.

Moment two: the joist run. Original plan had 4.2m joists at 450mm centres in F17 hardwood. Fine on paper. But the client wanted to put a heavy outdoor kitchen at midspan, fridge, oven, masonry-look benchtop. Suddenly the dead load at one point doubled. The span table assumption (uniform live load only) is now wrong. Re-ran the calc on a Saturday morning, dropped the joist centres to 380mm in that bay only, doubled up the joist directly under the kitchen island. Took two extra joists and forty minutes of re-cutting. Means the floor won't bounce when someone stands on it next to the oven.

Moment three: the hidden fixings. We use a clip system for hardwood decks, T-clip into a routed slot. The clip is rated to a certain pull-out load. On most decks the load is uniform, the clips share it, and you don't need to think hard. On this deck the boards ran from the house wall out toward the edge of a cantilevered section. Differential thermal expansion plus the cantilever moment meant the outer clips would see more pull-out than the calculator assumes. We doubled the clip density on the outer two boards. Looks identical from above. Holds when the timber tries to walk in February.

None of this is rocket science. It's just an engineer's instinct to ask "what's actually loading this?" before defaulting to the span-table answer. Twenty years of carpentry without the engineering would have built the deck the standard way and it would have been fine in 95% of weather conditions. Twenty years plus the engineering catches the 5%.