Notes from building models out of matchsticks
Dated notes from whatever is on the board. Stick counts are honest, including the ones that ended up in the bin.
Dry-fitted the second span and the diagonals were all short by a hair, consistently. Measured the jig: end stop had worn 0.4 mm. That is forty-eight pieces to recut and a lesson I had apparently not learned the first time. Rebuilt the stop with a brass facing glued on, checked against the rule, cut ten test pieces — all within a tenth.
Running total roughly 8,900 sticks. Bin total, for honesty, somewhere over four hundred.
Both trusses up and cross-braced. Went with box beams for the top chord and it was the right call — the whole span is noticeably stiffer than the mock-up I built in spring, at about the same weight. Cross-bracing between the trusses turned out to matter more than the bracing within them; before the cross-members went in the pair could be twisted by hand quite easily.
Weighted the deck flat overnight and it came off the board without any twist worth measuring, which after the windmill is a relief.
Decked the whole 940 mm in craft splints laid across matchstick stringers. Splints are the right material here and I should have used them earlier for anything wide — one splint replaces five matchsticks and does not show four glue lines.
Glued in three stages rather than in one go, which kept it flat. Stage gluing costs an evening and saves a rebuild.
Finished the tower and cap. The gallery deck twisted about 3 mm across 120 mm because I glued the whole ring in one sitting and did not weight it. Damped the concave face and left it under books for two nights; it came back to something under a millimetre, which I can live with.
Sails are removable and probably too fragile. If I build another I will laminate the spars rather than using single sticks.
Ran a proper comparison on scrap: white PVA, yellow aliphatic and thin cyanoacrylate, ten lap joints each, broken by hand the next day. PVA and aliphatic both failed in the wood rather than the glue line, which is what you want. Every cyanoacrylate joint failed at the line and failed suddenly. Switched to PVA for everything structural and have not regretted it.