Parts

Where the barn comes from. Its SHAPE is Barn Jig’s round plan, solved by Barn Jig’s own solver — diameter, plate height, the stable under the mow, the two gambrel pitches and where they break, the silo as a mast through the peak, the feed alley and the ring of cows facing across it. There is one reading of Fraser’s Fig. 19 in this toolkit and it lives there, so a barn tuned in either tool is the same barn. What Parts adds is the FRAMING: sections, spacings, courses, the ring purlins and the web — none of which the jig models, and none of which Fraser wrote down.

The source. Wilber J. Fraser, Economy of the Round Dairy Barn, Bulletin No. 143, University of Illinois Agricultural Experiment Station, February 1910 (read it in full at Project Gutenberg #38321). Fraser headed Illinois’s Dairy Husbandry department, and his bulletin is the document that carried the round barn across the Midwest; the three University of Illinois round barns (1907–1913) are its built argument. What the model takes from it: a self-supporting roof carried by a circular purlin plate at the gambrel break — F. H. King’s 1892 invention, which Fraser describes as “supported only by the braces which tie the joints” — rafters “sheathed with 1 × 6s with no space between”, a central stave silo running up through the roof, a mow floor over the stable, and cows standing in one ring facing the silo (the two boxes labelled FEED ALLEY in the section are that one circular alley, cut twice).

The proportions are scaled off Fig. 19, “cross section of 60-foot round barn”. The engraving carries no figured dimensions, only its own 60-ft scale bar, so the pitches and the break are measured off a printed picture — read them as ±a foot and move any slider that disagrees with you. The 7-ft feed alley is Fraser’s own figure. The barn is framed as an n-gon of flat facets, because that is how the jig frames, sheathes and draws it: every ring member lies in a facet and every sheathing panel is genuinely flat.

The windows. The ring of sash round the stable is ours, and it is placed from the herd rather than from the engraving: period dairy practice asked for roughly 4 sq ft of glass per cow, and the jig counts 30 cows in this barn, so 16 openings at 3′-0″ × 3′-6″ comes to about 168 sq ft — the right order, and the reason those are the defaults. Fraser argues for light in the stable and his photographs show windows; he gives no schedule of them, and Fig. 19 is a section that does not cut one. The mow stays blind on purpose — it is hay, and hay wants a dark dry loft. Each opening is framed the way a stud wall really is: the bay’s stud is cut, two trimmers carry a lintel, a rough sill goes under, and cripples fill the stud line above and below. Turn the ring to 0 and the wall closes up again, stud included.

What is ours, and is not Fraser’s. Every section size, spacing, panel course and web-brace pattern is a plausible reconstruction chosen to make an exploded drawing legible. The bulletin gives an argument and photographs, not a framing schedule. This is a teaching model of a round barn — not a measured drawing of any one barn, and not a shop drawing of any barn at all.

How the pieces are found. An imported mesh is split by welding coincident vertices and taking connected components, so a single joined layer still comes apart into its members. Two pieces are called the same drawing when a rigid motion carries one onto the other within the tolerance — tested by re-framing the candidate 24 ways (48 with mirrors) against a class representative, not by comparing names. The honest limit: a piece with a continuous symmetry (a turned column, a true cylinder) can refuse that test and split into a few types; the count then over-states, and never under-states, how many drawings you need.

What the plate is. Each piece is turned onto its thinnest face and spun until its footprint is the smallest rectangle that holds it (exact — the optimum always shares a side with a hull edge), then shelf-packed. The packing is a plain tallest-first shelf fill, not an optimal nest, which is why the percent covered is printed. Outlines are projected footprints, not unrolled developments — for boards, members and flat panels those coincide; for a doubly-curved panel they do not, and Panelizer is the tool that unrolls.