Fixed sizes from Saitō (1904) assume Meiji-era members of about 4寸 (120 mm); larger timbers may differ.

Ticks mark values documented in a source. Values marked placeholder are not yet read from the project record and borrow the denmoku proportions. Feasible zones are recomputed from the checks each time a setting changes.

Solid lines are visible edges; dashed lines are hidden edges inside the timber. Dimensions are written by their rule: as fractions of H or W where the sources give a proportion, and as sizes where they give a fixed size.

Where the traditional sequence and the tested specimen differ, both are given. Sketches are drawn from this model at its current proportions; hatched areas with a dotted outline are the material removed in that step.

What changes from manual fabrication: the milling cutter leaves rounded inside corners, so the square dowels and the tongue need hand finishing; the fit is set by a programmed clearance instead of paring; and the cutter chips the edges of the timber.

Redrawn in CAD from the cross-section in the restoration report (2008): grid, levels and repeating members regularised, ornament simplified. Tick the box to show where the report marks daimochi tsugi. Pinch or scroll to zoom, drag to move.

• Mieidō, Senjūji (専修寺御影堂, Tsu, Mie)
  Raised 1666 (寛文6年); Important Cultural Property 1961, National Treasure 2017; restored 2000–2007
  daimochi tsugi on the beam over the columns of axes 4, 6, 7 and 9, as marked in the section of the restoration report. Whether they are original or from a later repair is not yet read.
  Source: restoration report section; 文化遺産オンライン; Senjūji

From Wang (2024), who sampled 104 temple and shrine records at random in the Kyoto Institute of Technology library archive and checked the drawings for daimochi tsugi [WA 3.2.2–3.2.3, Figs 3.12–3.16, App. A.2].

Wang's caveats: a small random sample; Kansai is over-represented because the archive is in Kyoto; the listed year is often a repair or rebuilding year, not the founding year, so the Kamakura date marks the earliest record, not the origin; older drawings are sometimes hard to read. The exact counts by region and period are in her Figs 3.13–3.14 and Tables A.2–A.3.

Buildings recorded here

• Full-size specimen, Takenaka Carpentry Tools Museum (Kobe)
  Museum specimen; dimensions not yet read. Source: museum catalogue 2014

Derived from the contact limits between parts, for the current configuration. Repair here means replacing a part with as little taken apart as possible; temporary support is shown as practice, not calculated. "Show" plays it in the 3D view.

Evidence for repair

• Documented practice, general Columns are repaired or replaced locally by jacking up the structure, with the roof left in place: 瑞巌寺本堂, 三重県庁舎, 本山寺霊廟. None of these cases is at a daimochi.
• Geometry, this model The list above is derived from the contact limits between parts: with a post below and dowels, the post comes out with nothing else taken apart.
• Historical text and test The joint is held closed by load from above (Saitō 1904, via Shimoyama). In Wang's tension tests, the clamped joint pulled back into position (Wang 2024).
• Not claimed here Whether a given frame stays sound while the support is out depends on its loads and members.

Thresholds marked assumed are prototype settings, not source values. The structural check in the paper runs in CalculiX and is not reproduced here.

Axial tension with spring clamping to represent roof preload, 88.9 mm square section [WA]. One specimen per condition: indicative, not a statistical result. Cracks started at the scarf tips and around the dowel bearings.

denmoku-db data sheet [DM1]: stiffness and capacity reported per size. No peer-reviewed bending or shear test of this joint was found.

Test specimens in the model

Their dimensions appear as brown dots on the sliders; "Show in model" sets the model, drawings and checks to that specimen.

1. WA Wang, A. Join(t)ing the past and present. Senior thesis, Princeton University, 2024. Project geometry (Fig. 4.40: section 3.50, pieces 18.78, assembled 26.00, tips 0.53, dowel length 2.50 in), variants (Figs 3.2 to 3.7, wedges in Fig. 3.5), survey, hand and CNC fabrication, tests. Specifies a 45-degree scarf angle and blind 1 x 1 x 2.5 in ash dowels. Reconciled with Fig. 4.40 and her STEP: the 45 degrees is the reverse step at the centre (the scarf either side slopes 1:3), and the dowels are blind, 1.05 in square and 2.5 in long, stopping 0.5 in short of each face. With the project-specimen preset the viewer reproduces both, its solids within 0.05 mm of her STEP (joints/daimochi_tsugi/fe_viewer_geometry); with Dowels the pins are drawn 2.5 in long, with Through draw pins full height.
2. IS Ng, Wang, Ghorbani, Hackl. Design for Repair and Reuse, and Digital Fabrication of Traditional Japanese Timber Joinery. ISARC 2026. Expert CAD: 477 mm pieces, 660 mm assembled.
3. DM1 denmoku-db. Daimochi-tsugi tension data sheet, from the 2011 MLIT traditional construction report. Proportions as summarised; verify on the sheet.
4. SJ Report on the repair works of the Mieidō Hall, Senjūji. 2008.
5. SM Sumiyoshi, T. and Matsui, G. Wood Joints in Classical Japanese Architecture. Kajima, 1991. Related oblique scarf splices; the daimochi itself is not included.

The JSON keeps the source of every value. Fabrication code is produced by the project pipeline from the same parameter file and is not generated in this prototype.