Handoff: ND cabinet reconstruction for 3D modelling¶
Date: 2026-08-02 Goal: dimensions and geometry sufficient to build and 3D print scale models of Norsk Data machines, starting with the ND-100/CX Compact.
1. Where everything is¶
| What | Where |
|---|---|
| The record of findings | E:\Dev\Ronny\NDInsight\Hardware\ND-PHYSICAL-MODELS.md |
| Drawing set (sections, isometrics, plans, elevations, panels, logo) | https://claude.ai/code/artifact/b38e0d7d-8a21-475e-b309-79f5799bcbb8 |
| Open measurement worksheet with coded callouts | https://claude.ai/code/artifact/fc3c69c7-6625-43db-a93a-58cecfd4176c |
| Photographs of the real machines | \\Nas9t\data\NorskData\Pictures\ronny\ |
| Other ND photos, brochure, panel parts, logo artwork | \\Nas9t\data\NorskData\Pictures\ |
| Manual library (7.3 GB) | E:\Dev\Ronny\mirror-sintran-com |
| Panel PROM / part-number notes | E:\Dev\Repos\Ronny\nd-120\Code\68705\readme.md |
The repository author owns a running ND-100/CX Compact and an ND-100 Satellite. Direct measurement of those machines outranks every manual, and has already been shown to.
2. State of the ND-100/CX Compact¶
Measured, trust these¶
| Item | Value |
|---|---|
| Overall | 540 wide x 690 high |
| Depth, cabinet only | 710 mm (620 box + 90 wedge) |
| Depth, including rear fans | 745 mm |
| Front projection at the roof | +50 mm |
| Front projection at the lip | +90 mm, at 370 mm above the floor |
| Front projection at the grille face | +45 mm |
| Chamfer at the base | 45 degrees, 20 mm run, down to +25 mm at 10 mm height |
| Grille: flat top / louvred / count | 80 mm flat, 270 mm louvred, 17 louvres, about 16 mm pitch |
| Base frame | 40 mm metal, side panels rest on it |
| Castor clearance | about 10 mm, wheels not modelled |
| Rear fans | 3 off, 120 mm square, 35 mm out the back, 30 mm apart, tops 40 mm below the roof |
| Operator panel fascia | 455 x 53 mm (from the PCB pad grid, +/- 2 percent) |
| Cabinet plan | square - upper and lower sections are the same width |
Structure¶
- Five mouldings per ND-830102.1B section 2.1.1: front (one screw at the top), two side panels (two screws, half turn), top panel (four screws), rear panel (four screws).
- Two-tone: dark warm brown on the front panel and base frame; lighter greyer brown on the side and top panels; cream badge; black operator panel and recess linings.
- The top section is three planes: outer wedge face (badge), a vertical recess back wall carrying the black modules, and an outward-angled operator panel face.
- The badge is an applied label with rounded ends, 5 mm clear above and below - a decal, not moulded geometry. Model text varies: ND-100/CX, ND-110/CX, ND-5000/CX, tpServer A25.
- The ND mark is a dot matrix: 18 x 8 square grid, 120 dots, dot diameter 0.93 of pitch, and the gap between the N and the D is 1.5 pitches rather than 1.
Not measured¶
Everything on the open worksheet. The single most valuable one is C1, the recess depth - with it, the sloping recess wall angle can be calculated from the known 10 mm run instead of guessed.
3. State of the ND-100 Satellite¶
Only the 230 mm width has been measured, and it matches the published table. Front face layout is known from photographs: badge, drive bay with a 5 1/4 inch floppy turned 90 degrees on its side, cream keyswitch plate (LOCKED / ON / STANDBY) and cream button plate with six buttons in two groups of three, then the same grille profile as the Compact but narrower. Colours match the Compact, but the control plates are cream with dark legends, the inverse of the Compact's black panel.
Its section drawing is entirely borrowed from the Compact and is drawn dashed for that reason. 19 values are open on the worksheet.
4. Traps - things that already went wrong¶
- The grille face is +45, not +20. The 20 mm figure is the run of the base chamfer. This error propagated through several revisions and made the lip look twice as deep as it is.
- The fans are on the REAR, not the top. Their 35 mm projection is what reconciles the measured depth with the published 740 mm. Before that was known, the manuals looked wrong; they were not - the comparison was.
- Back-face culling matters in the isometric. The underside of the overhang faces downward and must not be drawn from an above viewpoint. Drawing it produced geometry that read as physically impossible.
- Always rasterise the SVG and look at it before publishing. Doing so caught label collisions and a blank grille face that inspection of the markup did not.
- The brochure cabinet photo cannot be measured. It is a phone photo of a bound brochure, so the page curves; width-scaling and height-scaling disagree by 4 percent and perspective rectification cannot fix page curvature.
- Two operator panel variants exist, same size, different keyswitch legends: ND-323163 / PCB 1835B reads LOCKED / ON / STANDBY; ND-323165 / PCB 1844B-2 reads ON LOCK / ON / OFF. The machine in the room is the 1835 variant.
5. The one technique worth reusing¶
PCB photographs contain their own ruler. Through-hole pads sit on a 2.54 mm grid. Autocorrelate the brightness along a pad row, fit about 30 autocorrelation harmonics through the origin, and the image yields an absolute mm/px scale with no assumed dimension anywhere. Two independent photographs of the same board agreed to 0.2 percent, which is what makes the 455 x 53 mm operator panel trustworthy.
It fails on sparsely populated boards - it was tried on the ND-322691 panel board and returned a broad envelope rather than a comb, so that panel has no published size, only a measured aspect ratio of 1.847 and an assumed width.
6. Next steps¶
- Collect the worksheet answers, starting with C1.
- Recompute the recess wall angle and redraw the top section as three correct planes.
- Confirm whether the dark area is one opening with a bar or two separate openings.
- Confirm whether the three ribs on the side panel top are open vents or solid mouldings.
- Then the Compact is complete enough to model and print; the Satellite follows once its 19 values are in.
Longer term and untouched: the large 1690 x 600 cabinet has no front detail at all, and the ND-100 6-module cabinet's six bay heights were never recovered.