KIM-1

How the model was made

It is cross-referenced from five photographs of two original Rev B boards, the Musée Bolo's and another on Hans Otten's site, and from a replica of the board's layout in KiCad, traced by its author from another photograph, scaled by its chips' 0.1 inch pin pitch, and built and run. Every photograph was registered to the replica: four corners marked by hand, refined by matching the photograph's copper to the replica's, then a smooth correction for the lens. On parts of the board held out of that fit, the median error is 0.1 to 0.21 mm.

The top face's tracks are a vote between three photographs: the Musée Bolo's board from each end, and the other board, so a track one of them hides is usually seen by another. Checked against each of the three in turn, left out of the vote, the tracks and that photograph's own tracing agree on 66% to 81% of their copper (intersection over union); the map traced from the one photograph this model had before agreed on 38% to 43%, even against the photograph it came from, because it traced the tracks wider and sat a little off where the registration puts them. Against the replica, which is of a later revision and so cannot match everywhere, the match went from 42% to 57%. Under the chips, the display and the keypad, where no photograph shows the board, 8% of the top face, the copper is the replica's.

The underside's tracks are traced from one photograph, of the other board's solder side, which shows 87% of it; they match the replica's underside on 54%. Before, the underside was plain.

Heights were triangulated between two photographs of the Musée Bolo's board taken from different places, with one camera fitted to both. A point on the board itself comes out 0.14 mm above it. The typical heights the model used before were off by 2.33 mm on average, and by 4.16 mm for the socketed 6530.

The parts soldered through the board stand on the replica's pads. Checked against two photographs, a chip's body is found 0.84 mm from where the replica puts it at the median, 1.38 mm at the 90th percentile; that includes the parallax of a chip's top, which stands above the board the photographs are registered on. The keys sit on a grid found from their legends, 0.26 mm from it at the median.

What it was traced and measured from:

• The tracks on the top face, as one of three photographs of it; the keypad's size and its keys' places, the display window, the crystal, the name and the red wire; and a check on where the chips sit: from the photograph of an original MOS Technology KIM-1, on display at the Musée Bolo, EPFL, Lausanne, by Rama & Musée Bolo, cropped by Tomer T, CC BY-SA 2.0 fr.
• The tracks on the top face, as one of three photographs of it, and the heights of the parts, by triangulation with the next photograph: from the photograph of the same board at the Musée Bolo, from its other end, by Rama & Musée Bolo, CC BY-SA 2.0 fr.
• The heights of the parts, by triangulation with the photograph before: from the photograph of the same board at the Musée Bolo, from a low corner, by Rama & Musée Bolo, CC BY-SA 2.0 fr.
• The tracks on the top face, as one of three photographs of it: a different board, so where a part or the red wire hides a track on the Musée Bolo's, this one often shows it; and a check on where the chips sit: from the photograph of another MOS Rev B KIM-1, its component side, on a white ground, by Hans Otten's Retro Computing site, which does not name the photographer, no licence stated.
• The tracks on the underside: the only photograph of it: from the photograph of the same Rev B board turned over: its solder side, by Hans Otten's Retro Computing site, which does not name the photographer, no licence stated.
• What every photograph is registered to; the places of the parts soldered through the board; the tracks under the chips, the display and the keypad, which no photograph shows; and the reference the tracks are measured against: from a replica of the KIM-1's Rev D board, in KiCad, by Eduardo Casino, CC BY-NC 4.0.

What it does not show, or shows less well:

• The underside is one photograph of a different board from the one the rest of the model follows, and only the replica checks it.
• Under the chips, the display and the keypad the copper is the replica's, which is of Rev D, not the photographed Rev B.
• Where the photographs disagree, 10% of what two or more of them see, the vote decides, and the photograph outvoted may be the one that is right.
• One part of each kind was measured for height, and the others of its kind take that height. Of the resistors, capacitors and diodes, two capacitors were measured; the rest take their sizes from their footprints, not from a photograph.
• Resistors, capacitors and diodes are plain cylinders without their bands or printing, and the transistors are round where the real ones are flat on one side.
• The red wire follows the path one photograph shows, drawn lying just above the board.
• The track map has 8 pixels to the millimetre, so close up a track's edge is soft.