Methodology and tests

Everything on this site is plain arithmetic that runs in your browser. Here are the formulas, the definitions they rely on, and how we test them.

Definitions

• 1 in = 25.4 mm exactly, by the 1959 international yard and pound agreement. 1 ft = 304.8 mm.
• A scale 1:N means model length × N = real length. Ratios a:b are read as N = b ÷ a, so 2:1 is an enlargement with N = 0.5.
• PDF pages are measured in points, where 1 pt = 1/72 in, so 1 mm = 2.83465 pt. A4 landscape is 841.89 × 595.28 pt, and US Letter landscape is 792 × 612 pt.

Formulas

• Model = real ÷ N. Real = model × N. Scale to scale: new = old × N_from ÷ N_to.
• Area factor = k², volume factor = k³, where k = N_from ÷ N_to.
• Figure ratio (heroic guide) = represented height ÷ figure height. The eye-level option uses 93.5% of standing height, an illustrative approximation.
• Ruler ticks: the smallest round real-world step whose spacing on paper is at least 0.8 mm, with labelled ticks at least 6 mm apart. Tick k sits at margin + k × step ÷ N, computed from the index each time, so there is no cumulative rounding.

How it's tested

The source includes automated tests (Node's built-in test runner) and browser checks. The current counts and the exact commands are in the test report that ships with the source code. In summary:

• Hand-checked examples: for instance 1,750 mm at 1:35 = 50 mm, 1 ft at 1:12 = 1 in, ¼ in at 1:48 = 1 ft, and 250 mm from 1:35 to 1:72 = 121.53 mm.
• Round-trip property tests: thousands of random values, units and scales converted there and back, matching to 1 part in 10¹².
• Boundaries and impossible inputs: zero, negative numbers, NaN, infinity, unknown units, malformed ratios such as 1:0 or 1:2:3, and out-of-range scales.
• Physical ruler checks: the tests parse the exported SVG (width/height in mm) and PDF (MediaBox in points, uncompressed line coordinates). They confirm the page size, the 100 mm and 4 in calibration lines, the card outline and every tick spacing, for both paper sizes and both unit systems. A separate check rasterises the PDF with Poppler and measures it.
• Import safety: malformed JSON, files from other apps, oversized files, prototype-pollution keys and invalid numbers are rejected or dropped.
• Headless Chrome on desktop and phone-sized viewports: each page loads without console errors, the tools produce results, and the downloads produce valid files.

Limits

• Results are only as good as your inputs. Kit and figure sizes don't follow their nominal scale exactly, and illustrative references are typical values, not specifications.
• A printed ruler is only as accurate as the printer and paper. Always measure the calibration marks.
• The site gives no structural, safety or engineering advice. It's for hobby modelling.

Who runs this site

Miniature Scale Lab is operated by SPAZ Power LLC. Corrections and suggestions are welcome at 55paz.power@gmail.com. We don't claim professional credentials, and the site has no sponsored content.