outerdiameter.com

Policy · v48d86b977106

Accuracy policy

A dimension table has a narrow set of ways to be wrong, and they are all boring: a typo, a rounded value, a copied table cited as independent, aminimum wall presented as if it were nominal. Here is what we do about each.

The rule — and what it covers

What is not published

27 size × designation values across the four series sit in a quarantine list: a single source, or two sources that disagree. They are reported as counts in the data index and named on the chart pages, and they do not appear in the tables. A reference that never has gaps is usually guessing.

Material families we do not cover at all are marked as such: copper tube (ASTM B88 K/L/M) is missing because every table we found fails the attribution test, and metric tubing (EN 10220) is missing because none of our inch-native sources carry it. CPVC and PEX sizes beyond the published ones are held for the same reason.

Minimum wall versus nominal wall

Steel standards publish a nominal wall, so the bore we derive is the exact arithmetic of the definition. Plastics publish a minimum wall with a tolerance above it, so the bore we derive is a maximum. That applies to ASTM D2241 and ASTM F876/F877 (CTS). Where a bore is a maximum, the size page and the table say so in the row itself — not in a footnote someone skips.

If you are sizing a fit, a liner or an insert, do not treat a maximum bore as the typical bore: real PVC/CTS pipe is commonly thicker than the minimum wall.

Precision and rounding

Editions and dates

No revision year is claimed for any standard here. Our sources agree with the content of ASME B36.10M / B36.19M, ASTM D2241 and ASTM F876/F877, but we did not verify which printing each figure comes from, so the pages name the standard without a year. The build rejects an invented year rather than print one.

The data set is stamped 48d86b977106 — a digest of the input files, so any page and any JSON export can be tied to the same snapshot. Sources were last checked 2026-09-25.

When this is wrong

Assume it can be, and check the two sources listed under the number — that is the whole design of the site. Every published value carries its sources on the page, and the published data files are downloadable above. If a source pair agrees on something that contradicts a standard document, the standard wins and this page is wrong.

There is no correction form yet (the site is pre-launch). Until there is, the mechanism is the data itself: the CSV-level provenance in the JSON export states which two publications produced each figure, so a dispute is checkable rather than a matter of opinion.

What this site is not