Methodology and editorial policy

A calculator is only worth using if you can find out how it decided. Most cannot tell you — they are a formula someone typed into a form, with no source, no test and nobody to write to. This page is the alternative: the process behind all 22 calculators here, written plainly enough that you can hold us to it.

1. Where the numbers come from

Every constant, table value and formula is transcribed from a published source, and that source is named and linked at the bottom of the page that uses it. In practice they fall into three groups:

• Standards. NFPA 70 (the National Electrical Code) for installation and code-compliance work; the IEC 60364 series for wiring rules; IEC 60063 for standard component values; IEC 60228 for conductor sizes; IPC-2221 and IPC-2152 for PCB trace sizing. Where a figure is edition-specific, the edition is stated in that page’s sources rather than assumed.
• Reference texts. Horowitz & Hill’s The Art of Electronics and Sedra & Smith’s Microelectronic Circuits for the electronics pages.
• Manufacturer data. Where a model approximates real hardware — the taper of a DC fast charge, for instance — the dataset carries the published figure and a test asserts our model reproduces it. We fit models to data. We do not ship a plausible-looking curve.

What is not a source: another calculator. Copying a competitor’s answer propagates their mistake and teaches you nothing about whether it was ever right.

2. How the maths is verified

Each calculator’s arithmetic lives in a pure function — no interface code, no shared state, nothing that can behave differently depending on what you clicked first. That function has its own test suite, and the tests assert against worked examples taken from the sources above, not against the function’s own output. A calculator does not ship until they pass, and they run again on every change to the site.

Three rules do most of the work:

• Textbook-verified values, not self-consistency. A test that only proves the code agrees with itself proves nothing.
• No plausible-looking wrong answers. When the inputs do not describe a real design, the correct output is no number plus a warning that says why — never a figure that looks usable. This is the one behaviour worth breaking a page over.
• Behavioural tests, not just point checks. Results that must rise, fall or stay ordered as an input moves are tested for that, which catches the class of error a handful of spot values will always miss.

Diagrams are checked by rendering them and looking at the image. Every diagram fault this site has ever had was found that way and none were found by a unit test, so the process reflects that rather than pretending otherwise.

3. Human review

Tests catch a mistyped constant or a misread table. They cannot catch an answer that is arithmetically perfect and wrong in a building. That is the job of the site’s technical reviewer, Salamot Hok, a working electrician with 10+ years of installation experience in Bangladesh and the wider South Asian region.

He reviews for practical realism: whether the default values are the ones people actually meet, whether each worked example describes a real job, whether a warning fires at the point an installer would stop and think, and whether the result is expressed in terms the trade uses — a cable size you can buy, a breaker that exists.

He is not the authority for the code tables, and no page implies he is. The scope of what he does and does not vouch for is set out in full on his profile, including the fact that his field experience is IEC-region practice rather than United States NEC inspection. A reviewer credited with expertise he has not claimed is worth less than no reviewer at all.

Where a page carries a reviewer’s name, that person has read it. Where one has not been reviewed, the page says so instead of borrowing the name — the byline is generated from the same record as the structured data, so the two cannot tell different stories.

4. What we refuse to publish

The omissions are as much a policy as the inclusions, and each of these is a thing the site could easily do and deliberately does not:

• No invented ratings. There is no star rating or review count in our structured data, because we have no reviews. Fabricated ratings are the most common way calculator sites earn a rich result they have not earned.
• No fake authors. One real, contactable reviewer, with a public profile you can check, rather than a roster of invented engineers.
• No traffic figures dressed as facts. Search-volume estimates inform what we build; they never appear inside a page as though they were measurements.
• No calculator for a question that is not a calculation. Some safety-critical topics are checklists, and packaging one as a calculator implies a precision that does not exist. We have turned pages down on exactly this ground.
• No advertising, affiliate links, sponsored calculators or paid placement. No third-party script of any kind runs on this site — see the privacy policy. If that ever changes, it will be disclosed here and there before it ships.

5. Corrections

A wrong formula is the one mistake this site cannot recover from, so error reports are answered before anything else in the inbox. If a result looks wrong to you, send the calculator name, the inputs, the answer you expected and — if you have one — the standard or textbook page you are checking against. The quickest way to send exact inputs is the Share link button: every value is encoded in the URL, so the link reproduces your calculation precisely.

What happens next is fixed:

1. The report is reproduced against the cited source.
2. If it is confirmed, the calculation function is fixed and a test is added that fails on the old behaviour — so the same error cannot return quietly.
3. The page is re-reviewed and its review date advanced, and the correction is described in the page’s own text where a reader could otherwise have acted on the old answer.
4. You get a reply either way, including when we conclude the original was right.

Report an error: [email protected], or via the contact page.

6. Review dates

Every calculator page shows the date it was last reviewed, and the same date appears in the page’s structured data. It moves when one of three things happens: a correction is confirmed, a cited standard is revised or withdrawn, or the page comes up in its scheduled re-read. It does not move because the site would like to look recently updated. A refreshed date on an untouched page is a lie told in small print, and it is the cheapest trust signal in the world to fake — which is the reason to be careful with it.

7. What this site is not

These calculators are engineering guidance. They are not a design review, an inspection or a code sign-off, and no result here substitutes for the governing standard or for the judgement of the qualified person responsible for an installation. The full statement is on the disclaimer page, and it is worth reading before you rely on anything here for work that could hurt someone.