RailCompat
SCOPE HO & N scale · every fact cited + dated · “not published” is shown as unknown, never guessed

Method & sourcing

What goes into a figure on this site before it's allowed to publish.

Where each number originates

Every fact on this site comes from a manufacturer's own current product page or category page — the kind of page you'd land on shopping for the track yourself — except the NMRA curve classes on the minimum curve radius page, which come from the NMRA's own published RP-11 document. We don't transcribe by ear from forum posts or old reviews. Each value is stored with the exact source URL and the date we checked it, and both show up next to the value on the page.

The site covers three HO/N scale track systems: Kato UNITRACK, Atlas True-Track, and Bachmann E-Z Track (Nickel Silver, gray-roadbed line). We picked these three because their own shop pages publish enough of the same kind of field — rail code, roadbed, joiner — to compare directly; see Known gaps on the track interchange page for what each one leaves out.

The DCC decoder sockets page is sourced from TCS's and SoundTraxx's own product pages: TCS's individual decoder product pages (its cross-brand "Find Your Decoder" chart is a client-rendered app whose data we couldn't fetch, so we used the server-rendered product pages instead) for socket types, current ratings and official WOWKit compatibility lists, and SoundTraxx's own factory-installed-decoders index and product pages for factory-install history and Drop-In boards. Digitrax was excluded from this site's sources: its Terms of Use limit use of its site content to personal, non-commercial use and prohibit public display or reproduction, which this public site can't satisfy. A further section on that page lists a handful of individual locomotive models per brand (Kato, Atlas, Bachmann), sourced directly from each model's own product page rather than from TCS or SoundTraxx — these are dated examples of how that specific model is described, not a claim about the whole brand's product line. For Broadway Limited Imports we list the model names only and link to the manufacturer's own page for the socket details, rather than restating them here. MTH is noted there too: only an O scale product page was reachable when we checked, so its HO/N entry says not found rather than borrowing the O scale finding.

The minimum curve radius page lists the plain curved pieces each line sells, taken from Kato's own UNITRACK catalog PDF (we cite the page number) and from Atlas's and Bachmann's own shop category pages, every page of each listing. Kato measures in millimeters and the other two in inches; the other unit shown next to each value is our own conversion at 25.4 mm per inch and is set in italics. A maker's smallest curve piece is not treated as a locomotive minimum: that column shows only what the maker itself says about its locomotives, or "not stated on the maker pages we checked". The NMRA section on that page shows four of the curve classes in the NMRA's RP-11 recommended practice, copied as printed on the PDF page we cite, inches and millimeters both; we don't convert those figures or work out a radius from degrees of curvature ourselves.

Coupler cross-reference is not published yet. Micro-Trains already maintains its own official conversion table on its own site, and we only checked 3 of Micro-Trains' 14 brand groups — not enough to add anything an official source doesn't already cover, so no coupler page was built this round.

What "compatible" means here

The track interchange page checks one thing: whether two manufacturers publish the same rail-height ("Code") number for a given scale. If both publish it and it matches, the page marks it same rail code. If both publish it and it differs, the page marks it different rail code and says a transition piece would be needed — while also saying, honestly, when neither manufacturer publishes an official transition track for that pair, because that's a separate fact from the rail-height comparison. If either manufacturer hasn't published a Code number for that scale, the page marks the pair unknown and says so in words ("not published by the manufacturer") rather than guessing from a similar-looking line.

Official transition track — a manufacturer-published product for joining its track to another brand's — gets one of two different negative statements, not one blanket "not found." Where we could exhaustively check a manufacturer's own catalog and site search and it turned up nothing, the page says the manufacturer confirms none published; everywhere else, a negative result is stated as not found on the pages we checked, meaning we didn't check every possible official channel, not that one doesn't exist. Separately, when a manufacturer publishes a transition line only for joining its own different rail-code products together (not for connecting to another brand), the page shows it in its own section with an explicit note that it isn't a cross-brand transition track — so it can't be misread as evidence that cross-brand compatibility exists.

This comparison intentionally does not cover roadbed height, roadbed width, or power-feeder pin compatibility. None of the three manufacturers publish those measurements on their consumer product pages we checked, so a real cross-brand fit could still depend on something this site doesn't check — that limitation is stated on the track interchange page, not hidden.

This logic lives in one file in the code (src/lib/compat.mjs), and every page that shows a status pulls it from there rather than re-deciding it. The test suite for that file includes cases that intentionally feed it a broken version of the logic, specifically so a change that quietly breaks the comparison gets caught instead of a green checkmark sliding through unnoticed.

What we won't do

  • A manufacturer's marketing copy or manual text doesn't get quoted here. Every sentence gets written fresh from the structured facts pulled off the source page, and the build itself won't finish if an automated scan turns up a long stretch of any published page reading the same as the archived manufacturer text it was built from.
  • We don't reproduce official product photos, logos or diagrams.
  • We don't publish a comparison page built from data we haven't sourced for the systems it covers. If a field isn't published by a manufacturer, the page says so instead of estimating it from a similar product.
  • We don't sell anything on this site and outbound links to a manufacturer's own page are plain, un-tracked links.

Staying current

Manufacturers revise product pages without notice. Every fact on this site shows the date it was last checked, so a stale figure is visible as stale rather than presented as current forever.

Independence

Nobody at Kato USA, Atlas, Bachmann or any other manufacturer named on this site had a hand in building it, signed off on it, or gets anything from it running — this is an unofficial, independently run reference project, full stop.