Dating a synth or keyboard

Not one of the 72 makers on this page publishes a serial-number table. Not Korg, not Roland, not Moog, not Sequential. We looked at every one that has a website, and the answer was the same everywhere.

So the serial number on the back is, for dating purposes, usually a dead end. What is inside is a different matter. Every chip in the instrument is printed with the week it was made, and a synth has dozens of them. Collect the codes and they bracket the build date far more tightly than any one of them could.

Why more codes is a better answer, not just more work

A chip was made before it was soldered in. So one code tells you the instrument is no older than that chip, and nothing else — the pedal page puts it the same way, because it is the same fact. One early chip on its own proves nothing at all: factories used old stock.

The newest chip is the one that constrains. An instrument cannot predate any part in it, so the latest code you find is the floor, and every older code is just stock that sat on a shelf. That is why this tool takes a list instead of one code: with twenty codes, the floor is usually within a few months of the real build date, because factories do not hold their whole bill of materials for a year.

And it only ever gives you a floor. It cannot tell you the instrument is not newer. A 1982 synth repaired in 1995 has a 1995 chip in it, and this tool will faithfully report that it cannot be older than 1995. That is not a bug — it is the honest reading of the evidence. Look at whether the suspect chip is socketed, or sits at an odd angle, or is the only one from its year.

Reading a code off a chip

Chips carry several markings and only one is a date. The part number says what the chip does; the date code says when it was made.

The quickest test there is: the last two digits are a week number, so they cannot exceed 52. If they do, what you are looking at is not a date code. That one check throws out most misreadings, and this tool applies it to every code you enter.

Four digits is much the commoner form: two of year, two of week. Three digits carries a single digit of year, which leaves the decade open — 215 is week 15 of 1962, 1972, 1982 or 1992, and the number does not say which.

Three-digit codes get resolved by the four-digit ones

This is where a list earns its keep. An ambiguous three-digit code has up to four possible years, but the instrument can only be one age. If a four-digit code puts the floor at 1982, then a 215 in the same box is week 15 of 1962, 1972 or 1982 — and 1992 would mean a part fitted ten years after the instrument was built.

The tool marks each ambiguous reading as consistent with the floor or not, rather than picking one and presenting it as the answer.

The pots are worth reading too

Potentiometers carry a different format — a three-digit manufacturer code followed by year and week — and it is decoded on the amplifier page, where it was built. It is not reimplemented here, because two copies of the same logic drift apart. A synth with original pots gives you a second, independent floor.

What the makers publish, which is nothing

The maker pool is 72 companies — synthesizer, organ, electric and electronic keyboard makers — taken from Wikipedia and listed in keyboard-makers.txt. Every one with a recorded website was swept for a serial table, a dating page or an age lookup, in seven languages.

And read the swept figure knowing how much was there to read. Of the 33 counted as swept, 20 offered this method fewer than 25 pages to look at, and 20 of those offered nothing from a sitemap at all. A host that answers, allows us, and then has almost nothing to search still counts as swept. It is the number of doors opened, not the number of rooms searched.

Zero is the finding here, and it is worth stating plainly. The piano page found two makers in 222 publishing figures. Keyboards found none in 72. That is not laziness on the makers' part so much as a difference in the instruments: an acoustic piano is a serial-numbered object built in a numbered sequence for a century and a half, and a synthesizer is a consumer electronics product.

And that caveat is not theoretical — it has since been proved. The same method was run over 146 band-instrument makers, came back empty for Ludwig, and Ludwig do publish a serial guide; their site serves no sitemap, so the sweep never reached it. It was found by hand. Read the zero above as this method found none, not as none exist.

What “nothing found” is not. It means this method found nothing. A maker who publishes a table inside a PDF, behind a search box, or four clicks from the home page would not be found by it, and would not therefore be publishing nothing. The distinction matters enough to keep saying.

Refused, or unreachable — named rather than counted

Five told us not to, and that was honoured. Electro-Harmonix and Elektron name our crawler specifically in robots.txt. Yamaha, Technics and PAiA return 403 on robots.txt itself, so their rules could not be read and the answer defaults to no. Casio served no robots.txt and then refused the home page. Nothing was taken from any of them.

Six more did not answer at all, under two different HTTP clients — Electronic Music Studios, Fairlight, Hammond, Prosoniq, Waldorf and Deutsche Wurlitzer. That is a failure to reach them, not a finding about them, and it is recorded as unassessed.

Moog — no register, but something better inside

Moog run no public serial register, and their own forum is where owners work it out between themselves. What their instruments often do carry is an inspector's tag under the panel, with initials and a test date, and assembly signatures inside the case. Where those survive they beat any serial table, because they were written at the bench on the day.

How the maker list was built, and what it nearly got wrong

“Keyboards” is not one Wikipedia category, and the obvious ways to build the list are both wrong:

• The synth list alone is 1,855 bytes and 52 names. No Hammond, no Rhodes, no Wurlitzer.
• Widening to the general instrument category caught Korg and Roland — and also Yamaha Drums and two acoustic piano firms. A wider net is not a better one.

Wikipedia files these companies by what they mostly are, not by what they make. Korg sits under guitar effects manufacturers. Roland sits under audio equipment manufacturers of Japan. Casio sits under calculators and clock brands. None of the three is in any synthesizer category, and nor are Hammond, Rhodes, Wurlitzer, Vox or Mellotron. A category sweep would have missed the three largest keyboard makers in the world and looked complete doing it.

So fifteen makers were added by name, each checked against Wikipedia first rather than assumed to exist — three of the names tried had no article and were dropped rather than invented. The list also had to have a person filter: category listings contain biographies, and two were in there. So was an article about medieval wandering minstrels, because Wikipedia's own synth list links the maker Gleeman to Itinerant poet.

Websites came from Wikidata's “official website” property rather than being guessed from company names. Guessing is how the piano page acquired two false “unreachable” entries, and it is not a mistake worth making twice.