Water Heater Age Decoder

• My Water Heater (Determine The Age of My Water Heater) (bradfordwhite.com/my-water-heater/#age-of-water-heater), Live consumer page, retrieved 2026-07-25. Published on bradfordwhite.com, fetched 2026-07-25, md5 a3dbbf36394e55cd84c80577b47d7c98, 92,525 bytes.
• Determining The Age Of A Water Heater (For the Pro bulletin, posted 05/24/2020 (updated 07/19/2024)), Retrieved 2026-07-25. Published on forthepro.bradfordwhite.com, fetched 2026-07-25, md5 1d15131b660135662614295f5651fb67, 126,929 bytes.
• How to read the serial number date code reference chart? (bradfordwhite.com FAQ), Retrieved 2026-07-25. Published on bradfordwhite.com, fetched 2026-07-25, md5 64c7c2f5fbe91c48d3a8654625c45727, 76,521 bytes.

"The first two characters of the serial number represent the year and month of manufacture. The remainder of the serial is a sequential production number, seven digits in length before December 2007 (DM), and eight digits in length after."

Defect in the document: Bradford White publishes THREE different year tables across its own pages and they do not agree. The consumer page prints A = 2004 or 2024 through Z = 2023 or 2043. The pro bulletin prints a first table whose A row reads A = 2004 or 2024 while every other row of that same table is the older window (B = 1985 or 2005), and then a second table that is entirely the older window (A = 1984 or 2004 through Z = 2003 or 2023). The pro bulletin's first table is therefore internally inconsistent: its A row was rolled forward twenty years and the other nineteen rows were not. How we handle it: None of the three windows is treated as authoritative. The letter is read as a position in a 20-letter cycle anchored at A = 2004, and EVERY year in that cycle from 1984 to the current year is generated as a candidate. That reproduces all three published tables rather than choosing between them, and it is why the tool can still be right after Bradford White rolls the window again.

Defect in the document: The digit rule's boundary month is not resolved. "Seven digits before December 2007 (DM), eight digits after" does not say which length December 2007 itself carries. How we handle it: December 2007 is admitted under BOTH lengths, so the boundary month is never used to eliminate a candidate. The cost is one extra ambiguous case; the alternative is silently excluding a real year on a reading the document does not support.

• How to Locate and Read Rheem/Ruud/Richmond Serial Numbers (Technical Service Bulletin 1207 (1207.doc), page 1), 2 pages, undated on its face; describes changes effective April 2 and June 4, 2012. Published on rheem.com, bytes served by cdn.globalimageserver.com (FetchDocument.aspx?ID=67e23d33-c085-48fa-8aa8-eab53d38296b), reached from https://www.rheem.com/whtechbulletins/, fetched 2026-07-25, md5 bc12e633adf9f7720a520a4a35ec3baf, 361,294 bytes.

Page 1 prints the worked example "09 | 12 | D | 01234" above the labels "Month Manufactured | Year Manufactured | Plant Line Code | Consecutive Serial Number".

Defect in the document: Page 1 says the ten-digit format applies to "all rating labels", and page 2 of the same document then describes a different format replacing it from April 2012. The word "all" is scoped by the page that follows it. The document also says "ten digits" for a string whose fifth character is a letter. How we handle it: The pre-2012 pattern is matched only when the serial is digits-first, which is disjoint from the 2012-onward letters-first shape, so the two schemes cannot claim the same serial. No upper bound is imposed on this format: the bulletin says each plant migrated "as they are migrated to Oracle" and gives no completion date, so inventing an end date would be inventing a refusal.

• How to Locate and Read Rheem/Ruud/Richmond Serial Numbers (Technical Service Bulletin 1207 (1207.doc)), 2 pages, undated on its face; describes changes effective April 2 and June 4, 2012. Published on rheem.com, bytes served by cdn.globalimageserver.com (FetchDocument.aspx?ID=67e23d33-c085-48fa-8aa8-eab53d38296b), reached from https://www.rheem.com/whtechbulletins/, fetched 2026-07-25, md5 bc12e633adf9f7720a520a4a35ec3baf, 361,294 bytes.

"Effective April 2, 2012 product being produced in Mexicali started using a new rating label... You will notice a different format based on the week and year of manufacture." Worked examples: "M141209135 Mexicali plant, week 14, 2012" and "Q231299996 Nuevo Laredo plant, week 23, 2012".

Defect in the document: The bulletin's prose example is the bare serial Q231299996, but the rating-label photograph on the same page reads SERIAL NO. RHLNQ231299996. The real label carries a four-character product prefix ahead of the plant code that the written key never mentions. This was visible only in the rastered page; the text extraction drops it. How we handle it: The 2012-onward pattern accepts an optional leading letter run and reads the plant code as the LAST letter before the digits, so both the bulletin's bare example and the label's prefixed form decode. Anchoring on the plant letter would have refused real equipment.

Defect in the document: The bulletin's own example disagrees with its own label photograph by about a week. It reads Q231299996 as "week 23, 2012", while the pictured label prints MFG. DATE: 30MAY2012, which falls in week 22 on an ISO calendar. Rheem's manufacturing week numbering is not stated on the document. How we handle it: Week-format decodes are published as an approximate month and never as an exact day, and the tool says the week is approximate. A one-week manufacturing-calendar offset does not move the month in most cases and never moves the year, which is what the warranty and age questions turn on.

The first characters of the serial number carry the date. Bradford White uses a year letter followed by a month letter; Rheem, Ruud and Richmond use either a month-then-year pair of digit groups on pre-2012 labels, or a plant letter followed by week and year on labels from 2012 onward. Enter the serial above and the tool shows the date, the age, and which characters it read to get there, or tells you that this particular serial cannot be resolved to a single date.

Because the serial genuinely does not contain one. Bradford White's year letters run through a 20-letter cycle, so the same letter means 1987 and 2007, and will mean 2027. Where the production number's length does not settle which, more than one real answer survives and picking the newer one would be a guess. A wrong year here drives a warranty claim and a repair-or-replace decision, so the tool shows both and says the rating plate settles it.

They are 20-year windows onto the same repeating cycle, published at different times. Bradford White's consumer page currently prints A = 2004 or 2024, while its For the Pro bulletin prints A = 1984 or 2004. Those two are each honest snapshots of one cycle. The bulletin's FIRST table is a different matter: its A row was rolled forward twenty years and its other nineteen rows were not, so that table is internally inconsistent rather than merely dated, and the receipt on this page says so. This tool reads the letter as a position in the cycle and generates every year that position can mean, which is why it does not go stale when the window rolls again.

Covered today: Bradford White, Rheem, Ruud and Richmond. Not yet covered: A.O. Smith, State and Reliance. A brand appears here only once we have read its date code off that manufacturer's own document. Where we have not, the tool says so rather than borrowing another manufacturer's scheme, because a scheme applied to the wrong brand returns a confident wrong year.

It tells you the input, not the answer. Rheem's service bulletin 1207 says of Rheem, Ruud and Richmond: "Your warranty is based on the date of manufacture and the warranty code listed in the model number." That is one manufacturer's statement about its own products, so this tool shows it only on those three brands. For any brand, take the decoded date to the manufacturer's warranty verification; their answer is the one that settles a claim.