Free tool
Tag memory
A passive tag is an antenna, a chip and a few hundred bits of memory split four ways. Pick a chip and rewrite its EPC. Then try to change the TID, lock a bank, and see what a stranger's reader can still do.
Poke the memory
Reserved
EPC
TID
User
| The reader tries to | Result |
|---|
How it works
What you’re looking at
Every Gen2 tag chip has the same four banks. The numbers in the map are 16-bit words, written in hex, with the word’s address down the side, which is how a reader asks for them.
Reserved holds two 32-bit passwords. The access password guards the locks: with it set, a reader has to know it to write a locked bank or change a lock. The kill password lets a reader switch the tag off for good. A new tag ships with both set to zero, which means no password at all. A zero kill password means the tag can’t be killed.
Not every chip keeps both. Impinj’s M700 and M800 chips keep one password and use it for both jobs, so setting an access password also makes the tag killable. NXP’s UCODE 9 has a kill password and no access password at all. The chip picker shows both cases.
EPC is what the tag says when a reader asks. The first two words are housekeeping: a CRC the chip works out itself, and a PC word that says how long the EPC is. Then the EPC itself, usually 96 bits. It’s writable, which is how a blank label becomes a label for your item, and why you lock it afterwards.
TID is the chip’s own identity. The first 32 bits say it’s a Gen2 tag, which maker made the chip and which model it is. Then a header word, and a 48-bit serial written at the factory. The Gen2 standard has the maker lock this bank for good, so no reader can write it. That’s what makes it useful when you suspect an EPC has been copied.
User is spare room. Many chips skip it entirely. Some give you a few words; a few give you kilobytes. Most deployments leave it empty and keep everything about the item in software, keyed by the EPC.
What it leaves out
Locking in Gen2 is finer than one switch per bank. A reader can lock the kill and access passwords separately from the EPC, and can make any of those locks permanent. The map shows the two locks people actually use.
The TID serial shown here is made up. A real one you only get by reading the tag.
Where the numbers come from
Bank sizes, sensitivities and TID model codes are from the makers’ datasheets and Impinj’s encoding guide, as listed in the source file for this page. The UCODE DNA’s TID code isn’t confirmed yet, and the chip picker says so. The TID layout follows the GS1 Tag Data Standard; the lock rules follow the Gen2 air interface standard. How the EPC itself is laid out, company prefix, item and serial, is part 10 of this series.
For how a tag gets its power in the first place, see wake the tag. For how far it reads, the range calculator.