Tool/software:
We are using TMP117AIYBGR parts in our sensor modules since few years, and now we would like to use TMP117 EEPROM content as unique ID for our internal traceabilility. About a year ago I checked few sensors and they had all 3 EEPROMx registers with non-0 content (true 48-bit ID). But in recently produced modules, the EEPROM1 register is read-out as 0x0000 - directly after soldering (no testing with some host software applied). We were always ordering the same part TMP117AIYBGR (marked as "E7").
I was using my own TMP117 device drivers on host MCU.
However, my colleague was performing in the past some duration testing with this Arduino library https://github.com/NilsMinor/TMP117-Arduino/tree/master where I have found interesting comment (which is more than non-obvious from reading TI TMP117 datasheet) : if CONFIGURATION register is written with unlockEEPROM, the u16 content is also written into EEPROM and will be restored on chip reboot. It is still unclear for me: is the CONFIGURATION register the only one stored into (which?) EPROMx or similar applies to say "temperature offset register" with EPROMy...
My suggestion that one of TI factories is performing some tests with similar library - which restores CONFIGURATION on repower - and then forgets to restore the EEPROM1 content.
Another explanation could be - there are different TMP117 ordering options: for customers who needs NIST traceability (supposed all 48-bit, or at least EEPROM1 and EEPROM3 are mentioned in datasheet as essential, while EEPROM2 as user scratchpad) and another ordering number for the customers who prefer to have "restore last used mode on repower". Unfortunately, if such ordering options exist indeed - they are not obvious from datasheet or TI website...
Could you please explain the situation with EEPROM1 content 0x0000, and how can we order the "proper" chips with full 48-bits unique content?
