This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMP102: Fail state by external stress

Part Number: TMP102

Hello Support team,

As a failure mode that can occur in the mass production process, damage due to external stress during PCB mounting or assembly can be considered.
Can TMP102 output abnormal values if it is damaged by external stress? Or may it be in a state where any output cannot be performed? Could you plesae let me know any fail state you can assume?

Best Regards,
Hirokazu Takahashi

  • Hi Hirokazu-san,

    Any IC temperature sensor (as well as all analog IC such as voltage reference) can have an output error induced by mechanical stress on the device package. For our temperature sensors, we typically observe less than 0.1C temperature error due to mechanical stresses. Mechanical stress can come from the solder process, from bending the PCB after assembly, or from pressing down on the device. 

    I would expect a device exhibiting complete lack of output due to mechanical stress to have obvious visual damage. In other words, you would need to physically destroy or separate the device to prevent it from responding.

    In case you're not just asking about mechanical stress, Electrical Over Stress (EOS) can also damage the device. Our datasheets contain Absolute Max Ratings which describe safe voltage levels. EOS damage to the SDA pin of TMP102 could prevent it from responding to temperature queries.

    thanks,

    ren

  • Hello Ren-san

    Thanks for your reply.
    Let me confirm below.
    If TMP102 may have any damage by mechanical stress, are there any case to output any abnormal temperature codes except 0x000 or 0xFFF?

    Best Regards,
    Hirokazu Takahashi

  • Hi Hirokazu-san,

    I don't expect mechanical stress to cause an electrical short inside TMP102 except in the extreme case of complete destruction.

    In the event that either of the I2C pins is stuck high or low external to TMP102, it's unlikely that 0x0000 or 0xFFFF would be read back. Most I2C hosts would recognize the failure, and refuse to begin transmission. If a very low-tech I2C host interpreted a stuck bus as 0x0000 or 0xFFFF, it would have nothing to do with TMP102.

    thanks,

    ren