• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 30 subscribers
  • Views 227 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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: Validity of the reading results

Atsushi Takahashi
Expert 1415 points
Part Number: TMP102

Tool/software:

Hi Team,

The TMP102 reading result is +90°C, but the PCB surface temperature is +50°C, which is a discrepancy.

The PCB surface temperature was measured with a thermoviewer and a thermocouple, and both results were +50°C.

Since the PCB surface can be touched by hand, they think +50℃ is correct.

Could there be any factor causing this discrepancy?

Best Regards,

  • 0
    TI__Mastermind 43785 points

    The TMP102 has an Extended Mode that could make temperature appear to double if the data is interpreted incorrectly. You would have to decode 12 bit data as if it were 13 bit data (EM=1) to get a double result.

    thanks,

    ren

  • 0 in reply to Ren Schackmann
    Expert 1415 points

    Hi Ren-san,

    Thank you for your play.

    The customer has not written to the configuration registers.

    Extended mode is disabled by default, so it seems that it is not used. don't think it is being used.

    Is there anything that you think might be the cause?

  • 0 in reply to Atsushi Takahashi
    TI__Mastermind 43785 points

    You are correct that you must write to Configuration register to enable 13 bit (EM=1) operation in TMP102. However, I stated that 12 bit data (as reported by default) would appear to be double if it were decoded as if it were 13 bit data. This decoding happens in software, outside of the TMP102 device. The customer could be expecting 13 bit data and reporting the double result when receiving (default) 12 bit data. If their software previously configured EM=1 and they recently stopped doing so, they would need to change their temperature decoding to match. 

    There may be rare situations where the TMP102 observes a glitch or noise as an additional clock (SCL) event. The additional clock would cause transmitted data to appear shifted by one or more bits. All I2C devices would be susceptible to this type of glitch. If this glitch happened during the address phase of the I2C transaction, it would cause the target device to NACK. This means there is very limited opportunity for these events to happen in the data phase, where data could be shifted, without also preventing communication from occurring.  

    Please ask the customer to share an oscilloscope capture of the I2C bus.

    thanks,

    ren