TMP006 emissivity correction

 I have just received a TMP006EVM and am quite impressed. The TMP006 User guides treat the subjects of local temperature and transient effects well and the ability to take different averages in the supplied software is very useful. However, one important parameter appears to be completely ignored. This is the effect of the emissivity of the object(s) in the TMP006’s field of view. The question of emissivity is not mentioned an any of TI’s documents

 I have looked through the forum posts and seen the emissivity mentioned only once: in a post by Chip Lukes on calibration of the TMP006. It would be very helpful if the TI engineers responsible for this sensor to modify the GUI software to allow emissivity to be input as a parameter. Also, for those writing their own software (the I2C interface invites this) some guidance on how to correct for different emissivities would be useful.

3 Replies

  • Hello,
    A correction to my previous post on emissivity. Emissivity is treated in the User guide SBOU007 with several typical values given for different materials. It is also recommended that the object being should have an emissivity >0.7 and preferably >0.9.

     It is not always possible to arrange for values in the above ranges, hence it would still be useful to have TMP006EVM software accept emissivity value input as a parameter.

  • In reply to Ian Patterson:

    Hi Ian,

    Thanks for your comments! 

    The User's Guide recommends an emissivity value greater than 0.7 to ensure that the TMP006 receives enough signal from the object to successfully detect the temperature.  The TMP006 operates at a fairly low SNR and when the object emissivity is too low the the "S0" term which represents the gain of the object calculation equation, must be increased.  This causes the noise to get gained as well and the signal will begin to get lost in the noise floor of the device. 

    A 2-point calibration procedure is described in the EVM User's Guide.  This procedure is the method that should be used to correct for all objects with emissivities less than ~0.96.  The procedure is used to properly set the S0 constant in the object calculation equation.  If you are using the EVM and the EVM GUI, you can modify the "S0" constant to see the effect of how you would compensate for an object with lower emissivity.  If you are writing your own software then you will need to modify the S0 term from the default value used in the EVM GUI to the value you determine to be more appropriate.

    Please let us know how we can help.

    Regards,
    Collin Wells
    Precision Linear Applications 

  • In reply to Collin Wells:

    Hello Colin,

     I agree that the calibration procedure takes into account  the object(s) emissivity. If I were designing a control/measurement system I would certainly use at least a two point calibration and would try for more points. Ian Williams' post to this forum including an Excel spreadsheet and 430 c-code (in a zip file) were very helpful in understanding the T calculation and the role played by the calibration procedure.

     However, I see some exploratory situations where a calibration is very inconvenient  or impossible. In this case it would be helpful to allow emissivity as a parameter to the TMP006EVM software. For what it is worth there are a number of emissivity values given in the following URL:

        Emissivity Coefficients of some common Materials  www.engineeringtoolbox.com/emissivity-coefficients-d_447.html

     There are no references cited for any of the values given, so caveat emptor.

     Regards and all the best in 2012.

     Ian Patterson