• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 1371 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.

TMP006EVM Accuracy

Jonthornham
Expert 1490 points
Other Parts Discussed in Thread: TMP006

I recently purchased the TMP006EVM.  I was excited to test it our but have found there are large temperature shifts in the readings.  I was using skin temperature as a source for readings.  I noticed temperature shifts over 20 degrees F.  Is this normal?  Am I doing something wrong?.  I'd like to clean up the accuracy but I am not quite sure how to do so.

Take care,

Jon

  • 0
    TI__Mastermind 27715 points

    Hello Jon,

    Shifts of 20°F are not normal behavior once the TMP006 has stabilized. Was the TMP006 experiencing any local transients (movement, airflow, temperature) when you observed these shifts? Was anything touching the part? Since the TMP006 die is exposed it's possible that the device has become damaged, in which case I can order you a new EVM kit to test.

    If possible, attaching a screenshot of the Graphs tab in the EVM software showing the sensor voltage, local temp, and object temp when you observe a large shift would also be helpful. 

    Best regards,

    Ian Williams
    Linear Applications Engineer
    High Performance Linear and Sensing Products  

  • 0 in reply to Ian Williams
    Expert 1490 points

    Ian,

    Thanks for the reply. I'll graph it over the next couple days and upload it. 

    Take care,

    Jon

  • 0 in reply to Jonthornham
    Expert 1490 points

    Ian,

    Ill also make sure there are no transients. I thought I was in a stable environment but I could be wrong. I know I wasnt touching the sensor. 

    Take care,

    Jon

  • 0 in reply to Ian Williams
    TI__Mastermind 27715 points

    Jon,

    Here's a screenshot of a working TMP006EVM as an example. I simply had the EVM on my desk with the TMP006 sensor side pointing upward, and I placed my hand a couple inches away from it. Around t = 100s, I removed my hand. You can see that the object temperature is accurate within ±1°C, which meets typical data sheet performance.

    Best regards,

    Ian Williams

  • 0 in reply to Jonthornham
    TI__Mastermind 27715 points

    Jon,

    What is the approximate temperature of the target object? If the difference between the TMP006 local temp and the target object temperature does not fall within the defined range of -10°C to 30°C then you will not achieve data sheet accuracy.

    EDIT: Sorry, I forgot you already said you're measuring skin temperature. The local vs. object temp difference should not be an issue.

    Best regards,

    Ian Williams

  • 0 in reply to Ian Williams
    Expert 1490 points

    Ian,

    The goal was to use the sensor to measure body temperature. However the body is warmer then 30 degrees C. Maybe that's my issue. Is there a recommended distance to take a reading?  I was holding the sensor pretty close. That may also be my issue. I'll run a series of tests tonight. I appreciate the responses. This is an incredible forum. 

    Take care,

    Jon

  • 0 in reply to Jonthornham
    TI__Mastermind 27715 points

    Jon,

    Since the body uses skin as a cooling mechanism, skin temperature in general is significantly cooler than internal body temperature. For a more accurate representation of the internal body temp, I'd recommend pointing the TMP006EVM into the ear or mouth. The ear would be the better of the two since your breath will cause local airflow and temperature transients which can degrade accuracy.

    As for the recommended distance, as long as the target object dominates the TMP006's field of view then you should be fine. Formally, our TMP006 User's Guide states that the optimal distance to the target object is (0.5 * radius of target object) or less.

    Best regards,

    Ian Williams

  • 0 in reply to Ian Williams
    Expert 1490 points

    Ian,

    I ran a series of tests tonight when I got home.  I noticed the following.  First, it took about 5 minutes for the results to calm down on any surface.  Once they results calmed down things like my hand, ear, desktop and cardboard kept a reasonable measurement.  However my forehead  or keyboard never settled down.  I played with the sampling rates and faster sampling made the stable objects worse.  They produced about an 8 degree C fluctuation.  I have attached some of the graph screen shots below.

    Take care,

    Jon

  • 0 in reply to Jonthornham
    TI__Mastermind 27715 points

    Jon,

    I apologize, but the resolution of the TMP006EVM GUI in your images is too small for me to interpret the graphs. Would you mind doing a similar experiment, but this time using Alt+Print Screen to take screen shots of only the GUI?

    Best regards,

    Ian Williams

  • 0 in reply to Ian Williams
    Expert 1490 points

    Ian,

    I worked the screen shots a little and put them under my files section in my account.  If you click on my user names and click the files tab you should see them there.  I'm sure you know that though since you work for TI.  =]

    Take care,

    Jon

  • 0 in reply to Jonthornham
    TI__Mastermind 27715 points

    Jon,

    Increasing the conversion rate does increase the peak-to-peak noise of the TOBJ reading, but the fluctuations you're seeing are higher than expected. I just did a forehead measurement with the TMP006EVM for comparison and saw a fluctuation of about ±2.5°C. I noticed in the first measurement that the TMP006 local temp changed by about 1°C, which can have a significant impact in the object temperature reading, so I re-ran the test in a newer version of the EVM software which includes a local temperature transient correction option. After enabling the correction, the fluctuation was reduced to ±1°C.  You can download the same version of the software at ftp://ftp.ti.com/pub/linear_apps/TMP006/TMP006EVM%20Installer.zip and experiment with enabling the transient correction algorithm to see if it improves your results.

    Forehead reading, transient correction disabled

    Forehead reading, transient correction enabled

    Best regards,

    Ian Williams