• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 31 subscribers
  • Views 749 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.

TMP006 Accuracy at cold

Arthur Thompson
Prodigy 105 points
Other Parts Discussed in Thread: TMP006

My TMP006 EVM is giving me poor accuracy at -40C ambient temperature, but the accuracy greatly improves at ambient to hot temperatures.  Please see the EVM software screenshots I've uploaded to my files.

My setup is as follows:  The TMP006 EVM is clamped to view the back side of a Series 450 DMD (the gray ceramic package) at a distance of about 1/8".  I have glued a surface mount resistor to the edge DMD to act as a heater, and I'm running the heater at a little under 1W.  There is a thermocouple on the opposite side of the DMD from the TMP006's viewing area.  The resistor heater is an equal distance from the TMP006 test point and the thermocouple test point.  The whole setup is inside a temp chamber (I used the extender ribbon cable so the USB interface board is outside the temp chamber).  This setup allows me to heat the DMD to about 10C hotter than the temp sensor PCB or the ambient temperature.  I can move the ambient temperature to wherever I want by setting the temperature chamber.

Accuracy is within about 2 C (thermocouple measurement vs TMP006 measurement) for ambient temperatures from 10C to 70C.  At colder temperatures, accuracy starts to drop, so at a -40C ambient the thermocouple reads -30C and the TMP006 object temperature reads -50C for an error of about 20C.

Please let me know any suggestions for debugging this further or if I am somehow violating a spec of the part or the EVM.

Thanks,

Arthur

  • 0
    TI__Expert 7945 points

    Hello Arthur,

    Thank you for the detailed description. I do not see any screen shots attached in your post. I have a few questions that would help me answer your problem better. When you are referring to the ambient temperature you mean the temperature of the DMD correct? Did you calibrate your setup and update the S0 coefficient or are you using the provided default values?


    Also, it would be great if you can attach a picture of your setup. it would help me with the understanding.

    Best Regards,

    Abhi Muppiri

    Applications Engineer

    AIP- Sensing Products

    Texas Instruments

  • 0 in reply to Abhishek Muppiri
    Prodigy 105 points

    Abhi,

    Please note that this test takes place in a temp chamber, thus the ambient temperature is the air temperature inside this chamber.  The DMD is hotter than the air temperature due to the heater resistor shown in the picture. The thermocouple confirms that the DMD is about 10C hotter than the ambient air temperature.

    I have not calibrated so I'm using S0=5e-14.  If you have access to the "My Files" section under my profile, I have linked the screenshots and graphs there.

    Thanks,

    Arthur

  • 0 in reply to Arthur Thompson
    TI__Expert 7945 points

    Hello Arthur,

    Thanks for providing the pic of your setup. I can now relate to your previous email describing the setup.

    Accuracy will drastically improve when you calibrate the part to find out S0 coefficient that works best for your setup. Please keep in mind that you need to keep the ambient temperature constant when you calibrate for S0. 

    Hope this helps!

    Best Regards,

    Abhi

  • 0 in reply to Abhishek Muppiri
    Prodigy 105 points

    Abhi,

    I would like to point out that the type of error I'm seeing cannot be corrected with a better linear approximation.  Note that the line currently "fits" for ambient to hot temperatures with the constant I'm using now.  But the line does not fit for cold temperatures.  There is no single line that could fit these data; there is something nonlinear happening when you transition from hot to ambient to cold.  Can you take another look and see if you agree?

    Edit: I'm adding a graph showing the nonlinearity.

    Thanks,
    Arthur

  • 0 in reply to Arthur Thompson
    TI__Expert 7945 points

    Hello Arthur,

    Please make sure the system's ambient temperature is constant when you calibrate by measuring the object at 2 different temperatures. I have created a few slides that will help you with calibration the process. Let me know if you have any questions.


    Best Regards,

    Abhi

    0724.Improve Accuracy with Software.pptx


  • 0 in reply to Abhishek Muppiri
    Prodigy 105 points

    Abhi,

    Are you saying I need to hold Tdie constant across various object temperatures?  The user's guide does not say anything to that effect.  Regardless, all a better calibration is going to give me is S0, and the graph I posted should make it obvious that no value of S0 will work to fit both the cold data and hot data.  The problem we need to be focusing on is why the hot data and cold data have different slopes, why the line has a bend in it.  If I fit S0 to the cold data, it's a slope of 2E-13.  If I fit it to the hot data, it's a slope of 4e-14.  No compromise value is going to give me good results across all temperatures.  The typical values are supposed to be between 5e-14 and 7e-14, so if one of the two calibration values is to be trusted, I'd trust the 4e-14 one.  So the problem we need to work on is why the cold values are so strange.

    If you could, please send me a PM and we can set up a phone call to discuss, I think we could make more progress than with this back and forth.

    Thanks,

    Arthur

  • 0 in reply to Arthur Thompson
    TI__Expert 7945 points

    Arthur,

    I concur with you that the user guide document doesn't mention Tdie/Tambient has to be kept at a constant temperature in order to get a correct S0 slope. For a given Tdie/Tambient temperature you will always obtain the same S0 slope values as opposed to obtaining an incorrect or negative slope when you fit the equation using readings at different ambient temperatures.

    Another important thing to consider is how fast you decrease the temperatures to -40degC.

    My email id is abhishekmuppiri@ti.com, do email me with your contact details so we can schedule a phone conversation to get this resolved ASAP.

    Best Regards,

    Abhi