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.

MSP430FR2033 - Internal temp sensor calibration

Other Parts Discussed in Thread: MSP430FR2033

Hello
We need to get a bit more accuracy from the internal temp sensor on the MSP430FR2033.

 We’ve made some measurements and found offsets in our temperature measurement of as much as +/-5°C, we’d prefer this to be around +/-2°C.

We’re using code based on the users guide that employs the 30 and 85°C calibration constants stored in the device descriptors:

Now we're trying to understand what accuracy we should reasonably expect based upon the datasheet (Table 5-19 : ADC 10-Bit Linearity Parameter).

Note 1 recommends a single point calibration.

Note 2 suggests using the device descriptor calibration constants, as we do.

What we're not certain on is if by employing the cal constants as per note 2, do we remove the offset error and therefore the requirement to perform a single point calibration as suggested by note 1?

 

If the device descriptor cal constants are not expected to remove the offset, then we can see this would explain our offset error.

 We also plan to double check we are complying with the tsensor min sample time, and also check we are not falling foul of errata ADC50.

 

Anything else we should look at?

Thanks
Bob Bacon

 

 

  • Hello Bob,

    Please review the following:

    1. Remember that the temperature sensor measures internal junction temperature, not ambient temperature. During the calibration process these two were equivalent, but for free air operation the power consumption and self-heating of the device will result in higher internal temperatures than ambient.
    2. The best that can be expected from each calibration value is +/-3°C according to note 2 and the internal 1.5-V reference parameters, more precise measurements may require additional calibration or an external temperature sensor.
    3. Be sure to cast data to the float type when dividing to make sure you are getting a valid multiplication ratio.
    4. Too short of a sample time can result in lower readings than expected.
    5. Allow some time for the reference voltage to settle before starting conversions.
    6. Oversampling might be utilized to increase resolution.

    e2e.ti.com/.../130904
    e2e.ti.com/.../273280
    www.ti.com/.../slaa694a.pdf

    Regards,
    Ryan

**Attention** This is a public forum