MSP430FR5969 Internal Temperature Sensor

Datasheet said:

A typical temperature sensor transfer function is shown in Figure 18-10. The transfer function shown in Figure 18-10 is only an example. Calibration is required to determine the corresponding voltages for a specific device.

Regards,
Maciej 

Yes, I've already read the calibration part but I don't know what value I must set for CAL_ADC_15T30, CAL_ADC_15T85, because CCS considers them as 2 undefined variables. I guess ADC(raw) is the value that ADC returns (ADC12MEM0 for instance). I'm sorry but I've just begun to work with these kind of MCUs.

You don't have to set CAL_ADC_15T30, CAL_ADC_15T85, they are part of the TLV structure and factory calibrated.

You shouldn't need to use a float (and all the overhead that comes along with that). Just store your cal at some factor (say, 1000x) and then divide that factor out after all the math is done.

Here is a DriverLib-based example:

   // Get the calibration data for the ADC and internal temp sensor
   TLV_getInfo(TLV_TAG_ADC10CAL, 0, &length, (unsigned int**)&ptr);

   // Get the calibration data from the chip's TLV data
   if (ptr != NULL)
   {
      calDataADC = (struct s_TLV_ADC_Cal_Data*)ptr;
      calADC20T30 = calDataADC->adc_ref20_30_temp;
      calADC20T85 = calDataADC->adc_ref20_85_temp;
      calTempScaleFactor1000x = ((INT32)(85-30)*1000)/(calADC20T85-calADC20T30);
      calADC_Gain = calDataADC->adc_gain_factor;
      calADC_Offset = calDataADC->adc_offset;
   }

And then something like:

#define ADC_TEMP_OFS_1000X 300000 // 30 degrees C is the lower calibration point
#define KELVIN_OFS_1000X 273150 // 0 degrees C = 273.15 Kelvin

UNSIGNED16 convertADCtoKelvin10x(void)
{
   INT32 result = 0;

   // Operations are done at 1000x to increase precision in the division
   // calculations below. 1000X factor is partially removed at end with
   // divide by 100, leaving result in tenths of a degree Kelvin.
   result = ( (INT32)SampleBuffer[PFM_TEMPERATURE] - (INT32)calADC20T30 ) *
             calTempScaleFactor1000x;
   result = (result + ADC_TEMP_OFS_1000X + KELVIN_OFS_1000X) / 100L;
   return((UNSIGNED16)result);

}

Thank you Kate; a last question: what is the best way to view the result of the temperature formula? I've tried to declare a "volatile int temp" in order to view it in the expressions table, but I can't view the correct result.  While ADC12MEM0 varies changing temperature, the variable temp remains the same having a busted value.

However I have declared the variable as volatile, otherwise I was unable to see it in expression table setting it as unsigned int temp (Error: identifier not found).

Thank you.

Hi Fabio,

Your variable is probably getting optimized out. You may want to try making it a global variable with the volatile keyword, or using the result somewhere to prevent the compiler from getting rid of it, or turning off optimizations for your initial testing/experimentation. See Jens-Michael's post here: http://e2e.ti.com/support/microcontrollers/msp430/f/166/p/227561/801191.aspx#801191

Regards,

Katie

Hi Katie,

I used your hints and I made my program, but when I define CAL_ADC_12T30 and CAL_ADC_12T85 (as you showed me above) I get the value 2676 and 65535 respectively. I think that CAL_ADC_12T85 is really bigger than CAL_ADC_12T30 and so the temperature formula gives always the same value because CAL_ADC_12T85 has a bigger weight in the formula. Any ideas?

Thank you again.

Fabio

Fabio Marinosci said:

when I define CAL_ADC_12T30 and CAL_ADC_12T85 (as you showed me above) I get the value 2676 and 65535 respectively.

The values cannot be >4095 for the 12 bit ADC12 or 1024 for the ADC10. 65535 is 0xffff = empty/unprogrammed.

Perfect, thank you for all. Grazie.

Fabio