PGA309: PGA309 T0...Tn Index What is the correct sequence ?

Thank you for the quick reply.
I have the EVM-board but I use it only to program the EEPROM and to control it with the register (block diagram).
Our piezoresistive capsules are already measured over the temperature and we have all the data from the sensor behavior so already.
We want to use this data to do without a calibration in the oven since this was just already made (only the capsule).
I calculate now in an excel sheet with the help of the raw data of the sensor the coefficients and all the other data for programming the EEPROM. After that we make a verification.
This works great ! The compensation works perfectly over the whole temperature range. I'm just wondering why I have to change the order of the data when I measure the temperature with an external diode connected. The compensation works perfectly under the condition that the data (Temp.ADC Counts) in the EEPROM for the temperature are ascending.
I followed the example (3.2 EEPROM Content and Temperature Lookup Table Calculation) on page 62 document SBOU024B. And the TempADC counts (Table 3-3) are inverse as soon as I use an external diode.

I found this out but is this really correct or am I making another mistake ?

Thanks for a quick reply
Roger

Roger,

1. What is the result of your calibration.  Is it complete non-functionality or just large errors?
2. One possible concern for your method:  The PGA309 is designed to be calibrated in the oven as a complete module with the sensor.  The reason that this is the recommended approach is that the errors of the PGA309 itself and the drift of these errors is calibrated along with the sensor.  Doing the full calibration assuming accurate gain and offset of the different blocks in the PGA309 will lead to errors.  Also, doing an over temperature calibration based on room temperature data and projections of the temperature sensor data will lead to drift errors.  It is possible that these errors will be acceptable to your requirements, but I want to mention this as a concern.
3. If you send me your spreadsheet, I can check the math.  The PGA does not require either ascending or descending data.
4. Note: I recently updated the document SBOU023 to revision C. The changes related to the section with table 3-3.  The issue with the old version is that some of the gain / offset DAC steps in the example exceeded the maximum allowable signed 2's compliment 16 bit integer.  The new document uses values that are in range, and also provides the comment "The allowed range for ZM is: -32768 ≤ ZM ≤ +32767 (16 bit signed integer)".  Beyond this I added section 5-4 providing the math behind the calibration procedure.  I do not think the changes are the source of your problem as the equations for Table 3-3 are correct.
5. One possibility to do a calibration without your spreadsheet would be to use SENSOREMULATOREVM.  This is the older manual version of the built in sensor emulator included in the PGA309EVM.  You could program the temperature sensor and bridge sensor outputs to match your sensor then disconnect the emulator and attach your real world sensors.  Alternatively, you could use two PGA309EVM and wire the built in sensor emulator from one PGA309EVM into the input of your PGA309 hardware.  The reason I suggest this is the most likely issue with your setup is that you are making some kind of math error in your spreadsheet.  The PGA309EVM software calibration page does all the math needed so that we can confirm your spreadsheet is correct.

I hope we can resolve your system issue soon.  Best regards, Art

Ok I will try to explain it again. Apparently I have expressed myself too imprecisely :)
To question 1
Do I program the data in the EEProm with internal temperature measurement active the T0 ... T6 values ascending then we measure a deviation over the whole temperature range from -40deg to +120deg a FSO deviation of <0.5% and offset <0.1 This is perfect and more than enough accurate.
If I program the data in the EEProm with external temperature measurement actively the T0 ... T6 values ascending then the compensation does not work at all ! The PGA always fetches the values from the temperature range T0=-40deg. So we had the suspicion that the PGA needs mandatory ascending TempADC values. Which is also the case with the internal temperature measurement. As already said with the external temperature measurement, however, the measurement results are inverse.

Solution: I have changed the pure sequence of the data so that the TempADC values are also ascending T0=120degrees .... T6=-40deg.
And now the system works perfectly !
Deviation over the whole temperature range from -40deg to +120deg a FSO deviation of <0.5% and offset <0.1 This is perfect and more than accurate.

I have no accuracy problem because the PGA reads the correct coefficients in the lookup table. I think it's not the calculation otherwise we would have accuracy problems.

It is more because of the fact that depending on the setting of the temperature measurement the data in the EEPROM is read in a different order.

Hi Art
You wrote:
"Neither ascending nor descending data is required for the PGA."
But this refutes itself. It does play a role.
Do you agree with this. That was my real question

Kind regards
Roger

Thank's a lot for the great and fast support !
Kind regards, Roger