mm9z1-638：calibration issue

[gumu]

Hello NXP team:

The project is currently in mass production, and our customer has returned a faulty component where the VSENSE3 voltage is approximately 500mV higher than the actual voltage. After checking the calibration data for each cell, we discovered that the gain and offset for VSENSE3 were both unusually high. The strange thing is that we did not detect any problems during the testing of VSENSE3 after calibration in the factory. However, after a period of time, when the product was delivered to the customer, the voltage deviation for VSENSE3 became significantly larger. Below are the calibration records and some data information:

1. VSENSE0 gain 515 offset 237, VSENSE1 gain 521 offset 238, VSENSE2 gain 522 offset 251, VSENSE3 gain 512 offset 0.

2. The external reference voltage provided is 1-12V, with a total of 12 sampling points and a step size of 1V. The voltages collected by the four senses are as follows:
   VSENSE0: 1.002V 2.00125V 3.0V 3.9995V 4.99975V 5.9995V 6.999V 7.999V 8.998V 9.9975V 10.9975V 12.0055V
   VSENSE0: After calibration algorithm, the calculated gain and offset are 0 and 0.
   VSENSE1: 1.00375V 2.00325V 3.003V 4.0015V 5.0015V 6.00125V 7.00025V 7.9995V 8.9995V 9.99875V 10.99875V 11.998V
   VSENSE1: After calibration algorithm, the calculated gain and offset are 1 and -4.
   VSENSE2: 1.0075V 2.006V 3.0055V 4.0045V 5.003V 6.0025V 7.002V 8.0005V 9.0V 9.999V 10.9975V 11.997V
   VSENSE2: After calibration algorithm, the calculated gain and offset are 2 and -4.
   VSENSE3: 0.603V 1.576V 2.562V 3.554V 4.547V 5.541V 6.538V 7.534V 8.53V 9.524V 10.521V 11.519V
   VSENSE3: After calibration algorithm, the calculated gain and offset are 13 and 102.

We then wrote the calibration data into the device and tested each cell using the calibrated data after restarting. The test results are as follows:
Input voltage 3.2V, actual voltage for VSENSE0 is 3.2V.
Input voltage 6.4V, actual voltage for VSENSE1 is 6.399V.
Input voltage 9.6V, actual voltage for VSENSE2 is 9.599V.
Input voltage 12.8V, actual voltage for VSENSE3 is 12.801V.

The testing results were good.

The actual problem is that the measured voltage for VSENSE3 is 400mV higher than the actual value.

Can you provide any analysis direction for this issue? Thank you!

IFR info:

[Q_man]

Hi,

have you checked (excluded) if the effect is maybe coming from the PCB board and not our component?

One method is the ABA swapping (e.g. https://forum.digikey.com/t/part-failure-vertification-a-b-a-swap-method/701)

W.

[gumu]

Hi @Q_man ,

After receiving the faulty device, I couldn't reproduce the issue, but it was confirmed that the calibration data for vsense3 was abnormal. Later, I recalibrated vsense3 using the same environment and power device,  the data and results were normal. The current issue is that based on the calibration data at that time, vsense3's measured data is about 500mV lower than the actual input value. Only vsense3 has this lower bias, while the other vsense readings are normal. I still haven't figured out where the problem is coming from.