I want to measure the Coulomb count of the MC3371 during its sleep period. Before going to sleep, I cleared the relevant registers of the coulometer and configured CYCLIC_TIMER = 2S at the same time. After sleeping for 16 seconds, I woke up the MC33771 using CSB and then read the three registers -cc_nb_samples, COULOMB_CNT1, and COULOMB_CNT2. CC_NB_SAMPLES is always 65535, and COULOMB_CNT1 and COULOMB_CNT2 are normal values. Shouldn't the count of CC_NB_SAMPLES during hibernation be based on the CYCLIC_TIMER?
The relevant screenshots are as follows
This waveform is the voltage waveform of VANA:
Dear Ye Xiaoer,
if coulomb counting during the entire hour is not critical, you can disable coulomb counting in sleep mode by setting the I_MEAS_EN=0 and enable it only during short active windows. This avoids overflow but sacrifices continuous integration.
With Best Regards,
Jozef
Dear Ye Xiaoer,
CC_NB_SAMPLES is not tied to CYCLIC_TIMER ticks.
It is the raw number of ADC2 samples that contributed to the coulomb integration. In low‑power/sleep operation the coulomb counter can continue integrating independent of your cyclic timer, so CC_NB_SAMPLES can quickly saturate to 0xFFFF unless you read/reset it often enough or force a fresh conversion after wake. Check the CC_OVR_FLT flag to confirm an overflow.
Please refer to the MC33771C datasheet and to the AN12633.
With Best Regards,
Jozef