[ESP32S3] v4.4 deep_sleep sleeping more than should (IDFGH-10088)

[KonssnoK]

Answers checklist.

• I have read the documentation ESP-IDF Programming Guide and the issue is not addressed there.
• I have updated my IDF branch (master or release) to the latest version and checked that the issue is present there.
• I have searched the issue tracker for a similar issue and not found a similar issue.

General issue report

issue is quite simple:

• we put 30 minutes sleep, device sleeps 40 minutes
• we put 1 hour sleep, device sleeps 1.5hours
• we put 1.5 hours sleep, device sleeps 3hours

Any idea on how to improve accuracy?

In the settings

- 150000 Hz if internal RC oscillator is used as clock source. For this use value 1024. 
- 32768 Hz if the 32k crystal oscillator is used. For this use value 3000 or more. 
- In case more value will help improve the definition of the launch of the crystal. 
- If the crystal could not start, it will be switched to internal RC.

should we consider the 8MHz as a 32kHz?

our RTC settings:

[igrr]

Linking #9455 which seems to be related.

[KonssnoK]

i forgot to put the v4.4 reference -> 3cec3a0

[10086loutianhao]

I will take a look at this issue

[10086loutianhao]

I see that the latest master branch has resolved this issue. Indeed, when using 8Md256, the improper PMU parameters caused abnormal counting in the lp_timer. The more appropriate parameters have been implemented on the latest master branch. (The cause of the malfunction is the same as in GH#11362)

[10086loutianhao]

If you want, you can also make direct modifications to the current code in rtc_sleep.c.
To make the requested modification in the C file on line 83,
change:
if (sleep_flags & RTC_SLEEP_USE_ADC_TESEN_MONITOR) {
to:
if ((sleep_flags & RTC_SLEEP_USE_ADC_TESEN_MONITOR) || (!(sleep_flags & RTC_SLEEP_PD_INT_8M))) {

[KonssnoK]

@10086loutianhao was this merged in v4.4?

[10086loutianhao]

@10086loutianhao was this merged in v4.4?
Yes, the solution here should be the same as the one in #11362. Please check it.