PMIC_PWR_EN signal of AM335x

I'm also confused.

The AM335x EVM uses the TPS65910 PMIC, which has a DEV_OFF register bit to transition the power state machine.

As another point of reference, the BeagleBone uses the TPS65217, which doesn't even have such DEV_OFF functionality to transition to OFF.  At least with the TPS65217, it seems the only way for the AM335x processor to turn things off is to somehow de-assert PMC_PWR_ENABLE.

The language in SPRUH73E (May 2012 revision) describes the PWR_ENABLE_EN bit of RTC_PMIC as:

"When enabled, pmic_pwr_enable will be controlled by ext_wakeup[3:0], alarm, and alarm2. ON to OFF (Turn OFF). By ALARM2 event OFF to ON (Turn ON). By ALARM event OR ext_wakeup[n] event"

Does that mean "ext_wakeup[3:0], alarm, and alarm2" can cause a "Turn OFF" and only "ALARM2" can cause a "Turn ON"?  If so, I don't know what the "By ALARM event OR ext_wakeup[n] event" sentence fragment at the end means.

FYI:

./drivers/rtc.c in the StarterWare install is ostensibly an implementation of controlling PMC_PWR_ENABLE, but it doesn't seem to work as the author intended.

For starters, the function RTCConfigPmicPowerEnable() is just a standalone implementation; it does not seem to be invoked by any of the StarterWare code.  That suggests that it hasn't necessarily been tested.

The comments indicate passing RTC_PMIC_PWR_DISABLE as a parameter is "to disable PMIC power".

However, the functions behavior with "disable" is just to clear the PWR_ENABLE_EN bit in RTC_PMIC, which the AM335x Technical Reference Manual indicates results in "pmic_pwr_enable will always be driven as 1, ON state".  ("disable" is the opposite of "ON")

Moreover, the function's behavior is to always temporarily clear PWR_ENABLE_EN, since it performs two writes to RTC_PMIC (instead of computing the new register value with a local variable).  If clearing PWR_ENABLE_EN it did actually control the PMC_PWR_ENABLE pin, it would cause the pin to glitch (temporarily telling the PMIC to disrupt power to the host CPU).  At least some of the PMICs seem to have de-glitching, so that bug may get concealed, but it wouldn't be good practice.

Hi,

Even we are little bit confused with the PMIC_PWR_EN pin.
Will it be required for the AM335x's Power-Up Sequencing as mentioned in the datasheet's
"4.1.1 Power-Up Sequencing" (is the below sequence correct in this case).
Our customer is designing their custom board with the below sequence.
They won't use the RTC mode feature and if the PMIC enable/disable is not necessary
can we just use the PWRON pin on the PMIC to start the Power-Up Sequencing?

　・1.8V Power is supplied to VRTC

　　　↓

　・RTC_PWRONRSTn

　　　↓

　・AM335x RTC is enabled and outputs PMIC_POWER_EN

　　　↓

　・ PMIC_POWER_EN enables the PWRHOLD and starts the PMIC

　　　↓

　・PMIC executes the necessary sequence and PMIC_RESETOUTn is high

　　　↓

　・AM335x receives PWRONRSTn

　　　↓

　・After a certain time WARMRSTN goes Low to high with weak pull-up

　　　↓

　・ After a certain time Internal Chip Resetn

Regards
prad

Hi Biser,

Thank you so much for the information.

Regarding the PWRON pin on the PMIC IC, as we don't need press and hold button,

Is it ok to leave PWRON as "Not Connected" or is it required to
pull up with 10K as mentioned in the "TPS65910Ax User's Guide For AM335x".

Regards.
Prad