• State
  • Locked Locked
  • Replies 18 replies
  • Subscribers 102 subscribers
  • Views 6328 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM335x - supporting RTC_ONLY mode using TPS65217?

AnBer
Mastermind 30120 points
Other Parts Discussed in Thread: TPS65217, TPS65910A, TPS65910, SN74AUP2G08, TPS51200

Team,

Could you please help with the following:

System is powered from battery only. Trying to use the RTC_ONLY mode to power off and on again the system and keeping the RTC alive.

------ 2nd description
We use "ALARM2" to pull PMIC_PWR_EN low, so the TPS65217A enters sleep mode.
Approximately 1 second after power down, the device reboots.
We can see something that looks like a glitch, pulling PMIC_PWR_EN high, which might be what causes the reboot.
Is this a known issue, with TPS65217A and AM3352ZCE, that it does not stay in sleep mode?

If we change the "OFF" bit in the PMIC "STATUS" register, the device no longer reboots, but there is no power to the RTC.
We then changed the PMIC power down sequence, so it doesn't turn off the power rail supplying the RTC and VIO + VCORE.
However LDO3 and LDO4 changes to LS1 and LS2, so presumably, at least some of the TPS65217A registers change values to their reset state.
When pressing the push button everything is powered off, instead of booting the device.
None of this behaviour is described in the documentation.
Any idea what is happening or what we can do to achieve power on RTC and VIO+VCORE only?

-------- 1st description
We are playing around with RTC and sleepmode, trying to figure out a way to preserve RTC date & time and still save some current BUT we need your help to figure out what is wrong.
We can get the system to power OFF.
If we configure the system to go to sleep (using RTC Alarm2) with standard settings for TPS65217 sequencer, the system enters sleepmode for approx.. 1 sec. then the RTC_PORz goes low, probably resetting the RTC logic and then PMIC_POWER_EN goes high again and the system reboots.
If we configure the system to go to sleep (using RTC Alarm2) with all power supplies ON (changing Sequencer settings in TPS65217, except for LDO3 which in this case is not used for the processor) we can get the system (TPS65217) to enter sleep, however it seems like the LDO4 changes state to LS2 (and thereby a different voltage level) and we are not able to exit sleepmode using RTC Alarm timer.
    VRTC stays at 1V8
    RTC_PORz stays high
    PMIC_POWER_EN goes low when entering sleepmode and stays low
    nWakeup from TPS65217 stays high
(see SLVU551E page 2 for the connection between AM335x and TPS65217).
If we then push the power button (low) then the TPS65217 enters OFF state after approx. 2 seconds this behavior is not described in the Figure 1 state diagram of slvs64e
What are we missing since we are not able to get the RTC circuit to pull PMIC_POWER_EN high ? (read how do we exit sleepmode ?)


- I have seen some other post but it does not seem to be the same issue:
http://e2e.ti.com/support/dsp/sitara_arm174_microprocessors/f/791/t/220444.aspx
Is RTC_ONLY operation supported on silicon rev 1.x provided the different silicon errata listed here?
http://e2e.ti.com/support/dsp/sitara_arm174_microprocessors/f/791/p/223272/789667.aspx#789667

Thanks and best regards,

Anthony

  • 0
    TI__Guru**** 393215 points
    Hi Anthony,
     
    I already sent you an email on this. I don't think your customer will be able to implement RTC-only mode on silicon rev. 1.x, due to Errata 1.0.5.
  • 0 in reply to Biser Gatchev-XID
    TI__Guru**** 393215 points
    I would also like to add, that Errata 1.0.5 could influence also EXT_WAKEUP:
     
    "A wakeup event applied to the EXT_WAKEUP input terminal is masked if the EXT_WAKEUP_DB_EN bit in the RTC PMIC register (0x98) is set to 1 which enables the de-bounce function for the EXT_WAKEUP input. This occurs because the 32.768-kHz clock is being used to clock the de-bounce circuit."

  • 0
    TI__Genius 16690 points

    Please make sure that PMIC VIO power supply input pin is connected to PMIC VLDO1 LDO output to keep RTC_PGOOD active while the system in the RTC-only mode. If it's powered by different power supply which goes off in the RTC-only mode, RTC_PGOOD goes low and AM335x RTC gets reset. 

  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Hello Biser.

    We are designing with AM335x using DDR3 with TPS65910A3. If we would like the possibility to have a rtc in either AM335x, or TPS65910A3, which should have a backup battery (or supercap), and continue to run at some low power mode, is then the only option to implement this in the TPS65910A3, due to the errata 1.0.5?

    If implemented in TPS65910A3 (or AM335x if this had worked), would the rtc continue to operate if main power is disconnected and only the rtc battery is connected? Or does main power need to be connected and the system operating in low power mode?

    The schema of TMDSSK3358 has a separate RTC LDO which supplies the VRTC. Does this circuit fulfill the requirement?

    It's a bit strange if both the cpu and the power chip contains a rtc but both alternatives are a bit tedious to work with. Then one should need a separate rtc chip (which seems a bit silly).

    Br Peter

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    Hi Peter,
     
    This question has recently been asked a lot. Errata 1.0.5 covers AM335X processor silicon Rev.1.x. It should be resolved in the next silicon revision.
    For the currently implemented AM335X internal RTC there are two major issues:
     
    One is Errata 1.0.5 - 32kHz clock gets gated off when VDD_CORE voltage is not present or PWRONRSTn input terminal is held low.
     
    The other is the relatively high VDDS_RTC consumption (0.35mA typ.) even when RTC is not used. This is also the reason why TPS65910A and A3 need an additional LDO on VDDS_RTC. Their internal RTC LDO is in low power mode (0.1mA max.) when the PMIC is in OFF mode and simply can not provide the necessary VDDS_LDO current.
     
    The TPS65910A3 internal RTC will work when main power is disconnected and only the RTC battery is connected. The PMIC has internal circuitry that switches between main and backup power and also provides charge current for a backup battery.
     
    The separate RTC chip is a solution that can cover both current and next AM335X silicon. When the next silicon is available (hopefully with impoved RTC) then this external RTC will be removed from BOM.
  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Hello Biser, thanks for answer.

    Please address some follow up questions in order to get this clear beyond doubt. Referring to the TMDSSK3358 with separate LDO for VDDS_RTC.

    1. If we decide *not* to use the RTC-functionality within AM335x (because in reality it can't be used due to errata), and not to use RTC-functionality within TPS65910A3, but use separate RTC-chip, is then the separate LDO still needed? Meaning, is AM335x, or some other part in TMDSSK3358, still dependent on VDDS_RTC?

    2. Trying to summarize the above thread for PMIC, if we decide to use the RTC in TPS65910A3, should this impose a problem? (Perhaps LDO needed anyway depending answer from question 1.) Easy/feasible to implement with RTC backup battery/super cap?

    Br Peter

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    Hi Peter,
     
    1. AM335X needs VDDS_RTC power supply even if internal RTC is not used. Regretfully, the consumption from VDDS_RTC is high in silicon Rev.1.X (0.35mA typ,), so an external LDO will be needed if TPS65910A3 is used. If you use TPS65910A31 you can do without the external LDO. Check out this document: http://www.ti.com/lit/ug/swcu093b/swcu093b.pdf
     
    2. I see no problem here. You will be able to access it over I2C, and backup will work.
  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Thanks again for fast reply.

    Why was not TPS65910A31 used instead of TPS65910A3 in TMDSSK3358 (and the LDO could be omitted)?

    Br Peter

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    I really have no idea. Perhaps it wasn't available then.
  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Hi Biser.

    I have studied the swcu093b.pdg as well as the "TPS65910, TPS659101 Data Manual Version Q (Rev. Q) swcs046q.pdf", and some clouds still remains.

    First, the main (only?) difference between TPS65910A3 and TPS65910A31 is that A31 will supply full power to the VRTC in OFF state, while A3 will not. (?)

    Then, when trying to connect this info to the different states of the state-machine at page 41 of "TPS65910, TPS659101 Data Manual Version Q (Rev. Q) swcs046q.pdf", I don't really get the main advantage of the OFF state compared to the Backup state, or the correlation to the No supply state.

    The text says that Backup state is when the main battery supply voltage is high enough to enable the VRTC domain, but not high enough to switch on all resources (what resources? the manual doesn't explicitly tell). Is then the Backup state not dependent on a power supply source, either coin battery or super cap, on VBACKUP?

    In the Backup state, A31 will not be able to supply MCU VDDS_RTC with enough power, but then, the MCU is not running (under reset?)?

    The No supply state; I recon this will continue to operate the PMIC built in rtc if there is a VBACKUP source?

    In which occasion should one need the PMIC A3 version and in which occasion should one need the A31 version?

    If the AM335X EVM 15x15 Baseboard should be equipped with a PMIC A31, then the U5 could be omitted, and R557 could be DNI and R53 could get installed?

    (In some forthcoming version of either the PMIC silicon and/or the MCU silicon, one could try to streamline all the different power on reset logic between the PMIC and MCU and try to get rid of the various SN74AUP2G08 on the different schematics (EVM, SSK, Beagle, etc.). )

    The above is the main obstacles before being able to finish our schematics.

    Then, if it is ok to mix up the thread with some other issues, please see the below questions. If not, I can submit another thread.

    1. The VTT regulator TPS51200 of TMDSSK3358 schematics rev 1.2A sheet 7; is this needed? I suppose it is. Why couldn't the PMIC handle this?

    2. The power on switch SW5 (same schema) on page 5; I have seen different threads and questions regarding this switch in the forum. Our design shall start/boot when power (VBAT) is applied and there should be no need to press any button for this. Will this be the case if we follow the TMDSSK3358 to some degree with the SW5 taken away?

    3. We are (obviously) not using the rtc in AM335x. I have in forum threads seen some recommendations on how to deal with the not used AM335x RTC_XTALIN. Some post said that TI will come out with new recommendation that this input should be left not connected if not used, due to internal pull-up which could add some unwanted current consumption. Is then the latest recommendation to leave RTC_XTALIN not connected if not used?

    Br Peter

     

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    Hi Peter,
     
    First, the main (only?) difference between TPS65910A3 and TPS65910A31 is that A31 will supply full power to the VRTC in OFF state, while A3 will not. (?)
    Yes, that's right.
     
    The text says that Backup state is when the main battery supply voltage is high enough to enable the VRTC domain, but not high enough to switch on all resources (what resources? the manual doesn't explicitly tell).
    The PMIC power resources.
     
    Is then the Backup state not dependent on a power supply source, either coin battery or super cap, on VBACKUP?
    If you check Figure 13 in the TPS65910A datasheet, Backup state is when (Main Battery OR Backup Battery > VBNPR) AND (Main Battery < VMBLO).
     
    In the Backup state, A31 will not be able to supply MCU VDDS_RTC with enough power, but then, the MCU is not running (under reset?)?
    Yes, that's right.
     
    The No supply state; I recon this will continue to operate the PMIC built in rtc if there is a VBACKUP source?
    No supply is when both Main Battery AND Backup Battery < VBNPR. If Backup Battery > VBNPR then it's Backup state.
     
    In which occasion should one need the PMIC A3 version and in which occasion should one need the A31 version? If the AM335X EVM 15x15 Baseboard should be equipped with a PMIC A31, then the U5 could be omitted, and R557 could be DNI and R53 could get installed?
    Yes, that's right.
     
     
     
     
    1. The VTT regulator TPS51200 of TMDSSK3358 schematics rev 1.2A sheet 7; is this needed? I suppose it is. Why couldn't the PMIC handle this?
    A VTT regulator has the ability to both source and sink current. This is a must for the VTT rail.
     
    2. The power on switch SW5 (same schema) on page 5; I have seen different threads and questions regarding this switch in the forum. Our design shall start/boot when power (VBAT) is applied and there should be no need to press any button for this. Will this be the case if we follow the TMDSSK3358 to some degree with the SW5 taken away?
    You will need this switch only if you intend to cycle power on/off. The PMIC will startup properly when power is applied for the first time, without the need to press this switch. Check Figure 4 in the TPS65910A datasheet.
     
    3. We are (obviously) not using the rtc in AM335x. I have in forum threads seen some recommendations on how to deal with the not used AM335x RTC_XTALIN. Some post said that TI will come out with new recommendation that this input should be left not connected if not used, due to internal pull-up which could add some unwanted current consumption. Is then the latest recommendation to leave RTC_XTALIN not connected if not used?
    I would suggest you tie it to GND through a 0Ohm resistor. The new recommendation is not official yet.
  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Thanks Biser for informative reply.

    A short addition; what is the advantage/purpose then of using TPS65910A3 instead of TPS6510A31?

    Br Peter

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    They are simply versions of the device with different EEPROM settings.
  • 0 in reply to Biser Gatchev-XID
    Intellectual 640 points

    Ok. But I am missing the main point here.

    When (what is the application requirements) do one need the A3 instead of A31? Is it to decrease power consumption when rtc is not needed in the off mode for some other MCU then AM335x?

    When A3/A31 is used with AM335x, the A3-version seems only to add an extra LDO, while the A31 can do without the extra LDO. As I understand, AM335x will need the VDDS_RTC in all circumstances anyway.

    Br Peter

  • 0 in reply to Peter T-Innova
    TI__Guru**** 393215 points
    Yes, VDDS_RTC must be present. Whether you use A3+LDO or A31 is up to you.
  • 0 in reply to Biser Gatchev-XID
    TI__Guru 62855 points

    I will confirm the new recommendation for RTC_XTALIN.

    When RTC oscillator is not used, the RTC_XTALIN terminal should be open circuit.

    Our AM335x power management experts are planning to add some example RTC power supply recommendations to the schematic check list.

    However, the new RTC_XTALIN recommendation will be put into the new revision of the data sheet.

    Regards,
    Paul

  • 0 in reply to Biser Gatchev-XID
    Intellectual 720 points

    Hi Biser,

    as far as I know the Errata 1.0.5 will be fixed on silicon Rev. 2.0. Can you tell me how much the typical / max current consuption in RTC only mode will be ? We are planning to use a backup battery and we need this value to chose the battery.

    Thank you for your reply.

    Reto

  • 0 in reply to reto hermann
    TI__Guru**** 393215 points
    Hi Reto,
     
    Rev.2.0 is still being validated. I don't have any details yet. I do know that the datasheet is expected to be updated soon.