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.

TWL6030 MANUALS - RTC ISSUE

Other Parts Discussed in Thread: TWL6030, TPS65950

We are trying to debug issues on Gumstix board which uses the TWL6030 for its power management and also RTC.
I contacted out local Ti Distributor to get the TWL6030 datasheets but no answer from them as to how to get the datasheet.

We are in Perth Western Australia

How do we get the NDA so we can get the datasheet and debug the configuration to determine an issue with the RTC dropping the time by indicating POR despite the RTC Battery being at suitable voltage (2.6V).  There is insufficient information in the general documentation on the TWL6030 page and no indication how we get the data, only on post on a read only forum which is someone else asking how to get the register map. 

Need to see if the Linux kernel is correctly configuring the RTC as we shouldn't be losing the clock but are and we need to retain the clock!

Help please!

  • Hi Barry,

    Documents can be provided once the NDA has been signed. Please work with your local sales representative to get this set up and they will be able to provide access at that point. If you don't know your local contact, please use our support page: http://www.ti.com/guidedsupport/docs/calltisupport.tsp

    In the meantime, I will try to help as best I can until it is setup. Some technical background questions:

    How do you identify that POR is occurring? Register 0x11 bit 7 in address 0x48?

    If so, is the bit 7 (POWER_UP) being cleared before main battery is removed?

    Are you making any changes to register 0x10 in address 0x48?

    Have you been able to take scope shots to ensure that the VBACKUP isn't dipping below 2.1V transient? 

    How close is the input cap for VBACKUP?

  • Just some further information for you Kevin

    There is no capacitor on our RTC Backup line.

    Gumstix recommendations are simply:
    VBACKUP provides battery backup to the real-time clock. It should be left open or connected to a rechargeable lithium coin cell which is trickle charged when the Overo is powered on. Do not tie this pin to ground as suggested in the PMIC documentation without disabling the trickle charge feature in software.

    TWL6030 datasheet on TI website gives no recommendation as to the backup battery circuit requirements.
  • Hi Barry,

    As a generic recommendation I always suggest input caps on all power inputs, though for a VBACKUP line it likely has no impact. As long as the battery is placed near the TWL6030 and has low series resistance it should be okay. Unfortunately, it makes measuring the VBACKUP state at the ball of the TWL6030 more difficult, if not impossible depending on your layout.

    Let me know the results of the digital questions and maybe we can rule out the need for the analog concerns. 

  • Thanks for the reply Kevin

    Doing some further research, it seems the GUMSTIX module uses the TPS65950a PMIC and not the TWL6030 as referred to in the Kernel documentation.

    I have managed to get contact number for TI person in Australia and he is helping with getting the NDA for the full documentation.

    Interesting thing is the 65950a basic datasheet say the RTC backup works till voltage below 1.8V, so very strange that we get an RTC loss when battery is sitting above 2.5V. That's a very large dip.

    Measuring VBACKUP at the PMIC is not going to be possible as it is on multilayer PCB which then plugs into our PCB.
    Lots of caps on the GUMSTIX but not sure any of them are for the RTC backup.

    Regards
    Barry
  • Hi Barry,

    In order for the TPS65950 to maintain the backup domain, one should note that the main battery voltage has a slightly higher input voltage requirement than the backup battery voltage.

    The Main battery can maintain the backup domain down to 2.1V. The backup battery can maintain the backup domain down to 1.8V. So it could only operate down to 1.8V if the main supply was shorted to the backup supply or there is a backup source.

    If the VBACKUP pin is floating, this could also cause uncertainty during the decision event when the main battery drops below 2.7V.
  • Hello Richard

    Thanks for the feedback.

    When you say main battery, that would in the case of a Gumstix module be the main supply?

    In our case there is only 1 physical battery, this is connected to VBACKUP and that is a 5mA lithium battery, which charges from the PMIC.

    This was measured to charge to 3.1V then over a 3 day period discharged to 2.6V (Which from the graphs is normal discharge curve).

    The main battery as you call it would be our board supply, shorting this to VBACKUP would serve no purpose other than to remove any backup capability. 

    The main supply was turned off for the 3 day period so no supply voltage here, normally it would be 3.3V.

    It would be a strange VBACKUP if it also requires the main supply in order to provide RTC backup.

    When the main supply was turned back on the Linux kernel reported a power on reset from the RTC and set the clock to its default of 1/1/2000.

    I am waiting on the datasheets so I can determine if the Linux kernel driver is reading / checking the correct status bits to determine the POR.

    From what little information is available without NDA, I understand POR should only occur if both main battery (supply) and VBACKUP are below their minimum levels. Am I not understanding this correctly?