• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 79 subscribers
  • Views 1738 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

Is RTC precision hampered by using different VCC and AUXVCC (MSP430F6723)

Anup Rajput
Expert 1920 points

Hi.

I'm being forced by a client to make RTC backup voltage (AUXVCC3) on MSP430F6732 same as VCC (3.3V). I may have to end up adding an expensive ultracap just to make the voltage same!

He thinks that if I use a CR2032 battery (3V), the oscillator will lose its precision while switching from 3.3V to 3V on power failure! I think the crystal runs on 1V generated internally and would continue to provide precise ticks irrespective of the change in input voltage.

And, of course, the precision in timing would remain the same irrespective of whether the source is 3.3V or 3V!

Am I correct in this understanding or is he right?

Thanks,

Anup

  • 0
    TI__Expert 7530 points

    Hi Anup,

    I recalled my earlier post to obtain further clarification on the exact device you are using.  My original thoughts were in fact correct.

    For chips that have the AUX module(Ch. 4 of the User’s Guide), the backup subsystem is only powered from AUXVCC3.  This backup subsystem contains the real-time clock (RTC) module with a 32-kHz crystal oscillator, backup RAM, and optionally (device-specific) up to two digital I/O pins.  There is no switching enabled with AUXVCC3 internally.  If you want to switch between powering the backup domain with AUXVCC3 and a battery, we recommend using a diode and the internal charger circuit (as shown below).  The only on-chip interface between the backup supply and DVCC is the resistive charging circuit.

    Can you describe in what way you are switching AUXVCC3 between 3.3V and the 3V battery?

    Walter

  • 0 in reply to Walter Schnoor
    Expert 1920 points
    Hi Walter, Thanks a lot for the detailed answer. The problem that the client envisages is that the crystal (32.768kHz on XIN/XOUT) will normally work on DVCC (3.3V) but as soon as DVCC is removed, will switch to AUXVCC3 in order to maintain RTC count. Is this understanding correct? Would this switching from DVCC to AUXVCC3 domain for XIN/XOUT cause discrepancy in RTC precision? Regards, Anup
  • 0 in reply to Anup Rajput
    TI__Expert 7530 points

    Hi Anup,

    The low frequency oscillator on XIN/XOUT is never powered by DVCC, nor is there any switching mechanism.  The oscillator, as well as the real time clock (RTC) and some backup RAM are always powered from AUXVCC3. 

    There is the capability to enable the charging circuit to provide current to AUXVCC3 from DVCC through a charging resistance (5-20kOhm).  Are you using this mechanism?  Can you provide a diagram clarifying how you are supplying AUXVCC3?

    I understand your concern is the oscillator precision, but to look at that problem we need to understand the power system.  AUXVCC3 should not be changing voltage unless you are tying it to DVCC externally through some external control circuit.

    Regards,
    Walter

  • 0 in reply to Anup Rajput
    Guru 227245 points

    The crystal always operates from AUXVCC3. If AUXVCC3 is not powered, the crystal won’t run even if VCC is present. As Walter said, no switching is done at all.

  • 0 in reply to Walter Schnoor
    Expert 1920 points

    Hi Walter,

    I'm sorry but I'm away from my desk and cannot post the circuit diagram.

    But, my DVCC comes straight from an LDO (3.3V), while my AUXVCC3 comes from a coin cell (CR2032 3V). Because the cell is unrechargeable, I have connected a diode "from" the cell positive towards my AUXVCC3 pin in order to ensure that charge transfer is only one way. There is nothing else on the AUXVCC3 path.

    There is absolutely no outside interaction between DVCC and AUXVCC3 and since you have clarified that there is only a charging resistor inside the MSP430, I think I'm good to go.

    Regards,

    Anup

  • 0 in reply to Anup Rajput
    Guru 227245 points

    Anup Rajput said:
    I have connected a diode "from" the cell positive towards my AUXVCC3 pin in order to ensure that charge transfer is only one way.

    Be sure that the leakage current of the diode is low enough. Ideal diodes are rare.

    I once used the leakage current of a Germanium diode as charging resistor for a Zinc-Air coin cell (which is partly rechargeable, like most batteries are). Low voltage drop on discharge and small charging current if main power is on.
    Kept my RTCs for the C64 running for many years without a battery change.

**Attention** This is a public forum