• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 79 subscribers
  • Views 735 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.

MSP430F1611Power Consumption LPM0

Paul Ehrlich
Prodigy 40 points
Other Parts Discussed in Thread: MSP430F1611

Hello,
I'm trying to run my MSP430F1611 in LPM0 but the power consumption is way too high compared to the data sheet. SLAS368F − OCTOBER 2002 − REVISED MAY 2009 Page 27
I'm using a frequency generator on XT2 to generate the 1Mhz SMCL and a 3.3V Power Supply. I also configured   BCSCTL2 |= SELM0; to get SELM=(0,1).
My measurement method seems to be right since it yields the right results for LPM2 and LPM3.
Thanks for any suggestions.
Cheers
eac

  • 0
    Guru 227245 points

    Looks like there are some components still active in LPM0/1 (e.g. sourced by SMCLK) which aren't anymore in LPM2 (when SMCLK is off). Especially if SMCLK is sourced by the DCO, LPM1 will not switch the DCO off.

  • 0 in reply to Jens-Michael Gross
    Prodigy 40 points

    Any hints what could be active. I don't use any component activly.Thanks eac

    #include <io.h>            //Definitionen der Ein und Ausgaberegister des Prozessors

    int main()                //main-Funktion
    {
     WDTCTL = WDTPW + WDTHOLD;   

     P2DIR |= 0xFF;
     P2OUT |= 0x08;
     P5DIR |= 0x70;// P5.5,6 output
     P5SEL |= 0x70;// P5.5,6 SMCLK, ACLK and MCLK output


      BCSCTL1 |= XT2OFF+RSEL2+RSEL1; /* ACLK is devided by 4; RSEL=6; no division for MCLK */

      BCSCTL2 |= SELM0;
      BCSCTL2 |= SELS;

      asm(
         " MOV.W   #5,R13\n"                                   // a
         " MOV.W   #5,R14\n"                                   // a
         " MOV.W   #10,R15\n"                                    // c
      );
      //__bis_SR_register(LPM0_bits);
    }

  • 0 in reply to Paul Ehrlich
    Genius 4925 points

    Hmm maybe it's simply because you are outputting 1 MHz on a port pin which will lead to an increased current consumption as with every switching there will be crosscurrents flowing through the output buffer stage of the pin.

    Or wait no, your outputting something quite larger than 1MHz if you RSEL=6 you will be in the order of 2MHz so the current will be even higher then - and not comparable to those in the datasheet as they are for 1MHz.

    In LPM2 or 3 you won't notice that as the DCO is turned off then.

    So maybe you should try and use the calibrated values for 1MHz (are they available on 1xx devices? I've never used one) and try again - if the current is still too high - turn off the pin output.

    And why is ACLK divided by 4, you don't set any DIVA bit, and the default value of BCSCTL1 is 0x84, which is XT2OFF, XT1 LF mode, ACLK / 1, and RSEL = 4.

     

  • 0 in reply to Bernhard Weller
    Prodigy 40 points

    Thanks for your suggestions.
    The output of MCLK and SMCLK isn't the problem, but it is nice to have on to monitor the CLKs.
    The Problem isn't either the RSEL since I didn't use the DCO for any CLK.
    The Problem gets fixed when setting SELS back to 0! This means that the MSP430 consumes 3 times more power when using a frequency generator at XT2 instead of the internal DCO to source the SMCLK in LPM0.
    Can anyone confirm my observation?
    eac

  • 0 in reply to Paul Ehrlich
    Guru 227245 points

    Power consumption ion the datasheet referes to CPU core (in increases significantly with speed).
    But driving a high speed crystal consumes even more. That's the price you pay for greatly increased accuracy.

  • 0 in reply to Jens-Michael Gross
    Guru 227245 points

    Jens-Michael Gross said:
    But driving a high speed crystal consumes even more.

    I just checked on a different MSPs datasheet. There the XT2 oscillator with 8MHz consumes about 300µA. But this is for a crystal. You wrote you're using a frequency generator, so it should be less, but then, the 1611 is a bit older and maybe consumes more. And the old BCS does not have a setting for 'bypass mode', so the oscillator circuit is still active, even if you don' tuse it at all.

  • 0 in reply to Jens-Michael Gross
    Prodigy 40 points

    Jens-Michael Gross said:

    . And the old BCS does not have a setting for 'bypass mode', so the oscillator circuit is still active, even if you don' tuse it at all.

    I did some measurements at 1,2 and 4 Mhz. The difference between using XT2 with an frequency generator as source compared to the DCO is 181µA, which isn't that dramatic but adds significant on top of the LPM0 value of the datasheet (75-90µA).
    cheers
    eac


  • 0 in reply to Paul Ehrlich
    Guru 227245 points

    Paul Ehrlich said:
    The difference between using XT2 with an frequency generator as source compared to the DCO is 181µA, which isn't that dramatic but adds significant on top of the LPM0 value of the datasheet (75-90µA).

    Definitely. It triples the power consumption.
    I don't think this is an issue anymore on the other families, as there then is a way to switch the crystal in bypass mode, so teh oscillator circuit is off. I actually expect a decrease in power consumption then when the DCO can be switched off totally.

**Attention** This is a public forum