• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 79 subscribers
  • Views 1932 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.

MSP430F5438A: Power consuming too much on the chip

DONGHYEON ROH
Prodigy 180 points
Part Number: MSP430F5438A

Hello, I developed a new MSP430-cored application device, and noticed there was an critical issue on power consumption.

; MSP430 chip consumes too much power than estimation.

============================================================================================

=  Condition                                                                                                                                                                            =

============================================================================================

Chip name: MSP430F5438a

Package type: BGA MICROSTAR JUNIOR

Mode: Active mode

System frequency: 25 MHz

External crystal frequency: 32.768 kHz

Supply voltage : 3.3 V

============================================================================================

=  Estimation of power consumption                                                                                                                                      =

============================================================================================

[1st way] Based on Features description (datasheet)

(Flash power consumption + RAM power consumption) = (230uA/MHz + 110uA/MHz) x 25MHz x 3.3V = 28.05 mW

[2nd way] Based on Maximum power consumption (datasheet)

(Flash Max. power consumption + RAM Max. power consumption) = (9,6mA + 4.9mA) x 3.3V = 47.85 mW

[3rd way] Based on Comparative table (online table)

Active mode power = 356uA/MHz x 25MHz x 3.3V = 29.37 mW

We chose the second calcululation result as estimation value because it has the highest value.

============================================================================================

=  Measurement practical power consumption                                                                                                                      =

============================================================================================

[1] MSP430 chip power consumption while operating in the application device : 103 mA x 3.3 V = 340 mW

[2] MSP430 chip power consumption only connected to 3.3V voltage regulator output pin and GND : 142 mA x 3.3 V = 468.6 mW

============================================================================================

=  Conclusion                                                                                                                                                                         =

============================================================================================

As a result, I estimated less than 50 mW power consumption,

but I measured more than 300 mW power consumption.

Why the MSP430F5438A chip highly consumes power?

Is this normal? I cannot understand why.

  • 0
    TI__Guru**** 202236 points
    Hello Donghyeon,

    These measurements are higher than expected, have you initialized all GPIOs as outputs driven low? Please follow the Table 1-3, Connections of Unused Pins, suggestions found in the User's Guide. I assume a custom EVM is being used and not a TI-provided one? Where is the current being measured? Other devices (LDOs) or tools (FETs) can be contributing to the elevated power numbers.

    Regards,
    Ryan
  • 0 in reply to Ryan Brown1
    Prodigy 180 points
    Hello Ryan,

    I tested again and found one connection error,
    Is the XT2IN pin important?

    Test contents
    =====================================================================

    1. Pin connection
    I followed all connections from User's Guide, except for XT2IN.
    In the table 1-3, XT2IN should be connected to VSS, but I didn't connect into any pins (Open).
    Is this XT2IN pin connection important?

    2. Test code
    On the test code,
    first of all, MCU initialized all the port,
    - setting : OUT=0, DIR=ON, SEL=0
    - target : PA, PB, PC, PD, PE, P11, PJ
    and ran a infinite loop (code: while(1);)

    , but it still consumed more than estimated value.


    3. TEST again
    - Connections for current measurement( (1) + (2) + (3)) :
    (1) 3.7V battery - CURRENT METER(Between plus pins - direct connection, not parellel) - LDO regulator - MCU
    (2) MCU - Bluetooth module (PAN1326B chip, TI, PANASONIC)
    (3) MCU - ADC module (RHD2132, Intan tech.) <- Just power on

    - LDO regulator : TPS73733 (Texas Instrument)
    -> We calculated and measured LDO's contribution;
    estimation : It will increase 15% total power.
    measurement : It increased less than 19% power in total. (in practical use) -(a)

    - Bluetooth chip power consumption (measured) : < 43.7 mW -(b)

    - ADC chip power consumption (measured) : = 31.9 mW -(c)

    - Total Power consumption (measured) : = 395.2 mW -(d)

    - Other modules (such as resistors, and capacitors) consumes almost zero power (steady state)

    4. Analysis
    From (a), (b), (c), and (d),
    MCU power consumption is
    P >= (Total power / LDO max. contribution ) - Bluetooth power - ADC chip power
    = 395.2 / 1.19 - 43.7 - 31.9
    = 256.5 mW*

    Because the supply voltage on MCU is 3.3 V,
    current on the MCU is more than 256.5/3.3 = 77.7 mA**

    ====================================================================================
    Re-test result
    1. One Pin connection error - XT2IN- OPEN (User guide's suggestion : GND)
    2. Measured again, but it still consumes much more power than estimation
    - Estimation : 47.85 mW
    - Measurement : 256.5 mW
    ====================================================================================
  • 0 in reply to DONGHYEON ROH
    TI__Guru**** 202236 points
    Hi Donghyeon,

    Do you have a TI-built EVM to test on, like the MSP-TS430PZ5X100 or MSP-EXP430F5438, and if so then do these yield similar results? Have you also tested with LPMs, different MSP430 CPU frequencies, and core voltages? Please provide the test code you use to evaluate the MSP430 current consumption.

    Regards,
    Ryan
  • 0
    Guru 46710 points

    DONGHYEON ROH said:
    [2] MSP430 chip power consumption only connected to 3.3V voltage regulator output pin and GND : 142 mA x 3.3 V = 468.6 mW

    It's nearly half a watt consumption for chip which is running just at 25MHz. I would immediately test against output short condition, either direct or indirect. - After init of clocks you shall disable interrupts, configure all pins of all ports as inputs and put CPU into while(1){} loop, measure current again. Expected result would be consumption as per specs. After that you just have to narrow your search by enabling functions of your firmware step by step until you find which part causes problem, then supposedly it is easy to find what's wrong

  • 0 in reply to Ilmars
    Prodigy 180 points
    Hello, Ryan,

    Today, I made regulator module and test on the Bluetooth Eval board and ADC Eval board.
    There was no specific result. All the current values were almost same as last measurements.
    What I only didn't test was the MSP430 MCU chip power consumption.

    =============================================================================
    This result was weird because I almost tested chips except for MCU.

    Application device circuit = Bluetooth EVM + ADC EVM + 3.3V regulator + MCU + 32.768 Crystal + capacitors, resistors, reset switch.
    So what I didn't test were as follow;
    1.MCU
    2.crystal
    3.Capacitors, resistors, reset switch.
    Capacitors were in steady state, so there was no current.
    one resistor was open (used for reset switch), and others were zero ohm resistors so there was no voltage difference.
    ==============================================================================

    On the next time, I will test the MSP430 chip. If finish, I will upload the feedback result.
    Maybe the BGA chip what I used in the application device is not a good choice for test power consumption.
    So next time, referring your suggestion, I will use MSP-EXP430F5438A and connect GND and VCC pins with regulator module pins, which I made today.

    ########################################################################################################
    *I have also considered possibilities on mis-connection error and mis-placement error.
    But I didn't find any error from now on.(already checked several times...)
    If find, I will tell you again.
    ##########################################################################################################
    Hello, Ilmars,
    Okay I will test!
    In my last test code, I set all the unused pins as 0 logic output pins.
    So from your suggestion, I will set as input pins, and also made the chip disable to use Interrupt.
  • 0 in reply to Ilmars
    Prodigy 180 points
    Hello, Ryan and Ilmars.

    I found why the MCU consumes large power.
    The reason was from mis-connection of Vcore with V3.3 (DVCC)
    Here is my experiment results
    ==========================================================
    Device : EXP-MSP430F5438A, and custom-made V3.3 regulator module (TPS73733)
    Test results
    1. Regulator ON
    - connection : 3.7 V battery - regulator
    - battery current : 0.40 mA
    2. EXP-MSP430F5438A also on
    - connection : battery - regulator - EXP430F5438A
    - current : 17 ~ 24.5 mA
    3. Vcore pin connection into V3.3
    - connection : '2. experiment's connection' + Vcore-V3.3(DVCC) connection
    - current : 95 ~ 100 mA
    =========================================================
    =Conclusion =
    =========================================================
    If Vcore is connected to V3.3 voltage source,
    the MCU can be downloaded code, and operate well.
    Maybe, this is because the MCU is in Active mode, and the 3.3 V is luckily in the valid range.
    However, the mis-connection causes high power consumption.
    .
    Thank you guys. I appreciate your help!

**Attention** This is a public forum