• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 80 subscribers
  • Views 963 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.

MSP430G2755: Analog input pins protection

Tudor Patroi
Prodigy 50 points
Part Number: MSP430G2755
Other Parts Discussed in Thread: LM9076

Hi,

we are having an internal debate whether additional external protection needs to be added to the analog input pins when measuring 24Vdc, if the micro is powered from 3.3V and we probe through a 10:1 high impedance divider. More details:

-the device is entirely self contained, no external inputs except external 24Vdc from a SMPS brick adapter. All data exchange is wireless.

-this 24V input voltage goes into a linear regulator with LM9076BMA which powers the micro with 3.3V

-the 24V input voltage is also sampled using a 10:1 series divider 44.2k with 4.99k to ground. Another option is to sample through 162k series and 18k to ground.

-Both options give a 10:1 ratio, so the output voltage is 2.4V. Equivalent circuit is a 2.4V supply with an internal resistance of 4.5k in the first case (44k2||4.99) and 16.2k in the second case (162k||18k).

-Datasheet specifies at page 19 Absolute maximum ratings "Voltage applied to any pin (2) –0.3 V to V CC + 0.3 V" and "Diode current at any device pin ±2 mA".

Under these circumstances, do we need to add clamping diodes to Vcc and GND at the analog inputs? What is the worst scenario if we do not have them? Looking forward for your opinions, thank you for looking into this.

  • 0
    Genius 3255 points
    It all depends on the application. If this is a mission critical part, then you should add protection.
  • 0
    TI__Expert 3520 points

    Hi Tudor,

    Based upon your system description I personally do not see an explicit requirement for external protection.

    It will be safer to have it, but if all voltages you describe are stable, DC voltages (unlikely in real world applications, but your system does sound quite contained) then no current should be flowing trough the GPIO's ESD structure and the +/-2mA diode current limit will not be exceeded. Given the voltage divider ratios no voltages will exceed the absolute maximum ratings.

    Choosing the second case resistor divider will result in a higher minimum tsample for the ADC measurement due to the increase in source resistance and will depend on your system requirements if this slower sampling rate is acceptable (see section 22.2.5.1 of the User's Guide for more details):

    If you exceed the diode current limits of the pin as a result of a system-level malfunction, then it can result in an increase in the supply voltage of the MCU as a result of the diode feeding current on to the VCC rail. This may result in the maximum supply voltage specification to be exceeded and damage done to the device leading to functional and parametric failures over time.

    A somewhat similar question was asked before related to an E-Meter protection circuit which had a response with a few nice resources you may review here.

    Best Regards,

    Mark-

  • 0 in reply to Mark Swanson
    Prodigy 50 points
    Thank you Mark and Mike for your insight. Since the topic of being able to use the internal clamps or not might interest other people too, I will leave this open a bit more.

    This is about an RFID reader that does statistics about the tags passing by. Nothing related to life or safety. So from this point of view is not "mission critical" as long as the micro continues to operate normally .

    Sample time is not important as we just measure some dc voltages after power on for the board self-test. If any of the 6 values is out of range then the self test fails, we record an error code and keep the "ok" led off. Later can issue a command to repeat the test.

    So at this time we would like to save the cost of the 10 clamping diodes, but there is a concern is that during power on the 2.4V input voltage is applied for maybe 50ms while the micro is kept in reset - due to the startup time delay of the 3.3V power supply. Sorry I should have mentioned that the LM9076 3.3V supply also generates the reset signal for the micro from its /RESET pin, so the micro comes out of reset only after its supply is OK and higher than the 2.4V input.

    Question is if applying external currents on the internal clamping diodes during this short time might perturb internal chip operation later after it powers on? Does this look like a possible failure scenario to you?
  • 0 in reply to Tudor Patroi
    TI__Mastermind 49692 points

    Our ESD Diode Current Specification application report should answer your questions.

    Regards,

    James

    MSP Customer Applications

  • 0 in reply to Tudor Patroi
    Genius 4350 points

    I guess I don't understand the concern.  It seems either of your divider options would keep diode current below 2 ma even if Vcc was grounded when the top of the divider was at 24V.  Also, it doesn't matter whether the processor is in reset.  It only matters what voltage is at Vcc - that's what the internal protection diodes are connected to.  And I don't know what you accomplish by adding external diodes.  Unless they are Schottky, you can't even be sure any current would ever flow through them.  As long as the divider keeps the current under 2 ma in the worst case, I think you're good.  I think you're also ok without the external delayed reset.  The MSP430s have all that circuitry built in.

  • 0 in reply to James Evans
    Prodigy 50 points

    Thank you James and everyone else, now the matter is clear. External Schottky diodes for additional protection are not required in our case as we use high value resistors to limit any current through the ESD diodes to ~0.5mA for ~50ms in the worst case. Tudor

**Attention** This is a public forum