• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 79 subscribers
  • Views 1994 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.

MSP430F5xxx Damaged I/O Driver - What Could Cause Shorted I/O Failure Mode?

Charles Gervasi
Expert 2270 points
Other Parts Discussed in Thread: MSP430F5506

I have an MSP430F5506 with a single I/O pin shorted to ground.  The short is in the part, not my board.  It persists when the parts is removed and ohmed out unpowered.  The part works apart from this pin being shorted.

On the board, this pin is connected only to a transistor (FDV303N) gate.  The transistor's source and drain are connected to a load well within the transistor's parameters.  The trace does not pass near other traces. 

I can't imagine anything in software could damage the I/O driver.  So something goin on on that pin is causing my problem. 

Question: What bad things could be happening to an I/O pin that would cause it to fail shorted to ground?  Does it sounds like ESD, overvoltage, excessive current sourcing, excessive sinking, etc?

Thanks for any help!

  • 0
    Guru 12160 points

    Hi Charles,

    ESD can cause this failure, but in your case I would be "shocked".  (Sorry.)  You were configuring that pin as an output (I assume), which makes it nearly impervious to ESD.  Plus the trace doesn't even go off board, which makes it doubly impervious to ESD when combined with being an output.

    I have never seen excessive current (pin as output) cause this failure.

    I can imagine burning out the substrate diode (aka protection diode) on the Vss side could cause this failure if the diode shorted.  But again, I can't see how your circuit could have caused that kind of driven "undervoltage" on the pin (driven @ less than Vss - 0V3).

    Maybe you just have an unlucky part?  Has the failure occurred more than once?

    Jeff

  • 0 in reply to Jeff Tenney
    Expert 2270 points

    We are going to add a series resistor between the I/O pin and the transistor gate, hoping that will protect the pin.  I hope we find out why it's happening.  It has happened to several boards that the software group is using for development. 

  • 0 in reply to Charles Gervasi
    Guru 227245 points

    A shortcut to GND or VCC on an I/O pin is indeed usually caused by an overcurrent throught the clam diodes that protect the pin from over- or undervoltage. These diodes can sustain a rated current of up to 2mA, but when stresse too much, they may melt, resulting in either a shortcut or a break (leaving the pin unprotected, most likely frying the port circuit).

    From your usage of 'surce' and 'drain', I assume you are driving a FET. usually, the gate of the FET should be isolated from source or drain. However, FETs are an easy prey for inverted voltages (this is why MOSFETs have a shunt diode). It is well possible that an undervoltage on source has made its way to the gate.

    Also, a often enjoyed cause for over/undervoltage problems is a dual power supply situation where the GNDs are not connected. Then a virtual GND level is generated through the circuitry or shielding, resulting in a completely unexpected situation according to relative voltages.

    However, a series resistor is alsways a good idea. I usually use 1k, which limites an over/undervoltage of 2V to the allowed 2mA. Since you're driving a FET where the gate voltage is important, not the base current, even higher values will be without negative effect for driving the transistor.

  • 0 in reply to Jens-Michael Gross
    Expert 2270 points

    You are right on your estimation of my problem: '5506's I/O pin is driving a FET's gate directly.  No negative voltages or > rated voltages are presented to the drain .  The source is connected to the same ground as the processor. 

    I don't see where any negative voltages are on the board at all.  All I can think of is something like the inductance of the load, which is connected to the FET, causing the voltage to spike and then couple through to the drain.  This is unlikely because there's not much inductance.  I'll post it if we solve the mystery. 

**Attention** This is a public forum