• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 80 subscribers
  • Views 1491 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.

P430F56xx GPIO input voltage tolerance?

Edward Averill
Genius 3100 points
Other Parts Discussed in Thread: MSP430F5659, MSP430F5438A

Erm.. I think I might have messed up my breadboarding and put 5V into one of my MSP430s GPIO inputs (color my face red).

What kind of damage will this do?  I see 0's on that input pin even when I give it a feed straight from the 3.3V power supply...

If I did toast it I have other chips, thankfully.

All info appreciated!

Ed Averill

  • 0
    TI__Guru 68116 points

    Hi Ed,

    I think it's safe to say we have all done something similar at one point or another. :-)

    The absolute maximum rating for the voltage on any MSP430F5659 pin (from the datasheet) is Vcc + 0.3V, and the maximum for Vcc is 4.1V. The absolute maximum rating is the point at which, not only is the part not guaranteed to run correctly while exposed to this out-of-spec condition, but it also could permanently damage the device - so you could have problems later even after you go back to within-spec voltages.(As a note, Recommended Operating conditions specs are a little different - with those, your part may not run correctly, but you will not permanently damage the part unless you hit the abs. max ratings).

    The damage behavior can be unpredictable but it sounds like from your part's behavior you have at least damaged the I/O. To be safe, it would be best to swap out for another chip - otherwise if you are trying to debug other issues down the line you won't be able to be sure if it's something else in your design causing your problem, or if it's just something behaving badly because the part was damaged/stressed in the past.

    Best of luck,

    Katie 

  • 0 in reply to Katie Pier
    Genius 3100 points

    Yep.. looks like I baked it.  The voltage divider circuit my HW guy set up let too much through, I saw a peak of bout 4.9 which appears to have killed it.

    I ordered some 3.3V regulators, that should prevent this from happening again.

    Net time.. test the outputs first!

    And thanks for the response - have a great weekend!

    Ed Averill

  • 0
    Guru 227245 points

    All MSP digital I/O pins have clamp diodes to VCC and VSS.
    If the input voltage exceeds VCC (plus the diode voltage), current begins to flow from the pin to VCC (or, in case of negative voltages, to GND). This will put a load on the signal source, which may (depending on the signals output impedance) cause a voltage drop and limit the input voltage. This effect can be forced by a series resistor.
    Since current flows into VCC, this will also rise VCC. So you can effectively power an MSP through its I/O pins.
    However, the diodes are limited to 2mA rated current. (this will also prevent VCC from being stuffed with too much current and rising above a critical voltage that would destroy the MSP).
    If the current is exceeded, the diode will melt. This may cause a permanent short of this pin to VCC or GND. Or the diode will simply break and the protection is lost, in which case the port pin input logic may take damage due to the overvoltage.

    If you need to receive signals from a 5V source, you have three choices:
    1) put a voltage divider on the input that lowers the voltage form 5V to 3V. This is the most obvious but least recommendable solution.
    2) use a series resistor. As soon as current begins to flow through the clamp diodes, voltage will drop on the series resistor. A 1k resistor will drop 2V on 2mA, so only 3 of 5V will arrive at the port pin. However, the port pin only needs <100nA for safely detecting a high value, so you can (in theory) use a series resistor as large as 20M (well, that much isn’t recommended due to pin capacitance and the resulting time constant). This is the simplest way    .

    3) You can use a pull-up resistor and a diode. The diode will turn the source into an open-collector output (only pulling low but not pushing high) and the pull-up resistor (you may use the internal one) pulls the pin up constantly. This is the “right” solution. It is, however, more expensive. And due to the clamp diodes, it is not really necessary.

  • 0 in reply to Jens-Michael Gross
    Expert 1240 points

    Has anyone tried the large resistor in series hack?
    In another thread, Jean recommends a 47K which everyone has floating about in dusty corners.
    Although I just purchased some proper 5V to 3.3V convertors, I would have tried this just prove the concept I was working on.
    Specifically, this for the MSP430F5438A installed in the TrxEB board which I DO NOT want to blow up before deadline.
    (He mentions even as much as a 20M would do, so maybe a very common 1M would be a good rule of thumb?)
    I've got these coming Sunday which are $2...
    www.amazon.com/.../ref=oh_aui_detailpage_o00_s00

  • 0 in reply to Zebryk
    Guru 227245 points
    Did I recommend 47k? Maybe. And likely for the reason you mentioned: everyone has one. The required value depends on the external voltage.
    When subtracting VCC from the maximum signal voltage and dividing the result by the resistor value, the resulting current must be smaller than 2mA and should be greater than 100nA. Anything that fits between is okay. However, smaller currents mean a larger delay on signal change detection due to input capacitances. Larger currents increase (obviously) current consumption
    Anyhting between 500R and 10M for each Volt of input overvoltage is within the allowed limits. In various projects, we have used 10k or 47k to interface 5V parts or even 12V input signals. We even had an applicaiton where the part was directly operating on AC mains, detecting the zero-crossing. But of course in this case, we didn't use a tiny SMD resistor, as it has to block up to 400V. IIRC, it was 4.7M 1W through-hole we used for this.

    The only additional thing to be considered is that the current that flows through the resistor and the clamp diodes is 'injected' into VCC. So if this current is higher than the circuit's total current draw on VCC, this will cause VCC to rise.
  • 0 in reply to Jens-Michael Gross
    Expert 1240 points

    Jens,

    Thought you retired or became a poet:-)

    Rather like 47K as it is somewhat memorable.

    Casting about now looking for something I can afford to blow-up.

    Thanks,

    Jay Zebryk

  • 0 in reply to Zebryk
    Guru 227245 points
    "Thought you retired or became a poet:-)"
    Both :)
    Well, I've already been a poet since school times. With only two or three things worth remembering.
    And - believe it or not - I continued work on my MSP book. Which isn't really poetry, but close...

    But for threads where I once wrote something, I will still get notifications on new posts. And if I have somethign to add, I'll sometimes do.

**Attention** This is a public forum