• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 102 subscribers
  • Views 380 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.

Original question:

AM335x GPIO input slow rise/fall times

AM3359: Input signal with long transition time

Daisuke Maeda
Guru 19835 points
Part Number: AM3359

Does a slow transition time signal to an input buffer with hysteresis cause permanent damage to the device? Or does it only affect device reliability?

On our customer's board, GPIO signal is input in a transition time of 40 msec.

A slow transition time signal should not cause permanent damage to the device because an input buffer with hysteresis does not generate a large through current in a short state.

Best regards,

Daisuke

  • 0
    TI__Guru**** 393215 points

    Daisuke Maeda said:
    A slow transition time signal should not cause permanent damage to the device because an input buffer with hysteresis does not generate a large through current in a short state.

    This is not true. If you check the AM335x Datasheet Rev. K you will see that input signal maximum rise/fall times for most peripherals are required to be in the range of 5-10ns. I2C is a notable exception, with 300ns. As GPIO's use the same I/O cells I would suggest that you provide external buffering for these slow signals.

  • 0 in reply to Biser Gatchev-XID
    Guru 19835 points

    Hi Biser-san,

    Thank you for your reply.

    Can a slow transition time signal cause permanent damage to the device?

    The input signal maximum rise/fall times for GPIO are undefined. Is it recommended that they are 300 nsec as well as I2C?

    Best regards,

    Daisuke

  • 0 in reply to Biser Gatchev-XID
    Guru 19835 points

    Hi Biser-san,

    Does the input buffer with hysteresis mean the schmitt trigger buffer?

    The schmidt trigger should not cause permanent damage because does not generate a large through current in a short state. Is my understanding incorrect?

    Best regards,

    Daisuke

  • 0 in reply to Biser Gatchev-XID
    Guru 19835 points

    Hi Biser-san,

    For GPIO, is it recommended that the input signal maximum rise/fall times are in the range of 5-10ns?

    If the input signal rise/fall times are exceed 10ns, what are acceptable times?

    Best regards,

    Daisuke

  • 0 in reply to Biser Gatchev-XID
    Guru 19835 points

    Hi Biser-san,

    Daisuke Maeda said:

    For GPIO, is it recommended that the input signal maximum rise/fall times are in the range of 5-10ns?

    If the input signal rise/fall times are exceed 10ns, what are acceptable times?

    Our customer asks this for this case as well as for other cases.

    Best regards,

    Daisuke

  • 0 in reply to Daisuke Maeda
    TI__Genius 15180 points

    Daisuke-san,

    We have not seen any problems where this transition time would cause permanent damage to the AM335x device.  However, as mentioned in the original e2e post, the best design practice for new designs is to buffer the signal with a Schmidt trigger device before passing it to the processor input. 

    For old designs, we’re not aware of any reliability concerns, but if your customer sees any issues you believe could be related to this transition time, you can submit those devices through our customer return portal and we can evaluate them for you.

    Regards,

    Melissa

  • 0 in reply to melissaw
    Mastermind 6365 points

    HI, Melissa-san

    I had the same question, so let me ask you a question.

    Sorry for repeating similar questions.

    If you check the GPMC_CSn1 (V14/U9) pin specifications in Table4-2 Pin Attributes in the data sheet, you will find the HYS function.
    Since there is a HYS function, there is no problem even if a 4ms rise / fall signal is input.

    or,

    Since this pin has HYS function, it is not necessary to add an external Schmitt buffer device.
    Is my idea wrong?

    best regards
    cafain