• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 368 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

PCA9306: Rising Edge Step

mark otto
Prodigy 10 points
Part Number: PCA9306
Other Parts Discussed in Thread: ISO1540

Tool/software:

Both the SDA and SCL signals have a 200nS rising edge step.  Struggling to understand where this behavior comes from, we've seen posts on this part discussing this type of behavior for falling edges but is there a similar known behavior for rising edges?  Below is a simplified schematic of the circuit as well as a scope shot showing the step.

  • 0
    TI__Mastermind 40526 points

    Hey Mark,

    This is a unique case- the PCA9306 is an open drain based translator, I wouldn't expect the device to cause this step on the rising edge. The schematic looks OK, but I do have several questions here: 

    What is VREF1 and VREF2 connected to? 

    Can we try disconnecting the load (in this case is it the ISO or processor) to see if step is caused by load?

    Where in the I2C data does this occur? Is it visible on the ACK bit or any other place in the waveform? 

    Have we tried using stronger pullups to achieve faster rise time (and also what is the data rate)? 

    Regards,

    Jack

  • 0 in reply to Jack Guan
    Prodigy 10 points

    VREF1 is connected to 1.8V and VREF2 is connected to 3.3V through a 200Ohm.  Let me get you some additional scope plots this week.

  • 0 in reply to mark otto
    TI__Mastermind 40526 points

    Hey Matt,

    Thanks for the clarification, please keep me updated with any findings.

    Regards,

    Jack 

  • 0 in reply to Jack Guan
    Prodigy 10 points

    Hi Jack,

    After further investigation it has been noted that the SDA and SCL nets do not have this step when iso1540 is not present.  When the iso1540 is populated the isolator adds the flat spot as parts of its normal operation due to its latch prevention strategy.  Thanks for the debug tips.  A few white board screen shots have been included to help explain how we believe the iso1540 works.

      

  • 0 in reply to mark otto
    TI__Mastermind 40526 points

    Hey Mark,

    Thank you for this insight! I will share this with my team and for any future boards facing a similar issue with PCA9306 + ISO1540.

    Despite this step in the signal, is it safe to say that communication is not corrupted? Or did you have to resort to an alt. solution? 

    Regards,

    Jack

  • 0 in reply to Jack Guan
    Prodigy 10 points

    Jack,

    We noticed this during our regular DV as something to look into further, it is causing no known issue.  Now that we understand what is going on we are more comfortable with what we are seeing and plan to take no action.  We believe that a slightly stronger pull up would shrink the amount of time it spends with a flat spot but the duration is currently so small that we do not have plans to investigate further.