• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 688 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.

TCA9548A: TCA9548 and TCA6146 non-monotonic falling edge on I2C

Zhizhao Niu
Intellectual 2315 points
Part Number: TCA9548A

When we tested I2C interface of TCA9548 and TCA6146, I found non-monotonic falling edge on both SCL and SDA. Its duration is about several nano-second.

Since the duration is very small, does it have an influence on threshold judgement ? Pls see the picture below. For the I2C interface, we use both standard mode and fast mode. Thanks!

Zhizhao

  • 0
    TI__Mastermind 49400 points

    Hi Zhizhao,

    Funny that you bring this up. I was just talking about this to a coworker about having an app note which discusses this in detail and to show customers that this kind of thing shouldn't cause alarm.

    What you're seeing is kind of like a by product of how a pass FET works. The TCA9548A uses pass FETs inbetween the secondary channel and the main channel. When a fast edge (usually the falling unless there is a rise time accelerator on the bus) drive the line on a pass FET, the signal driving will hang at the Vgate-Vth of the pass FET until the opposite side trails to the same voltage point. After which, the signal will then start to mimic the other side with some prop delay.

    The reason why I say this shouldn't be a point of alarm is because the I2C standard states I2C compliant devices need to incorporate a 50ns deglitch filter. I've personally never seen this byproduct cause a glitch larger than 10 ns so the deglitch filter should be able to take care of this blip.

    Thanks,

    -Bobby

  • 0 in reply to BOBBY
    TI__Intellectual 2315 points

    Hi Bobby,

    Thanks! From my understanding, this glitch is the inherent. I have looked through the datasheet and didn't found the pass FET description. Could you please help me to describe more about this part? Thanks!

    B.R.

    Zhizhao

  • 0 in reply to Zhizhao Niu
    TI__Mastermind 49400 points

    Hi ZhiZhao,

    You can see the pass FETs in the functional block diagram of 8.2 in page 13 of the datasheet. 

    -Bobby

  • 0
    TI__Intellectual 2315 points

    Hi Bobby,

    Thanks! I think you mentioned the input filter, main FET and the pass FET is shown as below, right? And the input filter here can also deal with <50ns glitch from the SCL and SDA line. right?

    BR

    Zhizhao

  • 0 in reply to Zhizhao Niu
    TI__Mastermind 49400 points

    Correct. The TCA9548A has an input filter on the SDA/SCL lines but this is for itself (when you communicate directly to the device), each I2C device on the bus that can be talked to should also include an input filter to filter out the glitch. (Once you enable a channel on the device, the filter on TCA9548A will NOT filter this glitch, the downstream I2C device should already have a filter built inside them)

    The main FET on SDA that you circled is for ACK'ing or driving the SDA line when you do a read from the TCA9548A device. It doesn't have anything to do with the pass FETs or the input filter.

    -Bobby

  • 0 in reply to BOBBY
    TI__Intellectual 2315 points

    thanks Bobby!