• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 33 subscribers
  • Views 419 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.

TCA9537: Reset not working

Karina Kring1
Prodigy 50 points
Part Number: TCA9537
Other Parts Discussed in Thread: STRIKE

Hi,

when we apply an ESD-pulse sequence to our chassis (air discharge 15kV), the TCA9537 will sometimes go to a state where the I2C is not responding and the device can not be initialized. However, the device is still correctly setting the interrupt pin INT when one of the Px pins is toggled. Pulling the RESET line low does not help in this cases- afterwards, the device I2C is still locked. Only toggling the supply voltage of the device will make the TCA communication work again.

Under "normal" conditions our reset pulse is working fine, it just happens on repeated ESD discharge.

Is this a known issue?

Best regards,

Karina

  • +1
    TI__Mastermind 49768 points

    Hi Karina, 

    We test our devices up to JESD 22 ESD standards. This includes human-body model and charged device model (CDM). 

    The P0-P3 pins and VCC are also HBM compliant up to +/- 4000V. 

    I would suspect that a 15kV air discharge might be outside the datasheet spec, so we wouldn't have any information on this type of test. I also don't know if the device can handle this type of transient. It might be possible that it could damage the device due to such high kV. External ESD protection might be required for this type of testing to work. 

    What it sounds like is happening is that the ESD strike is affecting the internal I2C control of the device locking up the SDA/SCL bus to a certain logic level. I believe the I/O port and /INT pin are somewhat separate to the I2C bus control. This is why the interrupt is still functional when a change on the p-port side is made. 

    In an attempt mitigate the effects of ESD, it might be wise o implement some type of external ESD to the I2C pins. A device such as the TPD2E2U06 might be suitable for this system. Series resistance and some filter capacitance on the I2C lines might also help with some of these transients, i.e. 33ohm series resistance + 100-200pF cap to GND (low-pass filter) on the I2C lines. 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 50 points

    Hi Tyler,

    thank you for the reply. 

    I hope the  IC itself does not see the 15kV but a lower pulse, since we apply the pulse to the chassis.

    Thank you for the idea with the filtering on I2C lines, we will try this.

    Best regards,

    Karina

  • 0 in reply to Karina Kring1
    TI__Mastermind 49768 points

    Hi Karina,

    Please let me know the results of the testing when you get to it. 

    Regards,

    Tyler

  • 0 in reply to Tyler Townsend
    Prodigy 50 points

    Hi Tyler,

    with the additional filtering, we still could get the device stuck, it only happened less often, so the performance was more robust. Might add this in the next revision.

    Best regards,

    Karina

  • 0 in reply to Karina Kring1
    TI__Mastermind 49768 points

    Hi Karina,

    I am glad the suggestion was beneficial. 

    Keep in mind that this device is only rated JESD 22 ESD standards, so conducting a test outside of those requirements is not guaranteed by the datasheet. This means that potential damage could be occurring to the device since the kV of the ESD is way higher than the ratings provided in the datasheet. 

    In terms of completely avoiding the stuck device issue, please considering adding more filtering or some type of external ESD chip like I mentioned above. 

    You might also find that I2C is not the best fit for the system. Differential protocols like CAN or RS-485 have more robust physical layers than I2C, and have better ESD TVS options for their designs since those types of transceivers deal on a daily basis higher amounts of ESD than I2C does. 

    Regards,

    Tyler