• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 36 subscribers
  • Views 125 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.

TCA4307: TCA4307 I2C hot-swap - Problem with devices getting damaged at hot-plug

Yannick Fortin
Prodigy 10 points

Part Number: TCA4307

Hi,

I'm facing an issue where hot-plugging a board into a live backplane causes damage to one or more TCA4307 devices located on the daugther cards.

Our design is a backplane with multiple daughter cards that are all hot-swappable. Each daughter card includes a TCA4307 device to protect their SCL/SDA lines that connect to the global I2C interface on the backplane. The daugther cards have staggered connector with ground-first contact.

The backplane also includes a separate 10-pin header that is used to connect a small control module. The connector for that module is a simple wire-to-board header, without staggered pins. The PCB of the module is at the other end of a 100mm AWG22 wire harness. That control module also includes a TCA4307 device and also connects to the backplane's global I2C interface.

When I do a hot insertion of the control module into the live backplane, it will frequently damage one or more of the TCA4307 devices located on the module itself and/or on the daughter card. A damaged device will hold the SDA and/or SCL line permanently low or have an intermediate voltage, resulting in a pernanently stuck interface for the entire system. We need to replace the broken TCA4307 devices to make the I2C interface functional again. No other devices are getting damaged in the system.

I was suspecting that the lack of staggered connection on the module could be the source of the issue, therefore I have tried attaching the ground of the control board to the ground of the backplane with a separate wire prior to system power-up, to ensure that the control board and the backplane were at the same potential throughout the experiment. Yet, the devices still got damaged after a hot insertion of the module into the backplane.

The TCA4307 devices all have their VCC pin powered from a unique 3.0V rail from the backplane. The card-sides are powered at 3.3V from on-board regulators. The shared SDAIN/SCLIN lines have 10K pull-up resistors to 3.0V on the backplane, and each daughter card/module has 10K pull-up resistors to local 3.3V on their SDAOUT/SCLOUT pins.

I have looked at the TI application notes and found some discussions about the possibility of stucking the bus during hot-plug, but did not see anything about a risk of damaging the devices. In my mind, the TCA4307 device is by itself the protection and should not die so easily.

Can you please advise on that issue? Can you think of a reason why the TCA4307 are dying and what could be done to protect them?

Note that although our backplane supports multiple daughter cards, at the moment I only have one daugter card and one control module in the system. We have limited supply of proto daugther cards, therefore I don't know yet if hot-plugging a second daugther causes similar problem.

Thanks,

Yannick.

  • 0
    TI__Guru 61061 points

    Hi Yannick,

    We are still looking into this and will provide a response shortly.

    Regards,

    Jack

  • 0 in reply to Jack Guan
    Prodigy 10 points

    Thank you Jack. I am attaching a simplified connection diagram for your reference.

    ComtechBackplane_Rev00.pdf

  • 0 in reply to Yannick Fortin
    TI__Guru 64948 points

    Hi Yannick,

    We will need more time to respond to you. Let us get back to you with some answer tomorrow eod. 

    Regards,

    Tyler

  • 0
    TI__Prodigy 360 points

    Hello Yannick,

    Can you confirm that the TCA4307 is operating in the recommended conditions shown below (Datasheet, pg. 4). Additionally, is it possible to provide oscilloscope captures of VCC, SCLIN, and SDAIN pins of the TCA4307 during the hot swap?

    Regards,

    Jason

  • 0 in reply to Jason Trinh
    Prodigy 10 points

    Hi Jason,

      I confirm that the TCA4307 is operating within recommended operating conditions under normal operation, but the hot-plug event can cause some signals to go above specification. You will see on the attached waveforms that the SCL and SDA lines can momentarily rise in the 7-10V range when hot plug occurs. Note that the SCL and SDA lines were measured on the backplane, not on the module. But as I indicated earlier, the TCA4307 devices located on daughter cards are dying as well.

      I am sharing two waveforms. Please note that the shape of the waveforms will vary from one time to the other. The ground wire was connected between  the module and backplane at all time when these waveforms were taken, this was done to compensate for the fact that the connector is not staggered. Also, note that for these specific measurement, the module's TCA4307 was powered from 3.3V generated on the module itself, not from 3.0V provided by backplane (different from schematic shared previously). Our PCB is actually designed to use the local 3.3V rail by default, the 3.0V VCC connection on the module was just a test that we did to comply with the indication from datasheet. The results were the same in all cases.

      Since the TCA4307 is described as a "hot-swap buffer", I would assume it should be able to handle the short surges associated with hot-plugging. I did not see indications in the datasheet about the need to use additional protection for transients. Maybe I missed something?

      Thanks,

      Yannick.

  • 0 in reply to Yannick Fortin
    TI__Prodigy 360 points

    Hi Yannick,

    Operating outside of the absolute maximum ratings (refer 5.1 of datasheet) may induce damage to the device. The TCA4307 does not provide any transient voltage protection, so a system level implementation to operate within the recommended conditions would be advised.

    Regards,

    Jason 

  • 0 in reply to Jason Trinh
    Prodigy 10 points

    Hi Jason.

    Could you please advise on the best way to protect the TCA4307 device from transients?

    I already did a test where I have installed 100 ohm resistors on the module, on SCLIN and SDALIN signals, in series between the connector pin and the TCA4307 pin. That seemed to have helped a lot, since it did not fail after 60 hot plug insertions. I was thinking of adding 3.3V TVS protection diodes as well. Would you have any other recommendation?

    Thanks,

    Yannick.

  • 0 in reply to Yannick Fortin
    TI__Prodigy 360 points

    Hi Yannick,

    A low capacitance TVS diode placed next to the pins of the TCA4307 would be advised. Please refer to this document for additional information, 

    How to select a Surge Diode.

    Regards,

    Jason