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.

TCA9406: TCA9406 damaged

Part Number: TCA9406
Other Parts Discussed in Thread: ESD321

We have a large number of "dead" devices. Reason unclear. Most devices do not work on first power up, but there was one (regarding to our tester) that was working, and did not work the other day anymore.

Dead devices work after replacing.


Is there anything incorrect to our circuit?

3.3V is powered by V5.0Bus. So V5.0Bus is powered before 3.3V.

There is a short amount of time during power down, where 3.3V is higher than 5V. It is unclear if this is a problem and if VCCA needs to be lower than VCCB only during normal operation...

OE does not have a pullup to 3.3V. This is done in the Evalboard. Is this needed ?

  • Hi Martin,

    Resistor on EN is'nt necessary but having one would help in terms of leakage current. 

    For power up, either side can power up without issues to the device. The important thing is, during actual operation VccB should be larger than VccA. 

    Are you hot plugging something into either side of the device? 

    When you say the devices are dead, what does this mean? (are devices just always HI-Z after power up? Are pins of SDA/SCL shorted to GND? what is the failure signature?)

    Is the device being handled correctly in terms of ESD? (handlers wearing ESD monitors?)

    How many devices are you seeing that are 'dead'?

    To double check, are the devices soldered one correctly? (make sure they aren't soldered 180 degrees flipped).

    -Bobby

  • Hi Bobby,

    thanks for your feedback. Here my updates on this:

    Are you hot plugging something into either side of the device?

    -> No hot plug

    When you say the devices are dead, what does this mean? (are devices just always HI-Z after power up? Are pins of SDA/SCL shorted to GND? what is the failure signature?)

    -> Device is not working. Further investigation pending

    Is the device being handled correctly in terms of ESD? (handlers wearing ESD monitors?)

    -> Actually yes. But it might be an ESD issue if there is no problem with the circuit. Should be consider to raise the 100 Ohm resistor to maybe 1k, to help lower the risk? Is there a drop in solution with higher ESD protection? We will go deeper in this.

    How many devices are you seeing that are 'dead'?

    -> So far we have ~8 of ~80 devices dead.

    Also an update from our tester:

    All devices were initially tested and working. Then stored in warehouse in later put in final device, where they were "dead" on first power up.

  • Some measurements on a non working device:

    all six pins measure a diode of~0.44V to ground . SCL and SDA lines have a high impedence of ~260k to ground...

  • Martin,

    Thanks for your patience, we'll work to get a response to you today.

    Regards,

    Eric Hackett 

  • Hi Martin,

    -> Actually yes. But it might be an ESD issue if there is no problem with the circuit. Should be consider to raise the 100 Ohm resistor to maybe 1k, to help lower the risk? Is there a drop in solution with higher ESD protection? We will go deeper in this.

    The 100 ohm resistor is technically adding a VoL shift to the I2C bus, making it larger would make it more difficult to send lows across it (VoL should always be lower than ViL of devices on the system). Changing it to 1k would also not help in terms of lowering the voltage seen by the device (only the current) which over voltage from the ESD could still damage the device. 

    Is there a drop in solution with higher ESD protection? We will go deeper in this.

    TCA9406 has a unique pinout so our existing portfolio doesn't have a replacement for it. I think the best approach may be to add an external ESD circuit if we are expecting the issue to be ESD. If you'd like, I can assign this thread to the ESD team in TI to make a recommendation.

    -> So far we have ~8 of ~80 devices dead.

    I can check with our quality team to see if they would recommend running any checks on the dead devices assuming you would be able to ship us a few of the dead units (and maybe 2 or 3 working).

    -Bobby

  • Hi,

    if the ESD team has an ESD solution that could be directly placed into the C47 and C47 position (which is a 0805 package) would be perfect. In this case we could remove the caps and solder further devices with this solution to see if it helps. We are still not 100% sure if the issue ist ESD though.

    We can send dead units to the quality team in the future.

  • Hi Martin,

    I've reassigned the ticket temporarily to our MPD product line (multiplexer and protection devices) to see if they can make a recommendation. They may reply on Monday since e2e tickets usually aren't handled on the weekends. Afterwards, the ticket will then be reassigned back to the group which owns TCA9406 (TRX-transceivers). 

    I suspect they may ask for more information about what ESD levels you would want to target protection for.

    -Bobby

  • Hi Martin,

    For I2C interfaces, we usually recommend TPD2E2U06 since it is a 5.5 V, 2 channel device that can protect both the SDA and SCL lines. Unfortunately, we do not have any single channel devices that are in an 0805 package. The closest device we have would be TPD1E10B06 in the DYA (SOD-523) package. You can check the package dimensions on page 21 of the datasheet and see if the leads would be able to make good contact with the current footprint for the capacitor. 

    Regards,

    Matt Smith 

  • Hi, thank you. TPD1E10B06 looks very promising. Would this also be a good device for 3.3V applications (I2C, but also general IO)?

    We will order some devices , do the testings and see if the issue is resolved. This may take a while.

    Thank you so far.

  • Hi Martin,

    It could work for 3.3 V applications, but it might not be the best. The general rule of thumb to to pick a device that has a Vrwm closest to your normal operating range to minimize clamping voltage. We do have some 3.6 V devices like ESD321, but none in the DYA package yet. 

    Regards,

    Matt Smith