• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 63 subscribers
  • Views 109 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.

LM3671: ESD (Electrostatic Discharge) Problem with no potential solution

Trevor Buhler
Prodigy 10 points
Part Number: LM3671
Other Parts Discussed in Thread: STRIKE

Tool/software:

I am using the LM3671MF-3.3/NOPB in a circuit for my company as a 3.3V regulator. In this design we have a USB 2.0 connector that uses Type A USB. We have the casing of the USB connected directly to our digital ground (same ground that our LM3671MF-3.3/NOPB uses).

When I provide an electric shock to the USB casing, the 3.3V regulator gets damaged and no longer works. When I replace the 3.3V regulator with a new one the board operates as normal. 

As you can see we have an ESD problem and it is affecting quite a few of our boards which results in unhappy customers. I have tried the typical solution of connecting a TVS diode directly across the input of the 3.3V regulator. This resulted in what seemed to be a more reliable solution, but still after shocking a few times the 3.3V regulator eventually fails. The tvs diode I used was the SP0502BAHTG. I tried both a unilateral and bilateral (by connecting 2 SP0502BAHTG's in series) connection for my TVS diode and it failed when shocking the USB case.

What I believe is the problem is that the LM3671MF-3.3/NOPB has an absolute max voltage of 5.5 volts and we input 5 volts into it which does not give us a big margin to work with. This means that if the SP0502BAHTG did its job perfectly it has a max clamping voltage of 8.5V which is 3 volts higher than the absolute max which I believe is causing damage. The hard part is there isn't really a TVS / Zener diode that meets the criteria to clamp at 5.5 volts, yet have a reverse standoff voltage of 5 volts. Even if there were such a diode I am not sure that I would be comfortable using it since there is only 0.5V margin between shorting the diode and the diode being an ESD protection.

Do you have any good solutions to ESD problems with the LM3671MF-3.3/NOPB? Have you seen this problem before?

Thanks,

Trevor

  • 0
    TI__Mastermind 21580 points

    Hi Trevor,

    The LM3671 is qualified for ESD strikes according to the JEDEC standard mentioned in the datasheet. The pulse widths of these ESD strikes are typically quite narrow, just a few ns.

    If the spike you see in your application is longer it may be more than an ESD strike and this may cause the device to get damaged. My suggestion is to experiment with a damped input filter using a ferrite bead ( https://www.ti.com/lit/an/slvafe0/slvafe0.pdf ). Small case size cap with low ESL may also be useful with the filter to to provide a low impedance path for the spike. You can monitor the voltage at the input pin of LM3671 with the extra filter to see if it is below ABS max spec.  

    In case the spike voltage can't be reduced below the ABS max rating of LM3671, then you may have to consider an alternate device with a higher ABS max rating.

    Best regards,

    Varun

  • 0 in reply to Varun John
    Prodigy 10 points

    Hi Varun,

    Since we do not have any ESD simulating hardware, the way that I am shocking the USB case is with a barbeque grill ignitor / shocker and its possible that the shock could be a longer duration than a typical ESD strike. I could attempt to add a ferrite bead with a low ESL cap to see if that makes a difference.

    I am also now questioning whether or not these defects are coming from ESD or some other source. These regulators keep failing and coming back from customers, but the only way I have been able to replicate it is by shocking the USB case. Would it be possible to get someone from Texas Instruments to run a failure analysis report so we can figure out what the root cause is? 

    Thanks,

    Trevor

  • 0 in reply to Trevor Buhler
    TI__Mastermind 21580 points

    Hi Trevor,

    Other than the electric shock you applied, could you check if there any voltage spikes at the input of LM3671 during plug in of the USB cable or normal running of the application?

    When adding the ferrite you might also have to add an extra damping cap (Cd with Rd) as the inductance of the ferrite along with the 4.7uF (C6) in your schematic can form an undamped LC. 

    It's possible for us to run an FA to see what is damaged in the IC. However the root cause for the damage is most likely going to be an electrical overstress at the input. 

    You can submit a return request here: https://www.ti.com/support-quality/additional-information/customer-returns.html

    Best regards,

    Varun

  • 0 in reply to Varun John
    Prodigy 10 points

    Varun,

    There is no spike on the input when plugging / unplugging the appliance. When measuring the input voltage with an oscilloscope I get these voltage measurements on the input while the appliance runs in its normal steady state:

    Max voltage: 5.36V, Min Voltage: 5.12V, Average Voltage: 5.22V

    These voltages are within the range of the absolute max for the IC, although 5.36V spikes are a little close to the absolute max. The only thing that runs off 3.3V in our application is the microcontroller, EEPROM, and keypad so I don't think current consumption is a problem. I could try adding some of these components recommended and see if that helps with anything.

  • 0 in reply to Trevor Buhler
    TI__Mastermind 21580 points

    Hi Trevor,

    Thanks for the info. The part is rated for 6V abs max and should be able to handle voltages up to that value. You can try out the electric shock tests with the extra filters and see if it helps filter out the spikes.

    Best regards,

    Varun