• State
  • Locked Locked
  • Replies 9 replies
  • Subscribers 36 subscribers
  • Views 954 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.

High Failure Rate on LMH0387

Dave Sharpe
Prodigy 140 points
Other Parts Discussed in Thread: LMH0387

I'm seeing a high rate of field failures associated with the LMH0387. This appears to be both Receiver and Driver, although mostly receiver.

Has anything changed with the fab process as of late?

While I don't think it's an ESD related issue, how sensitive is the BNC_I/O port?

Do you have the capability to de-cap chips and tell me root cause?


Thanks.. I'm investigating some other potential cause.. Voltage is within spec. But would like further input on power line noise susceptibility.

Dave

  • 0
    Prodigy 140 points
    Additionally.. The Max p-p input voltage swing (as EQ) is listed as 950mVp-p.. What level damages the input?
  • 0 in reply to Dave Sharpe
    TI__Intellectual 870 points
    Hi Dave,

    Are the "Failure Chips" damaged or your just are seeing errors on your system?
    How about the current failure rate?

    If this is one "damaged" issue, can you contact your local support to for requesting one FA(failure analysis)? The FA team will be able to figure out the root cause of the failure.
    The ESD spec of the highspeed IO is >6KV.
    I do not have the number for the absolute Max pk-pk input voltage at this moment. May i know what the current pk-pk voltage in your application is?

    If it is one "seeing errors on your system" issue, I would like to check both the schematic and PCB related with LMH0387.

    Regards,
    Liang
  • 0 in reply to Liang Liu1
    Prodigy 140 points
    I have seen about 8 fail in the last 30 days. All completely fail the SDI interface.. Most being the EQ portion.
    One concern I wanted to address was what might happen if composite video was supplied to the BNC. Typically a 1Vp-p signal which is larger than the 950mV specified in the datasheet. The nature of the product is that both composite and SDI are both present. I ran video into the EQ front end for about 6 hours and the 387 did not seem to have any problem with that.
    This is the same PCB/Board/Layout that has been in production for about 2 years with no significant problems until the past quarter.

    Dave
  • 0 in reply to Dave Sharpe
    TI__Intellectual 870 points
    Hi Dave,

    OK,
    I would propose that you go ahead and contact your TI's local support team for the FA process, so that the FA team will be able to figure out which part of circuit is broken inside the silicon.

    At the same time, I am checking with my colleagues for the absolute max pk-pk for the input signal on the SDI receiver interface.

    Regards,
    Liang
  • 0 in reply to Dave Sharpe
    TI__Guru* 93670 points
    Hi Dave,

    Per SMPTE spec the maximum input swing voltage is 880mV. However, in lab we have used 1V peak to peak and have not observed any issue so i am thinking this problem may not be related to composite 1V peak to peak signal.

    Do you have the capability to run return loss measurement? Do you see a difference between the good vs the failed boards? LMH0387 has internal return loss network and this could be helpful.

    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 140 points

    Thanks for checking on the maximum input swing. Still trying to identify the root cause. I do not have the ability to do return loss.. As close as I can get would be with an Eagle RLB-150 and spectrum analyzer with limit of about 1GHz. Unit/s can be provided as necessary..

    Dave

  • 0 in reply to Nasser Mohammadi
    Prodigy 140 points
    Hi Nasser,
    With further developments in trying to resolve this issue.. More data has come to light.. If you could expand on the maximum input swing at which damage will occur. While I don't know exactly what the customer is doing it appears there is a -2/+2.5V initial pulse or signal that is on the front end. I highly suspect that this is one potential cause wiping out the front end of the LMH0387..

    Regards, Dave
  • 0 in reply to Dave Sharpe
    TI__Guru* 93670 points
    Hi Dave,

    Maximum input voltage swing is about 950mV and here we are applying twice as much. It is possible this may have damaged the device. Truly this is beyond device spec, could the customer maintain the input voltage swing below 950mV to see if this takes care of this issue?

    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    TI__Expert 3050 points
    Hi Dave,

    Please refer to the Absolute Maximum Ratings Table in the datasheet. For input voltage abs max rating is −0.3V to VCC+0.3V.

    Exceeding this range may cause permanent damage to the device. This device is not designed to support the signaling levels you mentioned:
    " -2/+2.5V"

    There is no need to submit these units for failure analysis if the customer is supplying these input levels. The result with simply be electrical overstress (EOS) due to the violation of the Abs Max Table.

    As Nasser noted, the customer will need to bring the signal levels within the datasheet specified rang or consider another solution. If you can explain what the customer is trying to do in detial perhaps we can help find solution.

    Note: Any device that the customer has stressed with these large input signals should be considered over-stressed. Any further testing should be done on devices that the customer has not already electrically over-stressed.


    Mike Wolfe
    DPS APPS / SVA