• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 32 subscribers
  • Views 807 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LMH0307 SDI output short

Adam Toner0
Prodigy 40 points
Other Parts Discussed in Thread: LMH0307, LM3100

We have a design that uses the LMH0307 SDI cable driver.  On our second batch of boards we have discovered that the SDI output pin (pin 12) on some of the cable drivers is shorted to the 3V3 supply rail.  At first we assumed this was an assembly error, but further investigation shows that this pin (and associated PCB trace) are not adjacent to any 3V3 traces.  In order to isolate the issue, we then removed the passive components on this trace, leaving only the LMH0307, and the short remains.  So the short must be coming from the LMH037 itself.

In our last batch of boads we have 6 boards which are OK, and 4 boards with this issue.  The short was discovered on the boards when they were powered up for the first time, directly from manufacture. 

So my question is, what could cause this short between LMH0307 pin 12 and the 3V3 supply?

Adam

  • 0
    TI__Guru* 88835 points
    Hi Adam,

    I believe this could be due to some type of electrical over stress. I checked and this is not reported by other customers. This could be either due to ESD, big voltage spike on 3.3V or charged coax cable among other things. Also, please check to make sure DAP under the part is not shorted to pin 12.

    There is one experiment that could shed light into this. Please build 10 boards. Before we power up the boards, check to make sure there is no short to VCC or low impedance to GND. Then do not load 75 ohm to 3.3V or return loss network(75 Ohm R and 6.8nH inductor). In other words do not load any passive components attached to pin 12. Then power up and down many times to see if you see SDO pin 12 shorted to Vcc. This experiment will help us to further identify the root cause of this issue.

    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 40 points

    Hi Nasser,


    Thanks for the quick response.  I had also considered an ESD event as a possible cause, but to me this seemed unlikely since all four boards failed in exactly the same way, and direct from manufacture.  I did some more testing based on your advice:

    First, I can confirm that pin 12 is not shorted to the DAP, which is connected to ground. 


    I took one board and removed the passive components to isolate the trace to pin 12.  I checked that there was no short to 3V3 and then power cycled the board 100 times.  I again checked for the short and found that the board was still OK.  I next replaced the passive components (pull up and return loss network), and then power cycled the board 100 times more.  Again, testing for the short I found that the board was still OK.


    I decided to take a closer look at the power supply during power up/down.  Attaching a scope probe to the 3V3 line (close to the LMH0307), I power cycled the board repeatedly.  I saw that the power supply behaves nicely during power up (no overshoot).  However, during power down I see a spike on the rail 1mS before the rail drops.  I think this happens when the switcher (TI LM3100) shuts down.  In some cases this spike reaches 5V and has a duration of < 5nS.  Could this be enough to damage the part?

    Regards,

    Adam

  • 0 in reply to Adam Toner0
    TI__Guru* 88835 points

    Hi Adam,

    LMH0307 has passed Human Body Model ESD of more than 4KV. In this test a 4KV pulse with 1ns pulse width is generated. So i have hard time believing 5V with less than 5 ns pulse width could be causing this problem. A couple of thoughts:

    1). Do you use 10uF capacitor on 3.3V that is pulling 75 ohm ouput to VCC on SDO pin 12? If not can we add this ?

    2). Could it be that charge 75 ohm coaxial cable could be causing this high voltage spike? Can we try plugging in and out the cable?

    Please let me know so we can decide what we could do next.

    Regards,,nasser

  • 0 in reply to Nasser Mohammadi
    Prodigy 40 points

    Hi Nasser,


    To answer your questions:

    1. Yes there is bulk capacitance on the 3V3 rail, but it is not directly adjacent to the LMH0307.  The closest is a 22uF cap about an inch or so away.  I will add a local 10uF.

    2. I don't really know if this is possible?  However, I have tested plugging / unplugging the cable 100 times (with the unit powered) and there is no problem.

    Do you have any other thoughts on what could have caused this fault?  The six working boards that I have (and three from a previous batch) seem very reliable and I cannot get them to fail.  Howeverm the four faulty boards have failed at or before the first power up.  I am still a bit concerned that there is an undelying problem that could cause yield or reliability issues in the future.

    Regards,

    Adam

  • 0 in reply to Adam Toner0
    TI__Guru* 88835 points

    Hi Adam,

    1). Could it be the case where these four devices were already damaged before they got installed on your boards? 

    2). How about if we run the following test. You can get 4 samples of LMH0307 devices. Check to make sure there is no short on SDO pin. Then install these parts on your failing boards. Again before you power up the board please check to make sure there is no short. Then power up and down. Do you see any failure?

    3). Another test would be to take one of the parts from your 6 good boards and put them on one of the four bad boards. This tells us if the problem is moving with the board.

    Regards,,nasser

  • 0 in reply to Nasser Mohammadi
    Prodigy 40 points
    Hi Nasser,

    1. I suppose the parts could have been damaged before they were installed but I am not sure how? We used a contract manufacturer to assemble these boards and they are just as puzzled as we are.

    2. It looks like I can only order 1 sample through the website, but regardless we will replace the parts are re-test as you suggest.

    Thanks and regards,

    Adam