• State Resolved
  • Locked Locked
  • Replies 14 replies
  • Answers 6 answers
  • Subscribers 32 subscribers
  • Views 1282 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.

SD384EVK: Equalized 3G SDI signal has too much jitter on output signals, no eye

Joe Brackett
Prodigy 130 points
Part Number: SD384EVK
Other Parts Discussed in Thread: LMH0387, LMH0384, LMH0303

Hi

I have a commercial bidirectional HDMI to 3G SDI converter box.  If i take this box and hook a computer video output to is and drive the sdi signal into the SD384 eval board and its output into an oscope the eye diagram of the equalized signal is junk.  There is too much jitter on the lines and it causes the FPGA to see too many CRCs to recover the signal (when the FPGA is hooked to the receiver instead of the oscope).  If i have the FPGA drive signal out to the SD303EVK transmitter board and out the same HDMI to 3G SDI box the signal is recover fine and the image is transitted to an HDMI monitor cleanly.  I opened the converter box and found they use a TI part that I cannot find.  The chip has the markings: L0604A2 86U AX42.  This part seems to recover the 3G SDI signal far better than the LMK384 chip thats on the eval board.  What is the difference in these two parts?  Why would the LMK384 board be unable to create a clean signal out?

The cable connecting the SDI ports is a 75 Ohm cable and is only 2 feet in length.  No where near the hundreds of feet the SDI signal should be able to travel.  The same cable is used for both directions so the cable isnt the issue.  Also the FPGA has been tried as the source of the HDMI signal (FPGA drive out to the SD384 board then to the SD303 board and the equalized signal is still junk).  the FPGA output eye was verified to be very clean and have a very large eye well within spec.

Thanks
Joe

  • 0
    TI__Guru* 88925 points
    Hi Joe,

    1). You should be able to put your scope or video analyzer on the output of the L0604A2 device and check data rate and eye quality. Is this possible?
    2). Please let me know if L0604A2 is a 7mm by 7mm package device.
    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points

    Hi

    The device that uses the L0604 is no longer working because of something else.  The fact that the video is recovered perfectly tells me the eye is good.  it does look like a 7mmx7mm package.  The cable length doesnt matter correct?  SDI is designed for very long distances, could the short distance be the issue?  why would there be no eye on the LMK384?  there arent any controls on the part to mess with i would think the eye would be perfect at short distances.

    Thanks

    Joe

  • 0 in reply to Joe Brackett
    TI__Guru* 88925 points
    Hi Joe,
    L0604 part could be LMH0387 device.
    Cable length does matter and you can refer to each part's data sheet to find maximum cable reach versus data rate.
    Also, at short distance there could be some over-equalization that may cause issue. However, I have not observed this. On LMH0384 there is EQ BYPASS pin#7. You can pull this pin high to see if you see any improvement.
    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points
    Hi
    I will try and pull the pin high and see if it makes a difference. I am currently out of the office and it might take me a few days before I can try this. I will get back to you.
    Thanks
    Joe
  • 0 in reply to Joe Brackett
    TI__Guru* 88925 points
    Hi Joe,
    Sure. I close this case for now and please re-open it when you try this.
    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points

    Hi

    The bypass pin fixed the issue.  Thanks for your help.

    Joe

  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points

    The L0604 is not the LMH0387 part, its 4mmx4mm, i was mistaken.  It has 6 pins on each side totaling 24 pins from what i can see.  any other ideas what part this might be?

  • 0 in reply to Joe Brackett
    TI__Guru* 88925 points
    Hi Joe,
    Sorry I don't know what this L0604 device is. I am glad to see enabling BYPASS take care of this issue.
    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points

    Is there a better part we should use other than the LMH0384?  We want to be able to recieve and SDI signal on a variety of cable lengths from 1 foot up to 100 feet.  The bypass allows for the short cables to work, the equalization allows for the long cables to work but the in between length cables dont work.  seems like the medium length cables need some equalization just not all of it.  What would you recommed we should use that would allow us to operate in all these different senarios (ie dynamic equalization of a variety of equalization settings)?

    Thanks

  • 0 in reply to Joe Brackett
    TI__Guru* 88925 points
    Hi Joe,
    Actually LMH0384 should've worked and it is capable of adaptive equalization. Do you have standard video pattern generator? I am thinking perhaps there could be problem with your HDMI to SDI converter. I am assuming your SD384 would work with this type of generator.
    Also, you can do an additional test. Please take the output of your converter - through 1m short cable - to 75-ohm scope plug in. Please send us this scope shot.
    Regards,,nasser
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points
    Hi
    The HDMI to SDI is a standard off the shelf converter box by blackmagic.
    I do have standard video patterns to test the interface and I tried them in them manner i stated above, i used a SD386 sdi driver to convert the signal then back through the sd384 board where they eye pattern looked bad before i bypass the equalization. I will work on getting an eye diagram of what the 384 board creates.
  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points

    What horizontal and vertical resolution do you want to see? 

    i used 2 different scopes

    The first picture is 200mV/div 20ns/div (red plot)

     

    the second picture is 200mV/div 10ns/div (yellow plot)

    I am not sure what the sdi signal should look at or how to analyze it.

    some more information...  if i take the signal from the converter box and route it back into an sdi to hdmi converter box it works fine regardless of the cable length.  test patterns do not work either, same results.  if the receiver has adaptive equalization why would it be over equalizing the signal?  why would i have to bypass equalization for it to work?

  • 0 in reply to Joe Brackett
    TI__Guru* 88925 points

    Hi Joe,

    1). I didn't get your point saying "test patterns do not work either". Could you please clarify.

    2). With adaptive EQ and short cable, device should be able to equalize. In your earlier test, you took output of the FPGA , used LMH0303 cable driver, and a short cable. Under this condition, it worked fine. So I don't think there is over equalization here. I think there is something related to the signal on this HDMI to SDI converter and this is why we asked for scope shot. Can you please set your scope so we can capture output eye diagram. We need to see one or two unit interval of the eye diagram.

    Regards,,nasser

  • 0 in reply to Nasser Mohammadi
    Prodigy 130 points
    Hi
    1) the test patterns you mentioned would probably work, they don't. no video works through the SDI equalizer with a short cable unless equalization is bypassed
    2) when i took the fpga output to the LMH0303 i brought the SDI into a SDI to HDMI converter box and it worked fine which tells me its not the TX thats the issue. It never worked with the LMH384. The SDI signal is an analog type signal, how would you capture the eye on a signal like this? its not the typical unit interval. I have capture the eye on the 50ohm side of the equalizer but dont see a way to capture the SDI side.