• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 32 subscribers
  • Views 72 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.

DS25BR100: 4-channel LVDS signal with Cat5e cable

Hiroshi Sakurai
Prodigy 50 points
Part Number: DS25BR100

Tool/software:

I am designing a board that will drive the 1.28Gbps LVDS signal from DS92LV16TVHG using DS25BR100 and send it to DS92LV16TVHG on the other board. A 50cm Cat5e Ethernet cable is used between the two boards.
According to Figure 1 on page 7 of www.ti.com/.../ds25br100.pdf, a 1.28Gbps signal seems to be no problem. However, the end user has specified that each of the four pairs of Cat5e cable on this board should have an independent 1.28Gbps LVDS signal. Is it possible to pass 1.28Gbps LVDS signals through all four channels without a problem under these conditions? Personally, I am concerned about the possibility of the LVDS signal not being transmitted correctly due to crosstalk between pairs.

  • 0
    TI__Guru* 93440 points

    Greetings Sakurai-San,

    This device is targeted to run at up to 3.125Gbps on each channel. Typically we expect at least 40dB isolation between each channel. Additionally there has not been a report of cross talk related issue with this part. Therefor we don't see issue running this device at 1.28Gbps on each channel. Please note cross talk can get affected by pcb layout or connector as well.

    Regards, Nasser

  • 0 in reply to Nasser Mohammadi
    Prodigy 50 points

    Dear Nasser,

    > Please note cross talk can get affected by pcb layout or connector as well.

    I too am concerned about this point. And I am aware that the Cat5a cable is where the crosstalk impact is greatest. One reason is that it carries 10 times more frequencies than normal Ethernet communication, and another reason is that all four pairs are used for independent LVDS communication. If only one pair is used, I'm not so concerned about the effect of Cat5e cable.

    The near-end crosstalk (NEXT) of Cat5e cable is 35.3dB at 100MHz. According to TIA/EIA-568, the NEXT at a frequency of fMHz is NEXT(100)-15log(f/100), which means that at 1280MHz the value is negative.
    However, this value is based on a cable length of 100m, so it is not clear how it would change with a cable length of 50cm.

    Do you know what to think?

    Regards, 

  • 0 in reply to Hiroshi Sakurai
    TI__Guru* 93440 points

    Hi Hiroshi,

    50cm distance is shorter but both at the same time both aggressor and incoming signal are stronger as well. So it is difficult to predict.

    This all depends on SNR and your receiver sensitivity. I think if we target 15dB SNR at minimum this should give us a good margin to receive data error free. 

    Regards, Nasser