DS25BR440: use for 2Gbt diff-signal

Part Number: DS25BR440


To whom may concern,

I am using this IC as line driver for 2G LVDS signal. I test with 1.5 Meter cable and 3.8 meter cable. I Have following questions:

  • I tried with and without PE and EQ pin enabled, and see no difference in wave form. Seem the PE/EQ are permennat turned on? I dont know if this is international? See Ocsilloscope screen shot, first  is the outout signal of DS25BR440 before push the data to the cable (I can not see different with or without PE, look same). Second on is the resive side output of DS25BR440, signal recovered after the long cable. (I could not see different with or without EQ enabled)
  • the other question is about the clk recoverage. The oscilloscope have issue to recover the the clock especially with long cable, the Eye diagram is getting blurr with long cable (I used 3.8 meter) even the wave form looks ok (from my point of view).  With short cable has less issue. This shows errors at the data capture side.  I did not use any certificated CAT cable, have quit some delay on the cable(9.5nS/m ). But is it a known effect?  Does a 3+ meter cable (tweist pair) already over the  DS25BR440's limitation? 2_1_5Meter_Before _Cable_and_after_Line_driver_1_PE_Enabled.jpg

4_1_5_after_cable and_after_line_driver_2_LTB_connected.jpg

  • Hi Yangjie:

          You should see  pre-emphasis difference from OUT0-3  output signal  with PE ON or Off.

            You can also see  EQ  difference from OUT0-3 with PE is Off., but you can not see any difference  for PE or EQ from input signal IN0-3.

            What is your PE and EQ setting  for 1st and 2nd waveform,?

           What  is waveform  sent to IN0-3 pins?

      for the cable used for test, the cable is between DS25BR440 and scope? maybe 3.8M cable is too long for this application, DS25BR440  PE is about 6db

    Regards

    Brian

  • HI Brain, thank you to help me out again! (I asked question last, was you replied Slight smile) Both waveform are measured at output0 with PE and EQ, 1st is at the transfer side and second is at the receiver side. I can see clearly that the signal is enhanced at the high frequency part (over and undershot are added compared to the input signal). My question is I compare the PE/EQ on and off wave form, I saw bare difference, even PE/EQ are off, the signal looks still enhanced.

  • Here are two waveforms at transfer side with PE  on  and off. I could not till difference. 

  • Here are two waveforms at receiver side Out0 with EQ  on  and off. I could not tell difference. Both case signal has been repaired compared to the input signal of DS25BR440. (So form my observation, PS/EQ pins are not really controlling the PE/EQ on and off, they are all time on)

  • Hi Brain, 

    About the cable length, it is still under investigation. I got info there is success use case with 5+ meters. So I believe it is some sw /protocol issue from our side. 

    Kind regards,

    Yanjie

  • HI Yanjie:

       You should see difference between PE on and off. Sorry I don't have EVM for this device to test at our lab.

      Can you also draw block diagram  for signal measured and also where is the cable located?

    Best

    Brian

  • HI Brian,

    I measured again. Yes, I can confirm with PE on and off, I could see signal difference at output. With PE off, I still could see the IC worked on the input signal and enhanced the high frequency part, so it is not 1:1 repeat the input signal (this is probably the point confused me before). But with PE ON the effect are more obvious and the enhancement was stronger. 

    So far you answered all my question, thank you for your support. We can close this thread. Slight smile

    kind regards,

    Yanjie