2 Replies Latest reply on Sep 14, 2012 11:18 AM by weston_beal

    LVTTL CROSSTALK analysis

    sudhagar_k

      Hi All,

       

                     I have a 3 single end (3.3V LVTTL )nets placed at a distance of 7.8mils(almost 8mils). The center net(victim net) is operating @1MHz while the two adjacent aggressor nets are soperating @ 50MHz. DUring crosstalk analysis the center net is picking some noise in the range of 3.6V which is actually the maximum allowable voltage level.

                         Now as we know 50MHz is a very low frequecny,also the nets are spaced for almost 8 mils then i dont find any good reason for the cross talk to take place.

                          Also all the three nets are on the same plane, so theoratically, in inductive coupling the negative overshoot should be more in contrast of what am observing as positive overshoot.

       

                          Kindly explain.

       

      Regards,

       

      Sudhagar

        • 1. Re: LVTTL CROSSTALK analysis
          yu.yanfeng

          How long your traces?  What is the distance to reference plane? What is the setting for default Tr/Rf for IC? all of these take effects on crosstalk result.

           

          Yanfeng

          • 2. Re: LVTTL CROSSTALK analysis
            weston_beal

            Clock frequency does not matter. That just tells you when the crosstalk peaks should occur. The amount of crosstalk depends on the rising and falling times of the aggressor signals. You mentioned inductive coupling, but not capacitive coupling. It is interesting to study these separately, but in reality we always have both. Add in the capacitive coupling effect to your assumptions and see if it correlates better with the simulation results. Another common mistake is to look at only the separation between traces, neglecting the distance to a reference plane. The more important measure is th ratio of trace-to-trace space to the distance to reference plane. As this ratio increases the crosstalk decreases. Furthermore, we often study ideal coupling configurations with one uniform coupled section. This is very rarely the configuration we have in a real PCB layout. The position of coupling sections among other uncoupled or differently coupled section changes the overall crosstalk observed at the various IC pins.

             

            Weston