In the lab, both simultaneous switching noise (SSN) and crosstalk look the same.\u00a0 They appear as unwanted pulses of energy that line up with the (aggressor) signal edges.\u00a0 However, the mode of energy coupling is much different between SSN and crosstalk. \u00a0In the case of crosstalk, they are lining up with the edges because the signal edges are coupling energy onto the victim signal through electric (and magnetic) fields.\u00a0 This occurs from one trace to another, and increases the closer those traces are.\u00a0 SSN, however, couples noise through the power distribution network (PDN).\u00a0 If the impedance of the board PDN is too high at\u00a0the IC power pins, the switching current of the I\/O buffers will\u00a0induce a voltage onto the other I\/O lines.\u00a0 And because these current demands occur as the signals\u00a0are switching, the resulting SSN appears as a pulse that lines up with the signal edges.<\/p>\n
How can these two phenomena be distinguished?\u00a0 Well, in the lab, you could try to toggle only the nearest two bits to the victim signal.\u00a0 Most of the crosstalk on a single layer will come from the nearest two aggressors.\u00a0 And with only two bits toggling, if the problem is indeed SSN, there should be a significant reduction in the coupled noise.\u00a0The easier solution, however, is to run SI and PI simulations during the design phase to ensure such problems are avoided in the first place.<\/p>\n