I'm doing PCIe simulations (Gen1, one lane x1), and I wonder how can crosstalk effects be observed.
I'm going to use a cable that has 8 differential pairs, and besides transfering PCIe (1 TX, 1 RX) on it, there are also SATA II signals (1 TX, 1 RX) 3Gbit/s and, PCIe reference clock (differential signal) (100 MHz) on the same cable.
I would like to take into account how crosstalk from PCIe RX, and SATA II and reference clock signal will affects PCIe TX signal, and what effects this crosstalk will have on resulting eye at the receiver.
I have IBIS ver 4.2 model of the PCIe transmitter and I have an 32-port Touchstone model of the cable.
Could this setup work, If I want to see crosstalk effects on PCIe TX trace, then I terminate that trace on both ends with 50 Ohms resistors, and the other traces I terminate one end with 50 Ohm resistors and put driver model on the other end with corresponding IBIS driver?
Then by using HyperLynx SI Oscilloscope, I could see how much mV I get into the victim line, and using this formula: (Equation 1, page 18 in this paper: http://www.iconnect007.com/pdf/DesignCon_Greg_Edlund.pdf) convert this to jitter (ps)? How do I get how much has amplitude on the eye decreased?
Basically I would like to see how much margins are decrease due to crosstalk effects.
SInce there are several variables here I created a service request so that I can work with you on all of your questions. Once we have a solution I will post a summary.