Measuring SSN

[Altera_Forum]

Hi, 

 i have seen people trying to measure SSN with the toggling aggressors on the FPGA floating (not connected to anywhere). How accurate is that compared to aggressors with load?

[Altera_Forum]

You must have load caps in a SSO measurement. the discharge from the load capacitors flowing back into the device are what cause the ground bounce event. The larger the load the larger the bounce event. You need to choose the load that best represents what you expect as trace and load capacitance in your application.

[Altera_Forum]

Thanks Joe for the explanation :)

[Altera_Forum]

One effect that load caps have is reducing the amount of SSN seen at the far-end. If you toggle a bunch of aggressors and want to know whether the noise seen on a victim pin will result in a spurious far-end toggle, then you definitely need to include load caps. The load cap (or far-end chip) will serve as a low-pass filter, and thus reduce the magnitude of the peak noise event seen by the far-end device. 

 Bottom line: SSN is very complicated. You need to measure the SSN in your real system if possible, since the board thickness, I/O standards, loading, trace, far-end load, PDN decoupling, and even the oscilliscope and probing technique you use will all influence the magnitude of the noise you measure.  Regards,  Paul Leventis Altera Corp.