Signal Integrity

Power-aware Signal Integrity and EMI/EMC On High-speed Digital Chip-to-Chip Links

Signal Integrity header image 2


Posted August 18th, 2009 · 8 Comments · Application Note, Survey


Although scattering parameters (S-parameters) are widely used in RF and microwave engineering, it seems their adoption isn’t universal in the signal integrity field. I’m hoping this informal survey will air the issues and help us all understand what the barriers (if any) are. Also, please leave a comment if you have opinions not covered in the survey options.

Do you use S-parameters in your signal integrity work?

Loading ... Loading ...

For linear components, what factors prevent you from using S-parameters more? (Check all that apply)

Loading ... Loading ...

If you are in the “No” or “Not comfortable” camps, you might be interested in these PDFs on network analysis with S-parameters. They aren’t specific to signal integrity, but are nevertheless useful as a general introduction:

Tags: ··

8 Comments so far ↓

  • Jerry Johnson

    The old question of time domain vs. frequency domain. I’ve fought this battle for decades. First in the mainframe computer industry. Each side has proponents and argues for it’s viewpoint. My view is sort of like “which hand would you like to give up?”. Actually I’d prefer to keep both, using either or both when appropriate. Why give up either? Favor either when that one works best.

  • rick

    S params have limitations like finite freq range, limited # of entries, etc…. I use them in both freq and time domain.

  • Ike

    “S params have limitations like finite freq range?..”
    Can you advise us for a model with unlimited freq. range?

    • Colin Warwick

      Hi Ike, Thanks for your comment. I added a bullet at the bottom of the post about our method of extrapolating the frequency response in the S-parameter file by 2x to 10x in order to make the time domain response causal.
      Hope this helps,
      — Colin

  • Subramanian

    working on s-parameter model,with what respect signal quality can be changed

    • Colin Warwick

      Hi Subramanian,

      Thanks for your question.

      Can you be more specific?

      In general, building an end-to-end simulation based on behavioral models of the components (Tx, channel, Rx, crosstalkers, PI effects) allows you to explore the design space.

      Hope this helps!

      — Colin

  • Harish

    Trying to calculate NEXT & FEXT Crosstalk(2- Differential Pairs) using S-parameter generated with help od SI simulation tool for multiboard simulation,
    The Transmission link has a High speed connector in between, were the connector vendor has give S-parameter(s8p) for few connector pins.
    While simulating i have considered two separates parameter(S4p) models for two differential pairs terminating other 4-ports. so want to know the right procedure to use s-parameter while simulating.

Leave a Comment