extracting jitter from an eye diagram

Dear Svilen,

> I need to integrate this measurement in an script (ocean script perhaps). Is

> there a function that can do all this?

The cross() function might be your best option as you can measure the time between the threshold crossing of each data pulse and the ideal location of the respective threshold crossing. You know the ideal threshold crossing function from the data pattern you are supplying.

This will provide you with all the threshold crossing errors (analogous to the set of time interval errors for a clock signal). I assume this raw data can be used in whatever specific calculation you want to perform.

I hope this helps Svilen,

Shawn

Thanks Shawn,

I actually found a different way using the eye-diagram assistant from where I can get the histogram of the threshold crossings. Then I can do cross on the right and left sides of the histogram at a level of 0.5 and thus extract its span which is the width (within certain accuracy) of the eye-diagram edge crossings.

Svilen

Svilen,

You can also use the measurements at the bottom of the eye diagram assistant to get the standard deviation of the threshold crossing (and send this to ADE using the right mouse button menu over that measurement). A good rule of thumb to get the range of crossings is to multiply this by 6 (plus and minus 3 sigma). Note that you'd have to ensure that your eye diagram only has one transition region to do this.

Alternatively, look at

which provides a function abEyeCross that I wrote - you could use ymax(...)-ymin(...) of this function to find the peak to peak crossing times. The code can also be found at: - but the Cadence Support article gives more background.