How do I remove the "-nan(ind)" in the first entry of the analysis output csv from .measure statement?

pablo_sole...@brown.edu

unread,

May 6, 2019, 5:13:17 PM5/6/19

to xyce-users

I have a .MEASURE line calculating a variable X but when I include X in the .print line the first entry in the csv gives "-nan(ind)". How do I make that value zero or something that wont shoot an error through xyce when parsing the csv?

xyce-users

unread,

May 6, 2019, 5:20:15 PM5/6/19

to xyce-users

Could you provide an example netlist? Or at least the specific combination of .PRINT and .MEASURE lines (and analysis mode) that produced this behavior?

There are are some operators, like I() and P() that do produce NaNs for .AC analyses, and that "bug/feature" may be fixed in the next release. So, you may be running into one of those issues.

Pablo Soler Portilla

unread,

May 11, 2019, 9:57:55 PM5/11/19

to xyce-users

I have a square wave output:

.MEASURE TRAN Q_freq FREQ v(q) ON=1.65 OFF=0.1

If I include this line the first line after the headers in the output csv will be -nan(ind)

xyce-users

unread,

May 13, 2019, 5:09:51 PM5/13/19

to xyce-users

Pablo,

We haven't seen this effect in our testing (e.g, the test Netlists/MEASURE/FreqTest.cir in the Xyce Regression test suite). I just re-ran that test's netlist with FORMAT=CSV rather than FORMAT=NOINDEX on its .PRINT line. So, it's hard to diagnose whether you've found a bug, without an example netlist that shows the problem.

--The Xyce Team

Reply all

Reply to author

Forward