SOLVED - Problem to avoid the removal of some capacitors with low values with Spectre XPS tool while performing transient analysis

Dear VitoLima,

VitorLima said:

The Spectre XL is removing some capacitors and I need to avoid it,

To be honest, I am not sure what you mean by "Spectre XL". I have never heard of it - but I am not a Cadence tools expert. Do you mean ADE-XL with the spectre simulator? The version number you specify ("15.1.0.679"), if it refers to the spectre simulator appears to be quite old as the version I am using is:

$ spectre -W
sub-version 19.1.0.428.isr9

VitorLima said:

A message example:
The shorted instance X16830_g_stdcell_g32.CC16312 is removed.

Why do you need to include a capacitor that is shorted? It will have no impact on circuit performance.

You can prevent any version of spectre from removing instances by specifying its options from using any post-layout netlist reduction as well as preserving instances - options exist for both. However, I am not sure I understand the motivation for your question well enough (sorry!) to recommend an option(s) to meet your objective.

Perhaps other Cadence experts can shed some light on what you are calling "spectre XL". My apologies, once again, for my lack of understanding your question!

Shawn

It was my mistake, the correct is Spectre XPS. I'm working with power attacks in cryptography designs and I must ensure no interference.

There was no lacking and I truly appreciate your help. In fact, you gave me the answer of preserving all instances. I use this at Innovus and I didn't think to use here . 

Thanks a lot!

Just don't forget that these parasitics elements - capacitances and resistances - are calculated ("extracted") by parasitic extraction tools.

And their default settings often drop small capacitances, decouple small capacitances to ground, and do other things that are required to minimize the post-layout netlist size, but that can be very nasty for circuits sensitive to small parasitic elements.
If small parasitics are critical for your designs, I would strongly suggest to use a field solver (FS) - a built-in option in all extraction tools, with tightened accuracy settings (random walk-based field solvers allow that), on critical/sensitive nets.