ibm 130nm inductor ground-plane connection

[jlgr]

Hi there, in my circuit I'm using inductors of IBM-130nm (cmos8rf) with M1 ground-plane. When instantiating inductors we can select internal connection o external connection. For what I understand, internal connection should be used when one of the terminals is connected to an AC ground potential, which is my case. I supposed that the layout of such inductors would include contacts from M1 to RF Metal (E1 in this case), but it doesn't. Does anyone have worked with this? I will need to add vias from M1 to E1, but, where will be the best place to put them?
Thanks in advance

 

[shar2013]

If I recall well from that kit, internal connection means that the shield port is internal and you draw the path, where as external it adds a path to the inductor. In any case the shield port is not connected to the inductor ports

 

[jlgr]

Well, in both cases a path is added, the visible differences are:
1- When internal connection is selected this path is metal1, for external connection is metal2 with vias to metal 1, and
2- For internal connection this path is always under one of the inductor ports, while for external connection its orientation can be selected

- - - Updated - - -

This is taken from IBM’s doc, where explicitly suggests that shield must be connected to one port, specifically to the center:
“In many applications for integrated inductors, one end of the spiral will be at AC ground potential. When using the M1 groundplane inductor in an application of this type, it is suggested that the center of the spiral be placed at AC ground with the same connection that is used to ground the M1 groundplane”.

And is also added:
“If external connection is removed, the M1 groundplane’s central M1 connection should be connected directly to the innermost turn of the spiral through a low impedance via array”.

Now the same initial question, where to put this via array? Inside the inductor?

 

[jlgr]

By the way, if the via is drawn inside the inductor's area, we can add some layers over it so no drc violations exist, so, is possible to do so. But my doubt in that case is if the inductor's electrical model will remain to be valid. Any idea?

 

[jlgr]

well, I hope this helps if someone else will have the same doubt.
even if drc passes sucesfully when putting the vias inside the inductor, lvs doesn't, so I've decided to make the connection outside the inductor's area (don't know how else will accomplish "M1 connection should be connected directly to the innermost turn of the spiral through a low impedance via array").
anyway, I still have time for getting any suggestion, I will appreciate it.

thanks shar for your time