1 Reply Latest reply on Oct 3, 2011 7:55 AM by samantha_lizak

    virtual connect problem for PEX

    c_hapham

      Hi All,

      I have a subcircuit layout which is LVS cleaned and PEX is also running fined (for R only). In my layout there are two vitual connections by names: VDDHV, VSSHV(supply and ground)

      Using PEX Options->Netlist->Advanced and click/unclick the button for "Leave disjont net model fragments unconnected", then I check File->Control File->View to see

      PEX NETLIST VIRTUAL CONNECT YES/NO.

      The problem is when I check the Pt-to-Pt Resistance from some device to net VDDHV (in both cases), Calibre RVE give me the value of 1.0E+38 Ohms which tell me the net VDDHV is broken???

      while net VSSHV doesn't have that problem.

      Anyone could help me with this. Thanks. (my calibre xrc version v2009.4_23.21)

      best regards,

      ha



        • 1. Re: virtual connect problem for PEX
          samantha_lizak

          Hi Ha-

           

          I'm reaching way back in memory here, as I never did point-to-point with the GUI.  (I did check the docs, though, to see if there were any disclaimers.  For the batch mode, you have to use layout names rather than source, but as the CI manual does not mention this I assume it translates behind the scenes.)

           

          Is it possible that the port you are checking is not connected directly to VDDHV?  As you know, devices interrupt nets (from the net view) -- the net that goes in to a device is a different one than the one that comes out.   So a point-to-point resistance report (at least as originally implemented; may have changed) shows only a nominal resistance across a device.

           

          And just a reminder -- for final runs, be sure to leave disjoint models unconnected. (PEX Netlist Virtual Connect NO). It uses arbitrary nodes; usually safe, but who wants to rely on probability when taping out?

           

          Best regards-

           

          Sam.