5 Replies Latest reply on Jul 9, 2009 11:57 PM by daniel.ibanescu

    LVS of LDDP, LDDN devices



      I recently discovered that some HV mos devices with non-swappable source/drain were defined as standard mos and so LVS was allowing the source/drains to be swappable.



      After reading the documentation I discovered they should be defined as type LDDN instead of MN. After enabling this in the LVS deck the devices were extracted as subcircuits named LDDN.



      The problem is the source needs to netlist several HV nmos symbols as subcircuits of type LDDN, but when the different NCF files point to the same model (LDDN) it gives an error and doesn't netlist. I'm sure this is a common requirement so any help would be appreciated.



        • 1. Re: LVS of LDDP, LDDN devices


          Hi rjackson,



          Did you ever find a way to solve this problem?



          I am familiar with LDD usage in Calibre LVS but not very familiar with NCF/netlisting issues.



          I would be happy to collaborate with someone that is reading this, in case working together may help solve the problem.






          • 2. Re: LVS of LDDP, LDDN devices

            Hi, the statement to add to your cdl netlist is





            Select Check LDD in your cdl netlist out to print the statmement .LDD


            The LDD designator declares that source and drain have different characteristics and cannot be swapped.

            • 3. Re: LVS of LDDP, LDDN devices

              Can you tell me what's the difference between using M devices with "LVS Builtin Device Pin Swap" option  "NO" and using LDD devices?

              • 4. Re: LVS of LDDP, LDDN devices

                I think there might be several aspects we could discuss but one of the most important might be consideration of all devices in the design and how they might work together at the same time.


                LVS BUILTIN DEVICE PIN SWAP NO can affect many types of devices such as MN, MP etc. and so may prevent swapping even when you don't want to do that. The behavior can be overridden with DEVICE statements but then it may become difficult if you want to have some M devices swappable and others not.


                LDD devices on the other hand are non-swappable by default and blend nicely with MN, MP etc.


                I wonder if this is one of those cases where there are multiple ways to accomplish the same goal, and the best choice may be different depending on specific circumstances.


                Maybe others have thoughts or comments regarding good or bad experiences with either method?

                • 5. Re: LVS of LDDP, LDDN devices

                  Yes, i think the method you choose depends on the project / technology. We are using M devices with S and D on the same layer so both ways work, but with LDDs you have to use subcircuits in the schematic or $LDD designator.