Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/finfet-netlist-generation-tool.5973/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

finfet netlist generation tool

A

Amira

New member
Is there any one know how to generate a netlist for finfet transistor from TCAD simulation results??
 
Amira said:
Is there any one know how to generate a netlist for finfet transistor from TCAD simulation results??
Click to expand...

What type of netlist? If it is spice I think you'll need to fit a model to the TCAD results or do your own Verilog-A modeling. Don't think there is anything FinFET specific to the procedure.
 
Most transistor-level netlists come from either:

  1. Schematic capture tool, for pre-layout circuit simulation
  2. An extraction tool, for post-layout circuit simulation
 
yes , I need to generate spice netlist to use it in Hspice simulation, my available information is some I , V numerical values from TCAD device simulation, I need to use them to generate spice netlist based on BISM-CMG model, is there any tool can do that ??
 
ProPlus Design Solutions Inc.
bsimproplus and many other spice extraction tools can help. It belongs to general optimization scope. So you can use Matlab, Octave, R, and even VB inside excel can be used to do the job.
After transistor spice parameters extraction, you can turn on the mix mode simulation to compare with multiple transistors numerical simulation and optimize parasitic and any fancy stuff you want to top up the BSIM model.
My humble advice would be using VB in excel, more straight forward in the beginning. After you learnt quadratic programming by heart, you can easily play with other math package.
 
please advise is this tool free ? or I must have a license ??
Is there any other free tool ?
 
thanks for reply , sorry for bothering you
could you clarify more could I use( I, V )numerical values from TCAD simulation (for finfet ) as input to the Ngspice or DEVSIM to generate model card for the finfet ??
 
Amira said:
could you clarify more could I use( I, V )numerical values from TCAD simulation (for finfet ) as input to the Ngspice or DEVSIM to generate model card for the finfet ??
Click to expand...

Not really, you need to take some of the parameters of the SPICE model card and tune them so that the output of the SPICE simulation matches the (I, V) from the TCAD simulation. How much effort is needed will depend on the use case of the fitted model afterwards. If it is digital type simulation a more crude fit with some basic parameters for Ion, Vt, sub-treshold slope and gate input capacitance may be enough. If analog or RF type simulation much more effort and parameter tuning will be needed to match the first and even second order derivatives of your TCAD (I, V) curves.
From the other side I think it should also be possible to use the raw (I, V) and implement an interpolating model in Verilog-A although I heard from some people this can give convergence problems during SPICE simulation.
 
Last edited by a moderator:
please advise are these methodologies used before and result good results ??please state where if you know
 
please advise are these methodologies used before and result good results ??please state where if you know
 
Amira said:
please advise are these methodologies used before and result good results ??please state where if you know
Click to expand...

Sure, if you get a model deck from a foundry it will also have been calibrated to measurement results and doing it to measurements is more involved as to TCAD results as the former has some noise added to it.
I'm no expert but I did see the results of people dedicated to the task. What I do know:
  • It's no easy feat. there is no push-button solution; you will need to constantly feed back results and improve calibration procedure
  • After a while it seems to become a boring job. At imec it was very difficult to keep people at the calibration job.
 
You must log in or register to reply here.
Back
Top