STBJT_Model: STBJT Model

The Spectre bjtst model is translated to the ADS NPN or PNP STBJT Model. For translation information on the BJT device, refer to BJT Device for Spectre.

For more information on the ADS model, place the model in a schematic and choose Edit > Component > Edit Component Parameters to view the model parameters. You can also click Help in the component editor dialog box for additional information.

Example Spectre Command Line:

model \_stj1 bjtst type=npn is=1e-16 isn=1e-12 ...

Spectre Netlist Syntax:

Example ADS Command Line:

model stj1\_bjtst STBJT Af=1 NPN=yes Is=1e-16 Isn=1e-12 ...

ADS Netlist Syntax:

ADS Schematic Symbol:

MOSFET Models

The following information describes how the various MOSFET models from SPICE and Spectre are translated to the corresponding ADS models.

SPICE Models

All Mosfet devices in SPICE reference a model by its instance name. Each Mosfet model in SPICE has a keyword NMOS or PMOS, as well as a Level parameter.
The NMOS/PMOS keyword is used to determine what device to place in the schematic, MOSFET_NMOS or MOSFET_PMOS. The Level is used to determine which model is placed and what value is set for Idsmod .
In the ADS netlist, the model is always called MOSFET, with the appropriate keywords NMOS and PMOS set to [0|1], and the parameter Idsmod set as specified in SPICE Level Parameters Mapping Table (MOSFET).

The only exception to this is the Mosfet device which refers to an HSpice Level 50 model, the Phillips MOS9 model. In this case, the device placed in the schematic will be MM9_NMOS or MM9_PMOS and the model will be MOS_Model9_Process. The netlist component is called MOS9. For both the MM9_NMOS and the MM9_PMOS, the translator sets the parameter Type=2 to indicate that it is a process-based model.

Dependence Parameters

There are certain model parameters listed in the BSIM Models that reference additional parameters. These additional parameters are denoted in parenthesis using the letters L, W and/or P. As an example, the Dwg (L,W, P) parameter in the BSIM3 Model defines four separate parameters;

1. Dwg: Coefficient of Weff's gate dependence.
2. LDwg: Length dependence of Dwg.
3. WDwg: Width dependence of Dwg.
4. PDwg: Cross dependence of Dwg.

Each letter indicates a sensitivity parameter that exists in HSpice and ADS. These parameters are length(L), Width(W) and Cross(P). Refer to your HSpice and ADS component documentation for details.

HSpice Automatic Model Selection

Automatic Model Selection, also known as binning , is an HSpice feature that allows for the definition of a library of Mosfets over a range of lengths and widths. The following netlist fragment is an example of automatic model selection in HSpice. It is currently used only for Mosfets. Notice that the model name is a root name plus an extension, separated by a period (that is: NCHAN.2). Also note the special model parameters, LMIN, LMAX, WMIN and WMAX. They are used only for automatic model selection.
.OPTION WL
M1 1 2 3 4 NCHAN 10 2
M2 1 2 3 4 NCHAN 10 3
$$$$$$$ FOR CHANNEL LENGTH SELECTION
.MODEL NCHAN.2 NMOS LEVEL=2 VTO=2.0 UO=800 TOX=500
+NSUB=1E15
+ RD=10 RS=10 CAPOP=5
+ LMIN=1 LMAX=2.5 WMIN=2 WMAX=15
.MODEL NCHAN.3 NMOS LEVEL=2 VTO=2.2 UO=800 TOX=500
+NSUB=1E15
+ RD=10 RS=10 CAPOP=5
+ LMIN=2.5 LMAX=3.5 WMIN=2 WMAX=15
The ADS equivalent of this functionality is provided by a component called BinModel . Since ADS component names cannot include a period, the models are renamed nchanx2 and nchanx3. Then a BinModel component is placed, which lists the models by name, as well as listing the special length and width parameter ranges, as defined by LMIN, LMAX, WMIN and WMAX. The name of the BinModel component is the same as the root name of the models, in this case nchan. The ranges of all models with the same root name will be recorded on one BinModel component as shown in The ADS BinModel Component.

The ADS BinModel Component

Also note that the Mosfet devices that are placed in the schematic do not refer to the Mosfet models as they normally would. Instead they refer to the BinModel. During simulation, the length and width specified on the device are sent to the BinModel and checked against the ranges specified for each model, to determine which model to use. If the length and width on the device fall within the range of the length and width on more than one model, the first one that matches is used.

Spectre Models

All Mosfet devices in Spectre reference a model by its instance name. Each Mosfet model in Spectre has a name value pair type= [ n | p ].

The type value pair keyword is used to determine what device to place in the schematic, MOSFET_NMOS or MOSFET_PMOS. The Level is used to determine which model is placed and what value is set for Idsmod.

In the ADS netlist, the model is always called MOSFET, with the appropriate keywords NMOS and PMOS set to [0|1], and the parameter Idsmod set as specified in Spectre Level Parameters Mapping (MOSFET).

The only exception to this is the Mosfet device which refers to the Phillips MOS902 model. In this case, the device placed in the schematic will be MM9_NMOS or MM9_PMOS and the model will be MOS_Model9_Process. The netlist component is called MOS9. For both the MM9_NMOS and the MM9_PMOS, the translator sets the parameter Type=2 to indicate that it is a process-based model.

Dependence Parameters

There are certain model parameters listed in the BSIM Models that reference additional parameters. These additional parameters are denoted in parenthesis using the letters L, W and/or P. As an example, the Dwg (L,W, P) parameter in the BSIM3 Model defines four separate parameters;

1. Dwg: Coefficient of Weff's gate dependence.
2. LDwg: Length dependence of Dwg.
3. WDwg: Width dependence of Dwg.
4. PDwg: Cross dependence of Dwg.

Each letter indicates a sensitivity parameter that exists in Spectre and ADS. These parameters are length(L), Width(W) and Cross(P). Refer to your Spectre and ADS component documentation for details.

The following MOSFET models are translated in ADS:

• LEVEL1_Model:LEVEL 1 MOSFET Model

• LEVEL2_Model:LEVEL 2 MOSFET Model

• LEVEL3_Model:LEVEL 3 MOSFET Model

• BSIM1_Model:BSIM1 MOSFET Model

• BSIM2_Model:BSIM2 MOSFET Model

• BSIM3_Model:BSIM3 MOSFET Model

• BSIM4_Model:BSIM4 MOSFET Model

• BSIM3SOI_Model: BSIM3SOI_Model

• MOS_Model9_Process: Philips MOS Model 9 (Process Based)

Binning Process

In order for ADS to translate the binning process, each binning definition is translated as a separate model.

Binning Example

The following is an example Spectre binning statement for a bsim model:

model ModelName ModelType {
      1: <model parameters> lmin=2 lmax=4 wmin=1 wmax=2
      2: <model parameters> lmin=1 lmax=2 wmin=2 wmax=4
      3: <model parameters> lmin=2 lmax=4 wmin=4 wmax=6
   }

The Netlist Translator separates each binning reference into an ADS model. The example would result in three models with the names ModelName_1, ModelName_2, and ModelName_3. The models are then tied together by creating an ADS "BinModel" component and configuring the appropriate min/max values.