Code history

Model history

Since the updates for SiMKit 3.1 (May 2008) the code history is included in the model history. Older code history can be found below.

MOS Model 1102

Below is an overview of the relation between level and code versions.
The background of the compact model identification scheme can be read here.

MOS1102 SiMKit
level.version.implementation code revision release date
1102.7.1 1.279 4.0 2012-07-25
1102.7.0 1.229 3.8 2012-02-02
1102.7.0 1.224.2.1 3.7 2011-08-04
1102.7.0 1.213.2.2 3.6 2011-01-18
1102.6.2 1.199 3.5.2 2010-10-12
1102.6.1 1.180 3.4 2009-12-17
1102.6.0 1.178 3.3 2009-06-25
1102.5.4 1.169.2 3.2 2008-12-02
1102.5.3 1.166.2.1 3.1.2 2008-07-24
1102.5.2 1.165 3.1 2008-05-21
1102.5.1 1.156.2.2 3.0.3 2008-01-31
1102.5.0 1.142 2.5 2007-04-30
1102.4.0 1.131 2.4 2006-10-23
1102.3.0 1.114 2.3 2006-03-16
1102.2.0 1.97 2.2 2005-08-03
1102.1.0 1.91 2.1.1 2005-04-11
1102.0.2 1.71 2.0.1 2004-11-17
1102.0.1 1.55 1.3 2004-04-20
2012/07 Release of MOS MODEL 11, 1102.7.1
 
  • The interface of the model has been changed.
2011/01 Release of MOS MODEL 11, 1102.7.0
 
  • The scaling rule for KO has been updated.
  • A new instance parameter PRINTSCALED has been added to this model. When PRINTSCALED is set to any value other than zero (the default) the model will print the scaled parameters of the model when the DC operating point information is printed.
  • For Dcmatch calculations the parameter BET has been added to the OP output.
2010/10 Release of MOS MODEL 11, 1102.6.2
 
  • Mos Model 1102 is implemented in a thread-safe way
2009/12 Release of MOS MODEL 11, 1102.6.1
 
  • Minor change in the induced gate noise contributions: the contribution has been split up in two components to be able to manage source-drain interchange properly.
2009/06 Release of MOS MODEL 11, 1102.6.0
 
  • A gate resistance RG has been added to MM1102. This is done in a flexible way, similar to the gate resistance in PSP102 and PSP103, so if the gate resistance equals zero there is no extra node in the model.
  • The model has been extended with multiple finger support.
  • The gate resistance RG is added to the OP output of the model.

The default values of the new parameters are such that backward compatibililty is guaranteed.
2008/12 Release of MOS MODEL 11, 1102.5.4
 
  • minor code changes
2008/07 Release of MOS MODEL 11, 1102.5.3
 
  • minor code changes
2008/05 Release of MOS MODEL 11, 1102.5.2
 
  • For a certain set of parameters the initialization of the Newton process forthe surface potential could cause problems leading to floating point exceptions.These problems have been solved.
2007/04 Release of MOS MODEL 11, 1102.5.0
 
  • Induced gate noise limiting (i.e. the cut-off of the f² dependency of the induced gate noise at high frequency) has been improved in Spectre and ADS.
2006/10 Release of MOS MODEL 11, 1102.4.0
 
  • A bug in the parameter clipping for (short-channel) subthreshold slope for the actual transistor (MO) has been corrected in mos1102e. The MO high clip value is 0.5.
  • A problem occurred in the model equation for the saturation voltage VDSATshort (formula 14.233 in the modelbook). The combination of parameters and voltages led to a negative argument of a square root function. The equation has been rephrased to avoid the problem.
  • Some minor performance improvements.
2006/03 Release of MOS MODEL 11, 1102.3.0
  Changes in equations:
 
  • The units of the M1102 parameters pocgso, plcgso, pwcgso and plwcgso have been corrected to Farad- note that this has no influence on the calculations, only on the text output.
  • Induced gate noise limiting has been implemented.
  • Coulomb scattering has been implemented. The default values of the parameters are such that backwards compatibility is guaranteed.
2005/08 Release of MOS MODEL 11, 1102.2.0
  Changes in equations:
 
  • The calculation of the power dissipation Pdiss has been corrected in case of a source drain exchange (inverse bias operation).
2005/04 Release of MOS MODEL 11, 1102.1.0
  Changes in equations:
 
  • Default value of parameter NT has been corrected so that it corresponds to the default value of the reference temperature TR.
  • Implementation of noise equation BG corrected.
2004/10 Release of MOS MODEL 11, 1102.0.0
 
  • Addition of temperature dependence of the thermal resistance, Ath.
  • Simpler equation for variable VDBt is used, which results in a more physical description for VDS<0.
  • Definition of starting condition of surface potentials Ψs and Ψov has been slightly rewritten to increase numerical efficiency.
2004/02 Release of MOS MODEL 11, 1102.0, test version
  Differences w.r.t. Level 1101:
 
  • Iterative solution of surface potential.
  • More accurate, physics-based implementation of velocity saturation effect.
  • More accurate, physics-based equations for thermal noise, induced gate noise and their correlation.

MOS Model 1101

Below is an overview of the relation between level and code versions.
The background of the compact model identification scheme can be read here.

MOS1101 SiMKit
level.version.implementation code revision release date
1101.11.1 1.348 4.0 2012-07-25
1101.11.0 1.295 3.8 2012-02-02
1101.11.0 1.289.2.1 3.7 2011-08-04
1101.11.0 1.277.2.1 3.6 2011-01-18
1101.10.1 1.264 3.5.2 2010-10-12
1101.10.0 1.221 3.4 2009-12-17
1101.9.4 1.218 3.3 2009-06-25
1101.9.3 1.217.2 3.2 2008-12-02
1101.9.2 1.213 3.1.2 2008-07-24
1101.9.2 1.213 3.1 2008-05-21
1101.9.1 1.202.2.2 3.0.3 2008-01-31
1101.8.1 1.179 2.4 2006-10-23
1101.8.0 1.162 2.3 2006-03-16
1101.7.0 1.146 2.2 2005-08-03
1101.6.0 1.138 2.1.1 2005-04-11
1101.5.0 1.115 2.0.1 2004-11-17
1101.4.2 1.106 1.3 2004-04-20
1101.4.1 1.90 1.2.1 2004-02-20
1101.2.0 1.65 1.1.1 2003-10-27
1101.1.1 1.37 1.0 2003-04-30
1101.1.0 1.15 SMK/PMK 2002-09-25
2012/07 Release of MOS MODEL 11, 1101.11.1
 
  • The interface of the model has been changed.
2011/01 Release of MOS MODEL 11, 1101.11.0
 
  • The scaling rule for KO has been updated.
  • A new instance parameter PRINTSCALED has been added to this model. When PRINTSCALED is set to any value other than zero (the default) the model will print the scaled parameters of the model when the DC operating point information is printed.
2010/10 Release of MOS MODEL 11, 1101.10.1
 
  • Mos Model 1101 is implemented in a thread-safe way.
2009/12 Release of MOS MODEL 11, 1101.10.0
 
  • A gate resistance RG has been added to MM1101. This is done in a flexible way, similar to the gate resistance in MM1102, so if the gate resistance equals zero there is no extra node in the model. Furthermore the model has been extended with multiple finger support.
  Model parameters for these extensions are:
  For the geometrical and binning scaled models
 
  • RGO - gate resistance (Ohm)- default value is 0.0, no clip values;
  • RINT - contact resistance between silicide and poly (Ohm.m^2) - default value is 0.0, lower clip value is 0.0;
  • RVPOLY - vertical poly resistance (Ohm.m^2) - default value is 0.0, no clip values;
  • RSHG - gate electrode diffusion sheet resistance (Ohm/sheet) - default value is 0.0, lower clip value is 0.0;
  • DLSIL - silicide extension over the physical gate length (m) - default value is 0.0, no clip values;
  Instance parameters for these extensions for the geometrical and binning models:
 
  • NF - number of fingers (-) - default value is 1 - lower clip value is 1;
  • NGCON - number of gate contacts (-) - default value is 1 - lower clip value is 1, high clip value is 2;
  • XGW - distance from the gate contact to the channel edge (m) - default value is 1e-7, no clip values;
2009/06 Release of MOS MODEL 11, 1101.9.4
 
  • minor code changes
2008/12 Release of MOS MODEL 11, 1101.9.3
 
  • minor code changes
2008/05 Release of MOS MODEL 11, 1101.9.2
 
  • minor code changes
2006/03 Release of MOS MODEL 11, 1101.8.0
  Changes in equations:
 
  • The units of the M1101 parameters pocgso, plcgso, pwcgso and plwcgso have been corrected to Farad- note that this has no influence on the calculations, only on the text output.
  • One of the functions in the model has been made more numerically robust, thus solving floating-point exceptions on HP.
2005/08 Release of MOS MODEL 11, 1101.7.0
  Changes in equations:
 
  • The calculation of the power dissipation Pdiss has been corrected in case of a source drain exchange (inverse bias operation).
2005/04 Release of MOS MODEL 11, 1101.6.0
  Changes in equations:
 
  • Default value of parameter NT has been corrected so that it corresponds to the default value of the reference temperature TR.
2004/10 Release of MOS MODEL 11, 1101.5.0
 
  • Addition of temperature dependence of the thermal resistance, Ath.
2004/02 Release of MOS MODEL 11, 1101.4.0
  Changes in equations:
 
  • The expression for thermal noise density Sth can give erroneous results in deep subthreshold (near the accumulation region). This problem has been solved in the implemented equations.
2003/12 Release of MOS MODEL 11, 1101.3.0
  Self-heating has been included in the model.
  Changes in equations:
 
  • Bug fix for thermal noise density Sth.
2003/06 Release of MOS MODEL 11, 1101.2.0
  Changes in equations:
 
  • The 1/f -noise spectral density Sfl is inversely proportional to Gvsat instead of Gmob.
2002/12 Release of MOS MODEL 11, 1101.1.0
  Changes in equations:
 
  • Bug fix for overlap capacitance when VDS<0V (only important when CGDO is not equal to CGSO, e.g., in the case of a DMOS-transistor)
  • Gate-induced drain leakage (GIDL) implemented
2002/06 Release of MOS MODEL 11, 1101.0.0
  Small fixes:
 
  • Improved description of gate charge QG and bulk charge QB
  • Thermal noise spectral density Sth is larger than or equal to zero
  • Improved physical geometry scaling rule of parameter m0
  • Upper clipping limit of physical geometry scaling parameter Lmin changed to 2.5x10-6m
2001/12 Release of MOS MODEL 11, 1101.0, Test Version
  Differences w.r.t. Level 1100:
 
  • Two types of geometrical scaling rules available: physical rules and binning rules
  • Physical geometry scaling rules are slightly different
  • Temperature scaling implemented on the "miniset" level instead of the "maxiset" level

MOS Model 1100

Below is an overview of the relation between level and code versions.
The background of the compact model identification scheme can be read here.

MOS1100 SiMKit
level.version.implementation code revision release date
1100.3.14 1.190 3.8 2012-02-02
1100.3.14 1.188.2.1 3.7 2011-08-04
1100.3.14 1.179.2.1 3.6 2011-01-18
1100.3.13 1.173 3.5.2 2010-10-12
1100.3.12 1.163 3.4 2009-12-17
1100.3.11 1.162 3.3 2009-06-25
1100.3.10 1.161.2 3.2 2008-12-02
1100.3.9 1.159 3.1.2 2008-07-24
1100.3.9 1.159 3.1 2008-05-21
1100.3.8 1.148.22 3.0.3 2008-01-31
1100.3.7 1.132 2.5 2007-04-30
1100.3.6 1.123 2.4 2006-10-23
1100.3.5 1.114 2.3 2006-03-16
1100.3.4 1.104 2.2 2005-08-03
1100.3.3 1.101 2.1.1 2005-04-11
1100.3.2 1.83 2.0.1 2004-11-17
1100.3.1 1.78 1.3 2004-04-20
1100.3.0 1.70 1.2.1 2004-02-20
1100.2.1 1.50 1.1.1 2003-10-27
1100.1.1 1.18 1.0 2003-04-30
2011/01 Release of MOS MODEL 11, 1100.3.14
 
  • A new instance parameter PRINTSCALED has been added to this model. When PRINTSCALED is set to any value other than zero (the default) the model will print the scaled parameters of the model when the DC operating point information is printed.
2010/10 Release of MOS MODEL 11, 1100.3.13
 
  • Several minor code changes related to the noise implementation
2009/12 Release of MOS MODEL 11, 1100.3.12
 
  • minor code changes
2009/06 Release of MOS MODEL 11, 1100.3.11
 
  • minor code changes
2008/12 Release of MOS MODEL 11, 1100.3.10
 
  • minor code changes
2008/05 Release of MOS MODEL 11, 1100.3.8
 
  • minor code changes
2004/02 Release of MOS MODEL 11, 1100.3.0
  Changes in equations:
 
  • The expression for thermal noise density Sth can give erroneous results in deep subthreshold (near the accumulation region). This problem has been solved in the implemented equations..
  • Small change in description of flicker noise.
2003/12 Release of MOS MODEL 11, 1100.2.0
  Changes in equations:
 
  • The expression for thermal noise density Sth can give erroneous results in deep subthreshold (near the accumulation region). This problem was solved in this version, however not satisfactorily.
 
  • Thermal noise density Sth should be larger than or equal to zero for all possible parameter values.
2002/12 Release of MOS MODEL 11, 1100.1.0
  Changes in equations:
 
  • Bug fix for overlap capacitance when VDS<0V (only important when CGDO is not equal to CGSO, e.g., in the case of a DMOS-transistor)
2001/12 Release of MOS MODEL 11, 1100.0.0
  Small fixes:
 
  • Minor change in internal parameter Vlimit
2001/06 Release of MOS MODEL 11, 1100.0, Test Version

Code history

MOS Model 1102

The table below outlines the major changes to the file called "device_m1102i.c", which can be found in the source code.

2008/01/31 MOS Model 1102.5.1.
  Code changes
 
  • New flexible topology architecture implemented
  • Small improvement in noise implementation for non-correlated noise sources
  • Clipping added of internal parameter ETAMOB to 1e-12
  • Spectre only: DCmatch implementation for test purposes
  • Spectre only: model group selection feature is supported by the binning model
2007/04/30 MOS model 1102.5.0 identified in the code as revision 1.142. All code differences with the previous version (1102.4.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.5 in the model history above.
  Other code changes
 
  • In 64bit mode a floating point exception could occur in the calculation of the initial values for the iterative process to calculate the surface potential. By re-arranging the calculations, this is now avoided.
2006/10/23 MOS model 1102.4.0 identified in the code as revision 1.131. All code differences wit the previous version (1102.3.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.4 in the model history above
  • A gmin convergence aid has been implemented. This makes the convergence more robust. The results are not always only positive but also in some cases the convergence may be slower. The gmin convergence aid implementation has been made common to all simulators.
  • The performance of self-heating models has been improved by reducing the time spent in clipping.
  Minor code changes
 
  • There was a warning regarding specifying the working region when using spectre -h <model name>. This warning is removed.
  • In the past an unclear warning was given when the effective length and/or effective width of a transistor were clipped. This warning has been made more specific.
2006/03/16 MOS Model 1102.3.0 identified in the code as revision 1.114. All code differences with the previous version (1102.2.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.3 in the model history above
  Other code changes
  None
2005/08/03 MOS Model 1102.2 identified in the code as revision 1.97. All code differences with the previous version (1102.1.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.2 in the model history above
  Other code changes
  None
2005/04/11 MOS Model 1102.1.0 identified in the code as revision 1.91. All code differences with the previous version (1102.0.2) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.1 in the model history above
  Other code changes
 
  • Avoided possible numerical problems in function 'func_Vov
  • Removed Weff and Leff from the electrical parameter set.
  • Implementation of temperature limiting for self-heating
2004/11/17 MOS Model 1102.0.2 identified in the code as revision 1.71. All code differences with the previous version (1102.0.1) can be found in this diff file.
  Implemented model changes
 
  • See release 1102.0 in the model history above
  Other code changes
 
  • A floating-point exception has been solved by limiting the minimum temperature.
  • Implementation of variable VDSx has been rewritten to increase numerical efficiency.
  • Definition of starting condition of surface potential Ψov has been changed in order to prevent possible floating-point exception.
2004/04/20 MOS Model 1102.0.1 identified in the code as revision 1.55 All code differences with the previous version (1102.0.2) can be found in this diff file.
  Implemented model changes
 
  • Addition of self-heating.
  Other code changes
 
  • Bug fix for calculation of idb
  • Default value of ST;VFB corrected for electrical model
2004/02/15 Release of MOS Model 11, level 1102, test version

MOS Model 1101

The table below outlines the major changes to the file called "device_m1101i.c", which can be found in the source code.

2008/01/31 MOS Model 1101.9.1.
  Code changes
 
  • New flexible topology architecture implemented
  • Small improvement in noise implementation for non-correlated noise sources
  • Clipping added of internal parameter ETAMOB to 1e-12
  • Spectre only: model group selection feature is supported by the binning model
2007/04/30 MOS model 1101.9.0 identified in the code as revision 1.189. All code differences with the previous version (1108.1) can be found in this diff file.
  Implemented model changes
 
  • See release 1101.9 in the model history above.
  Other code changes
  None
2006/10/23 MOS model 1101.8.1 identified in the code as revision 1.179. All code differences wit the previous version (1101.8.0) can be found in this diff file.
  Implemented model changes
 
  • The description for the following parameters has been corrected. This description can be seen with spectre -h <modelname>. The parameters are unitless: swthesr, swtheph, swetamob, swther, swthesat, swtheth, slssf, swssf, swalp, sla1, swa1, sla2, swa2, sla3, swa3.
  • A gmin convergence aid has been implemented. This makes the convergence more robust. The results are not always only positive but also in some cases the convergence may be slower. The gmin convergence aid implementation has been made common to all simulators.
  • The performance of self-heating models has been improved by reducing the time spent in clipping.
  Other code changes
  None
  Minor code changes
 
  • There was a warning regarding specifying the working region when using spectre -h <model name>. This warning is removed.
  • In the past an unclear warning was given when the effective length and/or effective width of a transistor were clipped. This warning has been made more specific.
2006/03/16 MOS Model 1101.8.0 identified in the code as revision 1.162. All code differences with the previous version (1101.7.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1101.8 in the model history above
  Other code changes
  None
2005/08/03 MOS Model 1101.7.0 identified in the code as revision 1.146. All code differences with the previous version (1102.6.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1101.7 in the model history above
  Other code changes
  None
2005/04/11 MOS Model 1101.6.0 identified in the code as revision 1.138. All code differences with the previous version (1101.5.0) can be found in this diff file.
  Implemented model changes
 
  • See release 1101.6 in the model history above
  Other code changes
 
  • Removed Weff and Leff from the electrical parameter set.
  • Implementation of temperature limiting for self-heating
2004/11/17 MOS Model 1101.5.0 identified in the code as revision 1.115. All code differences with the previous version (1101.4.2) can be found in this diff file.
  Implemented model changes
 
  • Addition of temperature dependence of the thermal resistance, Ath.
  Other code changes
 
  • None
2004/4/20 MOS Model 1101.4.2 identified in the code as revision 1.106 All code differences with the previous version (1101.4.1) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
 
  • Speed improvements
  • Bug fix for calculation of idb
  • Default value of ST;VFB corrected for electrical model
  • Bug fix for division by zero if h mob=0
  • Bug fix in derivatives of Pdiss solved
  • Mult scaling added for self heating components RTH and CTH
  • Bug fix in Qov_L implementation and self heating conductance's
2004/02/20 MOS Model 1101.4.1 identified in the code as revision 1.90 All code differences with the previous version (1101.2.0) can be found in this diff file.
  Implemented model changes
 
  • The expression for thermal noise density Sth can give erroneous results in deep subthreshold (near the accumulation region). This problem has been solved in the implemented equations.
  • Improved description of the thermal noise density.
  Other code changes
 
  • None
2003/10/27 MOS Model 1101.2.0 identified in the code as revision 1.65 All code differences with the previous version (1101.1.1) can be found in this diff file
  Implemented model changes
 
  • Addition of self-heating.
  • Improved description of the thermal noise density.
  Other code changes
 
  • Addition of a high frequency series expansion for sig and sigth.
  • Bug fix for derivative of Igb with respect to Vs.
  • Correct model name in the copyright message.
  • Removed double clip warning.
2008/01/31 MOS Model 1100.3.8.
  Code changes
 
  • New flexible topology architecture implemented
  • Small improvement in noise implementation for non-correlated noise sources
  • Clipping added of internal parameter ETAMOB to 1e-12
2003/04/30 MOS Model 1101.1.1, identified in the code as revision 1.37
  The code differences will not be given in a diff file, because a complete new set up is made since the last version on the web
  Implemented model changes
 
  • See release 1101.1 in the model history above
  Other code changes
 
  • Solved problems on the noise model for Sig and Sigth when gm=0
2002/09/16 MOS Model 1101.1.0, identified in the code as revision 1.15
  All code differences with the previous version (1101.0.0) can be found in this diff file
  Implemented model changes
 
  • See release 1101.1 in the model history above
  Other code changes
 
  • none

MOS Model 1100

The table below outlines the major changes to the file called "device_m1100i.c", which can be found in the source code.

2007/04/30 MOS model 1100.3.7 identified in the code as revision 1.132. All code differences with the previous version (1100.3.6) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
 
  • None
2006/10/23 MOS model 1100.3.6 identified in the code as revision 1.123. All code differences wit the previous version (1100.3.5) can be found in this diff file.
  Implemented model changes
 
  • None
  Minor code changes
 
  • There was a warning regarding specifying the working region when using spectre -h <model name>. This warning is removed.
  • In the past an unclear warning was given when the effective length and/or effective width of a transistor were clipped. This warning has been made more specific.
2006/03/16 MOS Model 1100.3.5 identified in the code as revision 1.114. All code differences with the previous version (1100.3.3) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
  Minor code changes
2005/08/03 MOS Model 1100.3.4 identified in the code as revision 1.104. All code differences with the previous version (1100.3.3) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
  Minor code changes
2005/04/11 MOS Model 1100.3.3 identified in the code as revision 1.101. All code differences with the previous version (1100.3.2) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
 
  • Removed Weff and Leff from the electrical parameter set.
2004/11/17 MOS Model 1100.3.2 identified in the code as revision 1.83. All code differences with the previous version (1100.3.1) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
 
  • Minor code changes.
2004/04/20 MOS Model 1100.3.1 identified in the code as revision 1.78 All code differences with the previous version (1100.3.0) can be found in this diff file.
  Implemented model changes
 
  • None
  Other code changes
 
  • Speed improvements
  • Default value of ST;VFB corrected for electrical model
  • Bug fix for division by zero if η mob=0
2004/02/2 MOS Model 1100.3.0 identified in the code as revision 1.70 All code differences with the previous version (1100.2.1) can be found in this diff file.
  Implemented model changes
 
  • The expression for thermal noise density Sth can give erroneous results in deep subthreshold (near the accumulation region). This problem has been solved in the implemented equations.
  • Small change in description of flicker noise.
  Other code changes
 
  • None
2003/10/27 MOS Model 1100.2.1 identified in the code as revision 1.50 All code differences with the previous version (1100.1.1) can be found in this diff file
  Implemented model changes
 
  • Improved description of the thermal noise density.
  Other code changes
 
  • Addition of a high frequency series expansion for sig and sigth.
  • Bug fix for derivative of Igb with respect to Vs.
  • Correct model name in the copyright message.
  • Removed double clip warning.
  • Changed Gvsat to Gmob in calculation of noise density Sfl
2003/04/30 MOS Model 1100.1.1, identified in the code as revision 1.18
  All code differences with the previous version (1100.1.0) can be found in this diff file
  Implemented model changes
 
  • See release 1100.1 in the model history above
  Other code changes
 
  • none

Known bugs

There are no known bugs in MOS Model 11.

How to search?

Already registered to MyNXP? or Register

Feedback