29.08.2009, 02:33 PM
Fehlinterpretation !
Die alte Version läuft mit Krücken.
Damped Pseudo ist er dazu gekommen. Gmin stepping macht teilweise Problem weil er in zu kleine Zeitbereiche abdriftet, da gehen paar Interationen nicht auf.
Hintergrund ist wohl die genauere Umsetzung von Level 3 Mosfet Modellen, welche teilweise die Simulationen ins Nirvana schickt.
Und wenn man mal in die Changelog.txt schaut ... die Fehler werden weniger
08/17/09 Corrected an error cross-probing some busses through hierarchy.
08/12/09 Improved parsing of device currents in .four statements.
08/10/09 Fixed some numerical problems in the ferromagnetics implementation.
08/01/09 Revised the Yang-Chatterjee charge model to work over absurd user parameters.
07/22/09 Corrected a problem regarding level 2 MOSFET gamma.
07/16/09 Trapped a crash mode regarding thread timing.
07/15/09 Added warnings when using archaic MOS levels 1, 2, or 3 below reasonable dimensions.
07/15/09 Re-implemented MOS levels 1, 2, and 3 to use the Yang-Chatterjee charge model instead of the Meyer capacitance model.
07/15/09 Improved the accuracy of the computation of the Jacobian for MOS level 3.
07/14/09 Improved the implementation of AC current monitoring in levels 2 and 3 MOS devices.
07/14/09 Corrected a DC current monitoring error in the level 3 MOS device.
07/07/09 Improved BUS netlisting support in hierarchical designs.
07/07/09 Fixed a problem keeping window tab labels in sync with symbol file names.
06/30/09 Trapped a crash mode that could occur when sync'ing with multiple invocations of LTspice running.
06/23/09 Node "GND" is more uniformly accepted as a synonym for "0"
06/23/09 Fixed a problem in the .lib/.lib editor.
06/12/09 Modified the interpretation of .noise data in some .meas operations.
06/09/09 Fixed a problem in the waveform expression compilers.
06/04/09 Trapped a potential crash mode.
06/03/09 Improved the handling of the BSIM4 rgeomod parameter.
Die letzten zwe-drei Monate und ich hab die Banalitäten und LT Modellanpassungen raus genommen.
Alternativ mal mit der original PSPice Engine (OrCad) arbeiten ... da läuft aus dem Stand kein Schaltnetzteil.
Aber gut ...
Die alte Version läuft mit Krücken.
Damped Pseudo ist er dazu gekommen. Gmin stepping macht teilweise Problem weil er in zu kleine Zeitbereiche abdriftet, da gehen paar Interationen nicht auf.
Hintergrund ist wohl die genauere Umsetzung von Level 3 Mosfet Modellen, welche teilweise die Simulationen ins Nirvana schickt.
Und wenn man mal in die Changelog.txt schaut ... die Fehler werden weniger
08/17/09 Corrected an error cross-probing some busses through hierarchy.
08/12/09 Improved parsing of device currents in .four statements.
08/10/09 Fixed some numerical problems in the ferromagnetics implementation.
08/01/09 Revised the Yang-Chatterjee charge model to work over absurd user parameters.
07/22/09 Corrected a problem regarding level 2 MOSFET gamma.
07/16/09 Trapped a crash mode regarding thread timing.
07/15/09 Added warnings when using archaic MOS levels 1, 2, or 3 below reasonable dimensions.
07/15/09 Re-implemented MOS levels 1, 2, and 3 to use the Yang-Chatterjee charge model instead of the Meyer capacitance model.
07/15/09 Improved the accuracy of the computation of the Jacobian for MOS level 3.
07/14/09 Improved the implementation of AC current monitoring in levels 2 and 3 MOS devices.
07/14/09 Corrected a DC current monitoring error in the level 3 MOS device.
07/07/09 Improved BUS netlisting support in hierarchical designs.
07/07/09 Fixed a problem keeping window tab labels in sync with symbol file names.
06/30/09 Trapped a crash mode that could occur when sync'ing with multiple invocations of LTspice running.
06/23/09 Node "GND" is more uniformly accepted as a synonym for "0"
06/23/09 Fixed a problem in the .lib/.lib editor.
06/12/09 Modified the interpretation of .noise data in some .meas operations.
06/09/09 Fixed a problem in the waveform expression compilers.
06/04/09 Trapped a potential crash mode.
06/03/09 Improved the handling of the BSIM4 rgeomod parameter.
Die letzten zwe-drei Monate und ich hab die Banalitäten und LT Modellanpassungen raus genommen.
Alternativ mal mit der original PSPice Engine (OrCad) arbeiten ... da läuft aus dem Stand kein Schaltnetzteil.
Aber gut ...
"Ich hab Millionen von Ideen und alle enden mit Sicherheit tödlich."