Forums

Faux MOSFET... Behavioral Modeling

Started by Jim Thompson January 31, 2014
For sometime now I have been trying to build a behavioral model of
logic gates that was consistent with real-world behavior when mixed
into an Analog environment.

(Existing logic models, such as those in PSpice and LTspice are state
machines and do not behave properly when used for Analog applications,
such as when you build a crystal oscillator around a 74HCU04.)

So I now invite you to visit the Device Models & Subcircuits page of
my website...

     <http://www.analog-innovations.com/>

and take a look at FauxMOSFET.pdf, where I have finally created a
purely mathematical behavioral model of a MOSFET... using, of course,
my favorite math function TANH :-)

With all derivatives continuous, it is suitable for modeling logic
elements that can be run in an Analog simulator environment.  

Watch this space for further announcements as I develop inverters,
gates and complex functions that will run on any old analog simulator.
		
                                        ...Jim Thompson
-- 
| James E.Thompson                                 |    mens     |
| Analog Innovations                               |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Jim Thompson <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:
> For sometime now I have been trying to build a behavioral model of > logic gates that was consistent with real-world behavior when mixed > into an Analog environment. > > (Existing logic models, such as those in PSpice and LTspice are state > machines and do not behave properly when used for Analog applications, > such as when you build a crystal oscillator around a 74HCU04.) > > So I now invite you to visit the Device Models & Subcircuits page of > my website... > > <http://www.analog-innovations.com/> > > and take a look at FauxMOSFET.pdf, where I have finally created a > purely mathematical behavioral model of a MOSFET... using, of course, > my favorite math function TANH :-) > > With all derivatives continuous, it is suitable for modeling logic > elements that can be run in an Analog simulator environment. > > Watch this space for further announcements as I develop inverters, > gates and complex functions that will run on any old analog simulator.
At first I was all "Hyperbolic tangent, WTF does that have to do with SPICE?" Then I saw a TANH plot and the light came on. :) FWIW this guy has an intro to Analog Behavioral Modeling. http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm -- __ __/ \ / \__/ \__/ Don Kuenz / \__ \__/ \ \__/
On Sat, 1 Feb 2014 05:39:18 +0000 (UTC), Don Kuenz
<garbage@crcomp.net> wrote:

>Jim Thompson <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote: >> For sometime now I have been trying to build a behavioral model of >> logic gates that was consistent with real-world behavior when mixed >> into an Analog environment. >> >> (Existing logic models, such as those in PSpice and LTspice are state >> machines and do not behave properly when used for Analog applications, >> such as when you build a crystal oscillator around a 74HCU04.) >> >> So I now invite you to visit the Device Models & Subcircuits page of >> my website... >> >> <http://www.analog-innovations.com/> >> >> and take a look at FauxMOSFET.pdf, where I have finally created a >> purely mathematical behavioral model of a MOSFET... using, of course, >> my favorite math function TANH :-) >> >> With all derivatives continuous, it is suitable for modeling logic >> elements that can be run in an Analog simulator environment. >> >> Watch this space for further announcements as I develop inverters, >> gates and complex functions that will run on any old analog simulator. > >At first I was all "Hyperbolic tangent, WTF does that have to do with >SPICE?" Then I saw a TANH plot and the light came on. :) FWIW this guy >has an intro to Analog Behavioral Modeling. > >http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm
Thanks, Don, that's a nice intro and a nice composite listing of the different functions available. I've been at this long enough that I know that list by heart ;-) If your simulator supports VALUE, I'd avoid the POLY form altogether... it's not only obtuse, but difficult to keep your head and the coefficient order on straight for any order above 2 ;-) I settled on TANH while pursuing how to handle (simulation device) current sources which would pull voltages to infinity if unrestrained. My devices... <http://www.analog-innovations.com/CompliantParts.zip> will not pull below zero volts, thus not only behave "real world", but have no convergence problems (limiting implemented with a TANH function). ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
On Sat, 01 Feb 2014 08:34:34 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:

>On Sat, 1 Feb 2014 05:39:18 +0000 (UTC), Don Kuenz ><garbage@crcomp.net> wrote: > >>Jim Thompson <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote: >>> For sometime now I have been trying to build a behavioral model of >>> logic gates that was consistent with real-world behavior when mixed >>> into an Analog environment. >>> >>> (Existing logic models, such as those in PSpice and LTspice are state >>> machines and do not behave properly when used for Analog applications, >>> such as when you build a crystal oscillator around a 74HCU04.) >>> >>> So I now invite you to visit the Device Models & Subcircuits page of >>> my website... >>> >>> <http://www.analog-innovations.com/> >>> >>> and take a look at FauxMOSFET.pdf, where I have finally created a >>> purely mathematical behavioral model of a MOSFET... using, of course, >>> my favorite math function TANH :-) >>> >>> With all derivatives continuous, it is suitable for modeling logic >>> elements that can be run in an Analog simulator environment. >>> >>> Watch this space for further announcements as I develop inverters, >>> gates and complex functions that will run on any old analog simulator. >> >>At first I was all "Hyperbolic tangent, WTF does that have to do with >>SPICE?" Then I saw a TANH plot and the light came on. :) FWIW this guy >>has an intro to Analog Behavioral Modeling. >> >>http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm > >Thanks, Don, that's a nice intro and a nice composite listing of the >different functions available. I've been at this long enough that I >know that list by heart ;-) > >If your simulator supports VALUE, I'd avoid the POLY form >altogether... it's not only obtuse, but difficult to keep your head >and the coefficient order on straight for any order above 2 ;-) > >I settled on TANH while pursuing how to handle (simulation device) >current sources which would pull voltages to infinity if unrestrained. >My devices... > > <http://www.analog-innovations.com/CompliantParts.zip> > >will not pull below zero volts, thus not only behave "real world", but >have no convergence problems (limiting implemented with a TANH >function). > > ...Jim Thompson
I forgot to mention one thing that may be of interest. The transfer function of your everyday bipolar differential pair is TANH. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
On 2/1/14 8:04 AM, Jim Thompson wrote:
> On Sat, 01 Feb 2014 08:34:34 -0700, Jim Thompson > <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: > >> On Sat, 1 Feb 2014 05:39:18 +0000 (UTC), Don Kuenz >> <garbage@crcomp.net> wrote: >> >>> Jim Thompson <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote: >>>> For sometime now I have been trying to build a behavioral model of >>>> logic gates that was consistent with real-world behavior when mixed >>>> into an Analog environment. >>>> >>>> (Existing logic models, such as those in PSpice and LTspice are state >>>> machines and do not behave properly when used for Analog applications, >>>> such as when you build a crystal oscillator around a 74HCU04.) >>>> >>>> So I now invite you to visit the Device Models & Subcircuits page of >>>> my website... >>>> >>>> <http://www.analog-innovations.com/> >>>> >>>> and take a look at FauxMOSFET.pdf, where I have finally created a >>>> purely mathematical behavioral model of a MOSFET... using, of course, >>>> my favorite math function TANH :-) >>>> >>>> With all derivatives continuous, it is suitable for modeling logic >>>> elements that can be run in an Analog simulator environment. >>>> >>>> Watch this space for further announcements as I develop inverters, >>>> gates and complex functions that will run on any old analog simulator. >>> >>> At first I was all "Hyperbolic tangent, WTF does that have to do with >>> SPICE?" Then I saw a TANH plot and the light came on. :) FWIW this guy >>> has an intro to Analog Behavioral Modeling. >>> >>> http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm >> >> Thanks, Don, that's a nice intro and a nice composite listing of the >> different functions available. I've been at this long enough that I >> know that list by heart ;-) >> >> If your simulator supports VALUE, I'd avoid the POLY form >> altogether... it's not only obtuse, but difficult to keep your head >> and the coefficient order on straight for any order above 2 ;-) >> >> I settled on TANH while pursuing how to handle (simulation device) >> current sources which would pull voltages to infinity if unrestrained. >> My devices... >> >> <http://www.analog-innovations.com/CompliantParts.zip> >> >> will not pull below zero volts, thus not only behave "real world", but >> have no convergence problems (limiting implemented with a TANH >> function). >> >> ...Jim Thompson > > I forgot to mention one thing that may be of interest. The transfer > function of your everyday bipolar differential pair is TANH.
That's interesting. I might have thought arctan since the curves look similar, but I haven't done that math.
On Sat, 01 Feb 2014 11:37:58 -0800, Daniel Pitts
<newsgroup.nospam@virtualinfinity.net> wrote:

>On 2/1/14 8:04 AM, Jim Thompson wrote: >> On Sat, 01 Feb 2014 08:34:34 -0700, Jim Thompson >> <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: >> >>> On Sat, 1 Feb 2014 05:39:18 +0000 (UTC), Don Kuenz >>> <garbage@crcomp.net> wrote: >>> >>>> Jim Thompson <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote: >>>>> For sometime now I have been trying to build a behavioral model of >>>>> logic gates that was consistent with real-world behavior when mixed >>>>> into an Analog environment. >>>>> >>>>> (Existing logic models, such as those in PSpice and LTspice are state >>>>> machines and do not behave properly when used for Analog applications, >>>>> such as when you build a crystal oscillator around a 74HCU04.) >>>>> >>>>> So I now invite you to visit the Device Models & Subcircuits page of >>>>> my website... >>>>> >>>>> <http://www.analog-innovations.com/> >>>>> >>>>> and take a look at FauxMOSFET.pdf, where I have finally created a >>>>> purely mathematical behavioral model of a MOSFET... using, of course, >>>>> my favorite math function TANH :-) >>>>> >>>>> With all derivatives continuous, it is suitable for modeling logic >>>>> elements that can be run in an Analog simulator environment. >>>>> >>>>> Watch this space for further announcements as I develop inverters, >>>>> gates and complex functions that will run on any old analog simulator. >>>> >>>> At first I was all "Hyperbolic tangent, WTF does that have to do with >>>> SPICE?" Then I saw a TANH plot and the light came on. :) FWIW this guy >>>> has an intro to Analog Behavioral Modeling. >>>> >>>> http://www.ecircuitcenter.com/Circuits/ABM1/ABM1.htm >>> >>> Thanks, Don, that's a nice intro and a nice composite listing of the >>> different functions available. I've been at this long enough that I >>> know that list by heart ;-) >>> >>> If your simulator supports VALUE, I'd avoid the POLY form >>> altogether... it's not only obtuse, but difficult to keep your head >>> and the coefficient order on straight for any order above 2 ;-) >>> >>> I settled on TANH while pursuing how to handle (simulation device) >>> current sources which would pull voltages to infinity if unrestrained. >>> My devices... >>> >>> <http://www.analog-innovations.com/CompliantParts.zip> >>> >>> will not pull below zero volts, thus not only behave "real world", but >>> have no convergence problems (limiting implemented with a TANH >>> function). >>> >>> ...Jim Thompson >> >> I forgot to mention one thing that may be of interest. The transfer >> function of your everyday bipolar differential pair is TANH. > >That's interesting. I might have thought arctan since the curves look >similar, but I haven't done that math.
See... <http://www.analog-innovations.com/DiffPairMath-JT.pdf> One of my very first patents (filed 1965)... <http://www.analog-innovations.com/SED/Pat-3491301.pdf> takes advantage of that transfer function to accomplish harmonic mixing. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.