Forums

Explanation of SPICE Parameters

Started by Tim Williams September 25, 2017
I can't find any useful references for MOSFET SPICE parameters.  Does this 
information exist entirely in book form, or something?

Example: websites copypasting the SPICE 3f5 handbook.  This says 
more-or-less what the parameters are (e.g., doping density of so-and-so, 
etc.), but not how they're used, what they do (no formula, no description, 
no intent).  Even the ones that do (e.g., HSPICE manual gives some 
equations) have gaps between equations, and nothing explained.  I certainly 
haven't seen anything that says functionally what the parameters relate to, 
in the V/I/time/temp characteristics of the resulting device.

Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even though all 
manuals put it in the list of "these parameters apply to _all_ LEVEL=1,2,3,6 
models".  Also, parameters are hit and miss between all the flavors and 
makes of SPICE.  Just fantastic.

Adjusting the parameters one by one, I can't seem to find anything that 
doesn't simply shift the DC transfer curve up or down (in terms of current), 
left or right (equivalent to a change in VTO), or skew it (NFS appears to be 
the only parameter that actually causes subtreshold conduction in 
LEVEL=2,3).

No single parameter softens the drain output curves (the transition between 
triode and linear ranges), or causes Gm to drop at high current (analogous 
to BJT hFE degredation at high current).  IIRC, these were the reason for 
creation of the JFET + MOS model, which is sometimes seen.

Tips?

Tim

-- 
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
Website: http://seventransistorlabs.com 

On Mon, 25 Sep 2017 01:51:19 -0500, "Tim Williams"
<tmoranwms@gmail.com> wrote:

>I can't find any useful references for MOSFET SPICE parameters. Does this >information exist entirely in book form, or something? > >Example: websites copypasting the SPICE 3f5 handbook. This says >more-or-less what the parameters are (e.g., doping density of so-and-so, >etc.), but not how they're used, what they do (no formula, no description, >no intent). Even the ones that do (e.g., HSPICE manual gives some >equations) have gaps between equations, and nothing explained. I certainly >haven't seen anything that says functionally what the parameters relate to, >in the V/I/time/temp characteristics of the resulting device. > >Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even though all >manuals put it in the list of "these parameters apply to _all_ LEVEL=1,2,3,6 >models". Also, parameters are hit and miss between all the flavors and >makes of SPICE. Just fantastic. > >Adjusting the parameters one by one, I can't seem to find anything that >doesn't simply shift the DC transfer curve up or down (in terms of current), >left or right (equivalent to a change in VTO), or skew it (NFS appears to be >the only parameter that actually causes subtreshold conduction in >LEVEL=2,3). > >No single parameter softens the drain output curves (the transition between >triode and linear ranges), or causes Gm to drop at high current (analogous >to BJT hFE degredation at high current). IIRC, these were the reason for >creation of the JFET + MOS model, which is sometimes seen.
Rs? -- John Larkin Highland Technology, Inc lunatic fringe electronics
On 09/25/2017 08:51 AM, Tim Williams wrote:
> I can't find any useful references for MOSFET SPICE parameters. Does > this information exist entirely in book form, or something? > > Example: websites copypasting the SPICE 3f5 handbook. This says > more-or-less what the parameters are (e.g., doping density of > so-and-so, etc.), but not how they're used, what they do (no formula, > no description, no intent). Even the ones that do (e.g., HSPICE > manual gives some equations) have gaps between equations, and nothing > explained. I certainly haven't seen anything that says functionally > what the parameters relate to, in the V/I/time/temp characteristics > of the resulting device. > > Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even > though all manuals put it in the list of "these parameters apply to > _all_ LEVEL=1,2,3,6 models". Also, parameters are hit and miss > between all the flavors and makes of SPICE. Just fantastic. > > Adjusting the parameters one by one, I can't seem to find anything > that doesn't simply shift the DC transfer curve up or down (in terms > of current), left or right (equivalent to a change in VTO), or skew > it (NFS appears to be the only parameter that actually causes > subtreshold conduction in LEVEL=2,3). > > No single parameter softens the drain output curves (the transition > between triode and linear ranges), or causes Gm to drop at high > current (analogous to BJT hFE degredation at high current). IIRC, > these were the reason for creation of the JFET + MOS model, which is > sometimes seen. > > Tips? > > Tim >
you'll have to google around a bit. each different simulation model (level=) usually has some associated paper or something wot explains it. (that doesn't mean you or I will understand it tho')
On Mon, 25 Sep 2017 01:51:19 -0500, "Tim Williams"
<tmoranwms@gmail.com> wrote:

>I can't find any useful references for MOSFET SPICE parameters. Does this >information exist entirely in book form, or something?
Perhaps: "Operation and Modeling of the MOS Transistor" Yannis Tsividis, Colin McAndrew Oxford University Press, 2011 ISBN 978-0-19-517015-3
> >Example: websites copypasting the SPICE 3f5 handbook. This says >more-or-less what the parameters are (e.g., doping density of so-and-so, >etc.), but not how they're used, what they do (no formula, no description, >no intent). Even the ones that do (e.g., HSPICE manual gives some >equations) have gaps between equations, and nothing explained. I certainly >haven't seen anything that says functionally what the parameters relate to, >in the V/I/time/temp characteristics of the resulting device. > >Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even though all >manuals put it in the list of "these parameters apply to _all_ LEVEL=1,2,3,6 >models". Also, parameters are hit and miss between all the flavors and >makes of SPICE. Just fantastic.
My PSpice documentation shows Lambda in Level=3 equations
> >Adjusting the parameters one by one, I can't seem to find anything that >doesn't simply shift the DC transfer curve up or down (in terms of current), >left or right (equivalent to a change in VTO), or skew it (NFS appears to be >the only parameter that actually causes subtreshold conduction in >LEVEL=2,3).
Subthreshold behavior was not ever modeled adequately until PSpice Level=5 (or was it Level=6 ;-)
> >No single parameter softens the drain output curves (the transition between >triode and linear ranges), or causes Gm to drop at high current (analogous >to BJT hFE degredation at high current). IIRC, these were the reason for >creation of the JFET + MOS model, which is sometimes seen. > >Tips? > >Tim
The transition "bump" from triode to linear regions was fixed (in PSpice numbering) going from Level=5 to Level=6. I don't know the equivalent numbering in HSpice... perhaps 27->28? Did you vary ALL parameters ?:-)... .MODEL nch_16 NMOS ( LMIN = {1.000E-05-dxln} LMAX = 20.001e-006 + WMIN = {1.0000E-05-dxwn} WMAX = 900.001e-6 +LEVEL = 7 TNOM = 25.0 VERSION = 3.2 +TOX = {toxn} TOXM = {toxn} XJ = 1.6000000E-07 +NCH = 3.9000000E17 LLN = -1 LWN = 1.0000000 +WLN = 1.0000000 WWN = 1 LINT = 1.0000000E-08 +LL = 0.00 LW = 0.00 LWL = 0.00 +WINT = 3.0000000E-09 WL = 0.00 WW = 0.00 +WWL = 0.00 MOBMOD = 1 BINUNIT = 2 +XL = {-0.02E-6+dxln} XW = {0.0+dxwn} DWG = 0.00 +DWB = 0.00 ACM = 12 LDIF = 9.00E-08 + RSH = {rshn} RD = 0 +RS = 0 +VTH0 = {0.4365561+dvthn} WVTH0 = {0+dwvthn} PVTH0 = {0+dpvthn} +K1 = 0.5142301 K2 = 1.4797844E-02 +K3 = 0.00 DVT0 = 0.00 DVT1 = 0.00 +DVT2 = 0.00 DVT0W = 0.00 DVT1W = 0.00 +DVT2W = 0.00 NLX = 0.00 W0 = 0.00 +K3B = 0.00 VSAT = {9.0000010E04+dvsatn} +UA = -6.1227400E-10 UB = 2.3135539E-18 UC = 1.0876087E-10 +RDSW = 1.7000000E02 PRWB = 0.00 PRWG = 0.00 +WR = 1.0000000 U0 = 4.2203390E-02 A0 = 0.3700794 +KETA = 9.2634850E-03 A1 = 0.00 A2 = 0.9900000 +AGS = 2.0000000E-02 B0 = 0.00 B1 = 0.00 +VOFF = -0.1405780 NFACTOR = 1.0000000 CIT = -1.0000001E-04 +CDSC = 0.00 CDSCB = 0.00 CDSCD = 0.00 +ETA0 = 5.0000000E-05 ETAB = -5.0000000E-05 DSUB = 0.00 +PCLM = 0.8000001 PDIBLC1 = 1.0000000E-06 PDIBLC2 = 6.9999990E-04 +PDIBLCB = 1.0000000E-02 DROUT = 0.00 PSCBE1 = 2.0000003E08 +PSCBE2 = 1.0000000E-06 PVAG = 0.00 DELTA = 1.0000000E-02 +ALPHA0 = 0.0 ALPHA1 = 0.448150714 BETA0 = 11.59263 +KT1 = -0.2240464 KT2 = -3.0266136E-02 AT = 2.0000000E04 +UTE = -1.5799998 UA1 = 1.2240000E-09 UB1 = -1.1209314E-18 +UC1 = 1.1643822E-11 KT1L = 0.00 PRT = 0 +CJ = {cjn} PB = 0.6882682 MJ = 0.3595262 +CJSW = {cjswn} PBSW = 0.6882682 MJSW = 0.2003879 +CJSWG = {cjswgn} PBSWG = 0.6882682 MJSWG = 0.43879 +CGDO = {cgon} CGSO = {cgon} TCJ = 0.001040287 +TCJSW = 0.000645489 TCJSWG = 0.000645489 TPB = 0.001554306 +TPBSW = 0.001554306 TPBSWG = 0.001554306 +JS = 3.50E-07 JSW = 1.45E-12 NJ = 1 +XTI = 3 CAPMOD = 3 NQSMOD = 0 +XPART = 1 CF = 0 TLEV = 1 + TLEVC = 1 CALC ACM= 1 SFVTFLAG = 0 +DLC = 3E-9 LLC = -0.039 NOIMOD = 2 +NOIA = 2.00E19 NOIB = 1.2E04 NOIC = 2.5E-13 +EM = 3.0E07 EF = 0.874 ) ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I'm looking for work... see my website. Thinking outside the box...producing elegant & economic solutions.
On Monday, September 25, 2017 at 8:51:31 AM UTC+2, Tim Williams wrote:
> I can't find any useful references for MOSFET SPICE parameters. Does this > information exist entirely in book form, or something? > > Example: websites copypasting the SPICE 3f5 handbook. This says > more-or-less what the parameters are (e.g., doping density of so-and-so, > etc.), but not how they're used, what they do (no formula, no description, > no intent). Even the ones that do (e.g., HSPICE manual gives some > equations) have gaps between equations, and nothing explained. I certainly > haven't seen anything that says functionally what the parameters relate to, > in the V/I/time/temp characteristics of the resulting device. > > Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even though all > manuals put it in the list of "these parameters apply to _all_ LEVEL=1,2,3,6 > models". Also, parameters are hit and miss between all the flavors and > makes of SPICE. Just fantastic. > > Adjusting the parameters one by one, I can't seem to find anything that > doesn't simply shift the DC transfer curve up or down (in terms of current), > left or right (equivalent to a change in VTO), or skew it (NFS appears to be > the only parameter that actually causes subtreshold conduction in > LEVEL=2,3). > > No single parameter softens the drain output curves (the transition between > triode and linear ranges), or causes Gm to drop at high current (analogous > to BJT hFE degredation at high current). IIRC, these were the reason for > creation of the JFET + MOS model, which is sometimes seen. > > Tips? >
Does this help any?: https://imgur.com/a/lpo5Q From Andrei Vladimirescu, The Spice Book
On Mon, 25 Sep 2017 14:10:27 -0700 (PDT), Klaus Kragelund
<klauskvik@hotmail.com> wrote:

[snip]
>> > >Does this help any?: > >https://imgur.com/a/lpo5Q > >From Andrei Vladimirescu, The Spice Book
Just a few pages >:-}.... <http://www.analog-innovations.com/DeviceModelsSubckts/MOS_Spice_Model_Parameters.pdf> ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I'm looking for work... see my website. Thinking outside the box...producing elegant & economic solutions.
On Monday, September 25, 2017 at 11:10:37 PM UTC+2, Klaus Kragelund wrote:
> On Monday, September 25, 2017 at 8:51:31 AM UTC+2, Tim Williams wrote: > > I can't find any useful references for MOSFET SPICE parameters. Does this > > information exist entirely in book form, or something? > > > > Example: websites copypasting the SPICE 3f5 handbook. This says > > more-or-less what the parameters are (e.g., doping density of so-and-so, > > etc.), but not how they're used, what they do (no formula, no description, > > no intent). Even the ones that do (e.g., HSPICE manual gives some > > equations) have gaps between equations, and nothing explained. I certainly > > haven't seen anything that says functionally what the parameters relate to, > > in the V/I/time/temp characteristics of the resulting device. > > > > Also no explanation of why LAMBDA doesn't exist in LEVEL=3, even though all > > manuals put it in the list of "these parameters apply to _all_ LEVEL=1,2,3,6 > > models". Also, parameters are hit and miss between all the flavors and > > makes of SPICE. Just fantastic. > > > > Adjusting the parameters one by one, I can't seem to find anything that > > doesn't simply shift the DC transfer curve up or down (in terms of current), > > left or right (equivalent to a change in VTO), or skew it (NFS appears to be > > the only parameter that actually causes subtreshold conduction in > > LEVEL=2,3). > > > > No single parameter softens the drain output curves (the transition between > > triode and linear ranges), or causes Gm to drop at high current (analogous > > to BJT hFE degredation at high current). IIRC, these were the reason for > > creation of the JFET + MOS model, which is sometimes seen. > > > > Tips? > > > > Does this help any?: > > https://imgur.com/a/lpo5Q > > From Andrei Vladimirescu, The Spice Book
Better resolution: http://www.electronicsdesign.dk/tmp/BJTModel_Vladimirescu.pdf
"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in 
message news:n3lisch30ajfn09dlkpfvv8l6ql5u3pnq0@4ax.com...
> The transition "bump" from triode to linear regions was fixed (in > PSpice numbering) going from Level=5 to Level=6. I don't know the > equivalent numbering in HSpice... perhaps 27->28? > > Did you vary ALL parameters ?:-)...
I need 3f5 compatible parameters. Tim -- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
"Klaus Kragelund" <klauskvik@hotmail.com> wrote in message 
news:91e8ce47-666b-4982-bc11-4840c0032caf@googlegroups.com...
> Does this help any?: > > https://imgur.com/a/lpo5Q > > From Andrei Vladimirescu, The Spice Book
Well, no. "Nice copy right"? Also, MOS, not BJT. Did find this, https://www2.eecs.berkeley.edu/Pubs/TechRpts/1980/ERL-80-7.pdf which has been helpful. It's still not clear what the parameters and interactions are, but I seem to have done a satisfactory job. The remaining quirks are: - Triode-to-linear piecewise discontinuity - Subthreshold to strong inversion piecewise discontinuity (Gm peaks at the transition, slow clap) - Weird Gm profile, including sharp peak in the middle of operation (around Idmax) and sharp reduction at Idm (doesn't seem to be any way to accomplish this; I'm ab/using triode region to make it look alright at low voltages at least) - Output conductance can't be made high enough with LAMBDA (LEVEL=2) Which mostly seem to be limitations of the model itself, and also quirks of the device in question, which is very different from the kind of MOS they wanted to model in the first place. Plus, I now seriously doubt that any manufacturer has ever produced a model of this type (i.e., primarily using SPICE primitives) that's more than ballpark accurate to the real part. So I'm calling it good. And that's just forward DC parameters... But I have a pretty good idea what I need to make realistic capacitors, so that shouldn't be so bad. Unsure about body diode and recovery. I should maybe build a test circuit for that, or just ignore it entirely... Tim -- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
On Mon, 25 Sep 2017 17:39:52 -0500, "Tim Williams"
<tmoranwms@gmail.com> wrote:

>"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in >message news:n3lisch30ajfn09dlkpfvv8l6ql5u3pnq0@4ax.com... >> The transition "bump" from triode to linear regions was fixed (in >> PSpice numbering) going from Level=5 to Level=6. I don't know the >> equivalent numbering in HSpice... perhaps 27->28? >> >> Did you vary ALL parameters ?:-)... > >I need 3f5 compatible parameters. > >Tim
So contact Berkeley... <http://bwrcs.eecs.berkeley.edu/Classes/IcBook/SPICE/> Although it appears that 3f5 is obsolete. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I'm looking for work... see my website. Thinking outside the box...producing elegant & economic solutions.