Models

Q: Why doesn't .OPTION SCALE work in my simulation?

A: Check to see if your MOSFET and model card are wrapped inside a .SUBCKT call. For example:

.SUBCKT pmos d g s b w=1 l=1 m=1 mp d g s b p w=w l=l m=m

.MODEL p pmos LEVEL=.......

.ENDS

If so, check the model parameters for any dependencies on width or length. These parameters, in this case, are not scaled. .OPTION SCALE only affects width and length at the time they're passed into the model card as the physical length and width. The model parameter is using width and length before they get scaled.

Solution: 209

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Q: What is the process to get proprietary MOSFET models?

A: The following MOSFET model level numbers are vendor proprietary and require a special license to use:

AMD

AMD

LEVEL 32

Dallas Semi

DALLAS

LEVEL 19

Hitachi

HIT

LEVEL 43

IBM

IBM

LEVEL 48

Lucent

ATT

LEVEL 45

Motorola

MOT

LEVEL 31

 

MOTSSIM

LEVEL 29

National

NAT

LEVEL 33

SGS Thomson

SGSTHOM

LEVEL 46

Sharp

SHARP

LEVEL 36

TI

TI

LEVEL 37

 

 

LEVEL 41

Vitesse

VITESSE

LEVEL 6

As of July 12, 2000, these are the contact people:

AMD

Barbara Vervenne

(512) 602-5783

Dallas Semi

Roy Hensley

(214) 450-3858

Hitachi

Len Mizrah

(408) 456-2081

IBM

Cal Bittner

(902) 769-6528

Lucent

Roberta Kulp

(610) 712-2614

Motorola

Ralph Sokel

(512) 933-5264

National

Chen Young Tao

(408) 721-5665

SGS Thomson

Claude Frank

33 476 92 63 47 (Tel)
33 476 08 96 52 (Fax)

Sharp

Sigenobu Kiyoi

81 7 4365 4273

TI

Jeff Brunson

(972) 480-2481

Vitesse

Aubrey Grey

(805) 388-7517

For a customer to receive a license from Avant! (the vendors do not cut the licenses), he/she must contact the vendor and request that an authorization letter is submitted to Avant! Once Avant! receives the letter, a license file can be processed.

Solution: 263

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Q: Why doesn't BJT LV5 - FT agree with net analysis results?

A: FT is a simple estimate for a cut-off frequency of the internal transistor:

ft=abs(gm-go)/(2*pi*max( 1.0d-18 , C_be+C_bc+C_bx ) )

The parameters for the equation are from the equivalent circuit for BJT (see Vertical Transistor, BJT AC Analysis).
FT does not include parasitic elements (Rb, Rc, Re for example), and does not include the external circuit. It just gives a simple estimate for cut-off frequency of the internal transistor and it cannot replace results of actual simulation.
For another discussion on this, see "Semiconductor Device Modeling with SPICE" by Massobrio and Antognetti, McGraw Hill, 1993, page 99.

Solution: 560

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Q: What is the importance of the following excerpt from the attached .lis file:

*****************************************************************

**warning BSIM3v3 VERSION parameter does not appear in the model card!!!

Failure to explicitly set VERSION may result in inaccurate results!

*****************************************************************

A: Avant! implemented the BSIM3v3 models released from UC Berkeley in the forms of LEVEL 49 and LEVEL 53. Since implementation, UC Berkeley has released patches and Avant! has implemented them as 3.1 (default for 98.2), 3.2 (default for 98.2 and 98.4) and 3.22 (default for 99.2, 99.4 and 2000.2).

If your models were developed for the VERSION = 3.1 release and this parameter is not present in your models/libraries, then Star-Hspice defaults to 3.22 (in Star-Hspice 2000.2 release), which may give you unexpected results.

Solution: 562

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Star-Hspice Manual - Release 2001.2 - June 2001