MOSFET Modeling

MODEL Maker calculates parameters for many types of MOSFET devices. It has been tested with both power and small-signal devices, nmos and pmos types.

MODEL Maker assumes that the source and body, or substrate, of the MOSFET are connected.

NOTE: Be sure to input all MOSFET parameters as positive numbers. You should be cautious about setting any of the optional geometry parameters (especially l and w) on MOSFET cards that reference MODEL Maker MOSFET models. MODEL Maker calculates accurate MOSFET models using the default DR. SPICE geometry values. MODEL Maker calculates the value of a gate resistor for it’s MOSFET models. You must add this gate resistance to your circuit to obtain the highest accuracy in transient and ac simulations. Gate resistors aren’t needed for dc/operating point simulations.

There are three methods for specifying dc characteristics. Each method requires a different amount of information from the datasheet and each generates a different device model; the method that requires the most information generates the most accurate device model. If you don’t provide any data, DR. SPICE uses it’s default MOSFET dc characteristics. MODEL Maker automatically uses the most accurate method based on the data you provide. For MODEL Maker to select a method, you must provide all the parameters used by that method.

This is the simplest method for specifying dc characteristics. These parameters provide the gate threshold voltage and the forward transconductance from the datasheet. For highest accuracy, specify a transconductance evaluated close to your circuit’s operating point.

Parameter	Description
gfs	Forward transconductance with device turned on.
vt0	Gate threshold voltage with body tied to source.

This method generates a model based on two points from the output characteristic curves. The points are selected from the ”pinch off” or ”saturated” region of the curves. The saturation region is defined to occur when VDS > VGS - VT0.

Parameter	Description
ids1, ids2	Drain to source current for points 1 and 2.
vgs1, vgs2	Voltage gate to source for points 1 and 2.

This method generates the highest accuracy device model. These parameters provide six points from the output characteristic curves.

Point 1 is selected from the highest IDS curve at a point midway into the saturation region. Point 2 is selected from a curve with a mid-level IDS at a point midway into the saturation region. Point 3 is selected from the lowest IDS curve at a point midway into the saturation region. Points 4 and 5 are selected from the same IDS curve as point 1, as far apart in VDS as possible while still remaining in the saturation region. Point 4 has the lower VDS.

Point 6 is selected from the ”linear” or ”ohmic” region of the curves midway up the ohmic line. A reasonable estimate of the VGS of this curve generally produces good results. If you use a different graph (compared to points 1 through 5) in the databook to read this point, check that the two graphs have consistent values.

MODEL Maker generates device models with the most accurate ac/transient response when the six-point dc method (Method 3, above) is used along with an ac method. If you don’t supply any parameters for ac characteristics, DR. SPICE uses its default ac model values.

The following parameters must be specified for MODEL Maker to generate ac model values: crssh and cissh. In addition, one of the following must be specified: coss0 and crss0, or cossh. The following parameters are optional: ton, toff, vsw, and rsw. If you don’t include them, MODEL Maker assumes default values and generates ac MOSFET model values. However, the frequency response and rise/fall times of the model are more accurate if these parameters are specified.

For highest accuracy, coss0 and crss0 values should be read from a capacitance vs. VDS curve. Otherwise, they can be read from the datasheet text. Use data from the largest VDS values provided on the curve.

To specify the next set of parameters, you can use one of two methods. The first method, specifying coss0 and crss0, is the most accurate. This method uses values read from a capacitance vs. VDS curve. The second method, cossh, uses a single value at high VDS.

Of the optional parameters, ton and toff are read from the datasheet. Make sure you use the delay times, not the rise and fall times. Delay times are defined as the time from the gate input changes 10% to the time the output changes 10%. The last two parameters, vsw and rsw, are calculated from the test circuit used by the manufacturer to measure rise and fall times. The manufacturer usually specifies the component values used in this circuit. When calculating vsw and rsw, also include the effects of a voltage divider formed by the generator output impedance and the cable load.

To ensure the most accurate ac/transient simulation, you should give MODEL Maker consistent capacitances and on/off times-that is, specify either maximum values or typical values for both. (Many datasheets specify maximum values for device capacitances, and typical values for rise and fall times in the text. Capacitance graphs usually specify typical values.)

Parameter	Description
cissh	Input capacitance (ciss) at high VDS.
coss0	Output capacitance (coss) at 0 VDS.
cossh	Output capacitance (coss) at high VDS.
crss0	Reverse transfer capacitance (crss) at 0 VDS.
crssh	Reverse transfer capacitance (crss) at high VDS.
rsw	Resistance of switching test circuit as seen by MOSFET gate.
toff	Turnoff delay time in switching test circuit.
ton	Turn on delay time in switching test circuit.
vsw	Voltage of switching pulse as seen by the MOSFET gate.

The following .DATASHEET statement (mos1.she) generates a model for an mtm5p25 PMOS MOSFET:

* dc from mot db mtm5p25 pb 3-400* /* use six vgs, id, vds points if given. * points are selected from three output characteristic curves and * ohmic line. * p1 on highest id curve mid curve in sat region * p2 on middle id curve mid curve in sat region * p3 on low id curve mid curve in sat region * p4,p5 on highest id curve as far apart as possible on near straight * line segment, p4 lower vds both in sat region * p6 on the ohmic region line mid way up line */ * cap from mot. table max values for h, graph ?typ? values for 0 * sw from mot. dly+rise/fall. vswload wrong. .datasheet mtm5p25 pmos + vgs1=8. ids1=6.65 vds1=32 + vgs2=7 ids2=3.5 vds2=28 + vgs3=6 ids3=1.4 vds3=24. + vgs4=8. ids4=6.65 vds4=30. + vgs5=8. ids5=6.65 vds5=40. + vgs6=9. ids6=3. vds6=8. + crssh=75pf cissh=900.pf coss0=500.pf crss0=150.pf + vsw=10.0 rsw=25. + ton=40.e-9 + vswload=10. toff=130e-9

* dc from mot db mtm5p25 pb 3-400 * /* use six vgs, id, vds points if given. * points are selected from three output characteristic curves and * ohmic line. * p1 on highest id curve mid curve in sat region * p2 on middle id curve mid curve in sat region * p3 on low id curve mid curve in sat region * p4,p5 on highest id curve as far apart as possible on near straight * line segment, p4 lower vds both in sat region * p6 on the ohmic region line mid way up line */ * cap from mot. table max values for h, graph ?typ? values for 0 * sw from mot. dly+rise/fall. vswload wrong. * .model mtm5p25 PMOS ( level = 1 vt0 = -4.30386 kp = 0.973269 lambda = 0 rd = 2.42998 rs = 0 cgso = 8.25e-006 cgdo = 7.5e-007 cbd = 3.5e-010 pb=1.) *rg = 53.9708

Parameter	Description
ids1, ids2, ids3, ids4, ids5, ids6	Drain to source current for points 1 through 6.
vds1, vds2, vds3, vds4, vds5, vds6	Voltage drain to source for points 1 through 6.
vgs1, vgs2, vgs3, vgs4, vgs5, vgs6	Voltage gate to source for points 1 through 6.