Drain-Source Current

The drain-source current model in Agilent EEFET3 is comprised of various analytic expressions that were developed through examination of g_m vs. bias plots on a wide class of devices from various manufacturers. The expressions below are given for V_ds > 0.0 V although the model is equally valid for V_ds < 0.0 V. The model assumes the device is symmetrical, and one need only replace V_gs with V_gd and V_ds with -V_ds in order to obtain the reverse region (V_ds < 0.0 V) equations. The g_m, g_ds and I_ds equations take on 4 different forms depending on the value of V_gs relative to some of the model parameters. The I_ds expression is continuous through at least the second derivative everywhere.

if V_gs V_g and V_DELT 0.0

else if V_DELT > 0.0 and V_gs > V_gb

else if V_gs V_t

else

where

The preceding relations for I_dso, g_mo and g_dso can now be substituted in the following equations that model the current saturation and output conductance. This portion of the model can be recognized from the work of Curtice [1].

These expressions do an excellent job of fitting GaAs FET I-V characteristics in regions of low power dissipation; they will also fit pulsed (isothermal) I-V characteristics. To model negative conductance effects due to self-heating, the thermal model of Canfield was incorporated [2]. With this final enhancement, the DC expressions for I_ds and associated conductances become:

where

Qualitatively, the operation of the drain-source model can be described as follows:

The V_ds dependence of the equations is dominated by the parameters VSAT, GAMMA, KAPA, and PEFF. Isothermal output conductance is controlled by GAMMA and KAPA. The impact of GAMMA on output conductance is more significant near threshold. At V_gs=VCH, the output conductance is controlled only by KAPA. The parameter PEFF provides a correction to the isothermal model for modeling the self-heating effects manifested as a negative resistance on the I-V curves. The parameter VSAT represents the drain-source voltage at which the current saturates and output conductance becomes a constant (approximately).

The overall impact of VCH on the I-V characteristics is second order at best, and many different values of VCH will provide good fits to I-V plots. For most applications encountered, it is our experience that the default value of 1.0V is an adequate value for VCH. Similar to VCH, VDSO is a parameter that should be set rather than optimized. At V_ds=VDSO, the drain-source model collapses to a single voltage dependency in V_gs. It is recommended that the user set VDSO to a typical V_ds operating point in saturation. At this point, many of the parameters can be extracted right off a I_ds-V_gs plot for V_ds=VDSO or preferably, a g_m(DC)-V_gs plot at V_ds=VDSO.

When V_ds=VDSO and PEFF is set large (to disable the self-heating model), the significance of the parameters VTO, VGO, VDELT, GMMAX are easily understood from a plot of g_m(DC)-V_gs. GMMAX is the peak constant transconductance of the model that occurs at V_gs=VGO. The parameter VTO represents the gate-source voltage where g_m goes to zero. If VDELT is set to a positive value, then it causes the transconductance to become linear at V_gs = VGO - VDELT with a slope equal to that of the underlying cosine function at this voltage. The parameter definitions are illustrated in the following figure.

Figure 21 Agilent EEFET3 g_m-V_gs Parameters