I have imported the nonlinear model of a microwave transistor lp6836 (filtronic company) in the microwave office 2004 and ADS2005.
then for approve of imported transistor I calculated s parameters of this transistor in ADS and AWR. but the results differ with s2p of lp6836 very much especialy in phase of s parameters.
the biase condition is similar and the model is curtice-ettenberg.
it is a regular?
it is a due of small signal calculations and the model is only for large signal?
How much nonlinear models are accurate?(than S2ps)
Nonlinear modesl are generally not fitted to measured s-parameter.Because, s-parameter -by definition- are small signal parameters and nonlinear models are curve fitted or approximative models.Therefore, small signal parameters are not valid anymore except under certain bias condition and small signals are existed.
Another explanation can be parasitic elements might not be added to nonlinear models.
Briefly, use the proper model which you design on.
in discrete design usually use the s parameter file to design ur amplifier
always nonlinear models used to design the bias point , but not the gain and noise figure
it is better that u get the transistor and by network analyzer extract ur s parameter file , and noise performance it make the design ittirations short
for oscillator and power amplfier always use the nonlinear model or u can extarct the large signal Sparameter files with power sweep options in some network analyzers
You must always use non-linear models for your amplifiers and oscillators design, unless the models that you are going to use for design aren't available. The advantages of using the non-linear models than the s2p files are that they give more accurate simulation results, more flexible and reduce lots of design time. That is why those simulation tools cost so much because of the non-linear models.