What is the difference between data transistor and nonliear transistor(spice model)

Dear Sir,
I want to know the difference between data transistor and nonlinear transistor(spice model). What happend , If I go library and choose data transistor (Example ATF54143 AVAGO), we will get s-parameters , noise parameters of particular transistor and also Vds value and Id value. If we design Low Noise Amplifier using the Data transistor , set the biasing values of particular values Vds and Id , we are getting the NF and Gain around values,whatever they have given in the noise paramete data sheet. If I go the nonlinear same transistor ,there s-parameter and noise parameter are not availble, but spice model is avaible, If I set the same Vds and Id for the nonlinear transistor for designing LNA, I didn't get the values of NF and Gain as they have given in the data sheet, But when I went the DATA transistor , I got the NF and Gain as per data sheet. I want to know the difference between Data transistor and Nonlinear transistor.

The data transistor as you call it is the linear S-parameter model of the device and the Nonlinear transistor is the SPICE model, probably based on the Gummel Poon model of a transistor.
The S-parameters are data that has been measured from the device at the operating frequency and bias given in the file.
With a network analyser S-paramters are easy to measure.
I have never made a SPICE model but Iexpect the data has been derived from measurements at low frequencies.
Both models are used differently, teh SPICE model for biasing and the S-paremter model for matching the input and output impedances.
Before youask why do they give different answers. They are only models, NOT the device itself. Use them as a guide and within their limitations.
The only true model is the device itself.
Peter

Dear sir,
Which transistor I need to take , am confused, because nonlinear (spice) transistor is giving good NF and Gain by varying Vds and Id within the Maximum limit . But data transistor(subcircuit) , we have to adjust bias as per data sheet values Vds and Ids, will get the NF and Gain as specified in the data sheet. We are not able to go above or below Vds and Id , even though if we go , we may not get the NF and Gain as mentioned in the datasheet. But nonlinear transistor (spice) , we can set the Vds and Id. within the maximum limit, will get good the NF and Gain. I am confused, which transistor , I need to choose, whether I have to choose the nonlinear (spice) or data transistor . I want to know the exact difference between these two. Can you help me

If you want to know what the difference between the models is then look up S-parameter models and SPICE models either on the net or in a book, you will get a more detailed explaination than I can give.
As to which one to use; my personal preference is for S-parameters. I can check them myself by direct measurement of a device. I have used SPICE models in the past but not found them to be very good. Things may have changed with better non linear simulators now being available.

Ther are application notes available for this device. Simulate these circuits with both models and see which one gives results that are closest to those presented inthe application note and choose that model.
failing that design a circuit with one or other of the models, build it and measure it.
You are going to have to build a prototype of your circuit, to verify the models of both the active and passive components in your circuit as weel as the layout.
Unless you can find someone who has direct experience of those particaulr models you will have to gain it yourself by experiment.
Peter