Aqui está a página que achamos que você queria. Veja o resultado da pesquisa ao invés de:


Contate um especialista

Esse produto foi substituído por um novo:

Principais recursos e especificações

Key advantages of the NeuroFET:

  • The model can be used in all bias condition (including Vds<=0 as it happens in switches)
  • Improved DC and RF convergence, compared to table-based models
  • Improved distortion simulation at low amplitude, compared to table-based models
  • Less degradation outside the measured data region, compared to table-based models
  • More accurate S-parameters versus bias, compared to compact models
  • Better distortion and power-added efficiency (PAE) simulation, compared to compact models
  • The model is general and works for both HEMT and MESFET devices
  • It is available in Keysight's Advanced Design System (ADS), the industry standard RF simulator


The W8531EP IC-CAP NeuroFET Extraction Package license includes extraction for the Keysight NeuroFET model for FET and HEMT devices. The NeuroFET is a measurement-based model developed by the Keysight Technology Center.

Unlike other measurement-based models (e.g. Keysight Root model) that use splines to interpolate constitutive relations data stored in multi-dimensional tables, the Keysight NeuroFET model uses an artificial neural network (ANN) to describe the measured states. The ANN is made up of simple, interconnected functions, called neurons, with weighted connections. The functions are infinitely differentiable, providing a good distortion behavior. They also work well on the boundaries and outside the measured regions (extrapolation) resulting to a better convergence behavior in the simulator.

The IC-CAP NeuroFET Extraction package enables users to make all of the necessary DC and S-parameter measurements that are necessary to extract the model (including some basic de-embedding). A dedicated acquisition procedure avoids introducing device degradation during measurements, until the high power region is measured. The ANN training is automatically accomplished through a dedicated procedure that maximizes CPU usage by multi-threading the error optimization processes. The output model file can then be simulated using ADS within the IC-CAP environment and users can verify the model versus the original data; or if available, compared to other measured data, such as non-linear measurements (not provided).

The IC-CAP NeuroFET Toolkit

Figure 1. The NeuroFET IC-CAP Extraction Package