Project Details
Description
The goal of this research is to give the designer of power amplifiers the necessary insight into the behavior of the circuit. Therefore, models of the amplifiers need to be created first. Afterwards, also a methodology for using these models needs to be proposed.
The applicability of this technique will be illustrated by means of the realization of some amplifiers. This allows for the verification of the assumptions made in a realistic setting. This verification also makes the techniques more accessible for industrial application.
The goal can be split up into four tasks:
- Finding the good modeling approach. In this phase, existing modeling techniques are adapted so that they are able to describe the strong nonlinear behavior of power amplifiers. First, an extension of the Volterra theory was tried. It was shown that this description is not accurate enough to describe power amplifiers. Second, the best linear approximation was validated. This model was successfully extended towards MIMO systems and subsequently applied to a two-stage power amplifier with an adaptive biasing circuit.
- Evaluating the usefulness of the models. In this phase, it was verified whether or not the models provided the necessary insight for a circuit designer. This was shown by the analysis of the two-stage power amplifier by means of the models. The dominant nonlinear contributions and their causes - such as the nonlinear interaction between the different sub-circuits - could be detected.
- Using the models for system-level simulations. These models allow for fast, straightforward and accurate simulations of the complete front-end. Therefore, rules for cascading and placing in parallel the models need to be proposed.
- Proposing a design methodology starting from the developed models. The possibilities to detect the nonlinear contributions and to cascade different models are combined. This results in new methods that help the designer to take the specifications on the nonlinear behavior into account from the beginning of the design.
The applicability of this technique will be illustrated by means of the realization of some amplifiers. This allows for the verification of the assumptions made in a realistic setting. This verification also makes the techniques more accessible for industrial application.
The goal can be split up into four tasks:
- Finding the good modeling approach. In this phase, existing modeling techniques are adapted so that they are able to describe the strong nonlinear behavior of power amplifiers. First, an extension of the Volterra theory was tried. It was shown that this description is not accurate enough to describe power amplifiers. Second, the best linear approximation was validated. This model was successfully extended towards MIMO systems and subsequently applied to a two-stage power amplifier with an adaptive biasing circuit.
- Evaluating the usefulness of the models. In this phase, it was verified whether or not the models provided the necessary insight for a circuit designer. This was shown by the analysis of the two-stage power amplifier by means of the models. The dominant nonlinear contributions and their causes - such as the nonlinear interaction between the different sub-circuits - could be detected.
- Using the models for system-level simulations. These models allow for fast, straightforward and accurate simulations of the complete front-end. Therefore, rules for cascading and placing in parallel the models need to be proposed.
- Proposing a design methodology starting from the developed models. The possibilities to detect the nonlinear contributions and to cascade different models are combined. This results in new methods that help the designer to take the specifications on the nonlinear behavior into account from the beginning of the design.
| Acronym | OZR1462 |
|---|---|
| Status | Finished |
| Effective start/end date | 1/01/07 → 31/12/07 |
Keywords
- Measuring
Flemish discipline codes in use since 2023
- (Bio)medical engineering
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.