Re: How to model class-B distortion of a complex signal

[Nick Foster]

Anish,

If you really just want to model crossover distortion in a class-B amplifier, you could use a pair of Rail blocks:

If you're trying to model the distortion created by a real solid-state amplifier, and not trying to correct it with predistortion, there are a million different models which attempt to accurately characterize distortion in SSPAs: Rapp, Saleh, and Ghorbani seem to be most widely used. They're all pretty simple.

Here's a very simple, memoryless model (Ghorbani) implemented as an embedded Python block:

https://gist.github.com/bistromath/75d30dd9691fd18211b306db630d8c4f

Some basic information about the model can be found here:

https://www.mathworks.com/help/comm/ref/memorylessnonlinearity.html

Nick

On Tue, Jan 25, 2022 at 12:03 PM Wheberth Damascena Dias <wheberth@gmail.com> wrote:

Hello Anish,

Non-linear  power amplifier modeling and pre-distortion is a vast subject.

If you want to go deep in this matter I would suggest you to take a look at Baseband polinomial models.

 Here a reference I have used some time ago:

https://repositum.tuwien.at/handle/20.500.12708/13980

If you just need something quick, you may want to lookout for AM/PM models, such as the Saleh model.

Good luck!

Best Regards

Wheberth Dias

On Tue, Jan 25, 2022, 20:44 Anish Mangal <anishmg@umich.edu> wrote:

Hi,

How should I go about modeling the distortion introduced by a class-B amplifier stage that clips the signal when its value is below zero?

Example:

I want to combine two FM modulated signals, pass them through this "class-B distorter", demodulate them, and analyze the quality of the signal produced (compared with the undistorted one).

How should I go about doing it.

Created the flowgraph attached, but this is obviously wrong. :)

--

Anish