Linearization of diode in frequency (s) domain

[FusionITR]

What is the common practice of dealing with a diode in the circuit when developing a s domain transfer function?

 

[crutschow]

If the diode is being used as a low-level RF signal detector, then it can be generally represented with an AC to DC logarithmic transfer function

 

[MrAl]

Hi there,


It depends on your application. For example, sometimes you can represent the diode as a simple DC voltage source, where the diode looks like a DC voltage drop. For another example, if the non linear diode drop is not a detail needed for the application but the diode still provides a switching action, sometimes you can use a large signal approximation where you consider only the switching action...an example of this would be for say a full wave rectified sine wave rectified by diodes where the voltage drop may be of less importance than the overall waveshape out of the rectifier...in this case you might consider the output Laplace from of the waveshape and use that as the input to the next stage which may be completely linear. In the rectifier example where the input is a sine wave A*sin(pi*t/b) the output is abs(A*sin(pi*t/b)) and the Laplace form would be something like: abs(A)*coth(b*s/2)*pi*b/(b^2*s^2+pi^2), where 2*b is the time period of the sine wave.
That may or may not help, and also if you intend to use that exact expression for something you should check it for complete accuracy before using.
In any case, the idea is to find the function of s that appears after the diode(s) and use that as the input to the next stage.