Switching frequency stability and ripple reduction in a boost converter

I am simulating a high voltage boost converter (270Vdc -> 500Vdc) that is feeding a constant power load (several kilowatts), and so far, my control law seems to be working. I'm using a cascaded PI controller (inner current loop, followed by outer voltage loop), and the output voltage is tracking my reference voltage.

--> I have a lot of input current ripple (about 12 amps of ripple) at a switching and sampling frequency of 10kHz. I want to try and reduce this ripple by increasing the switching frequency to 50 to 100kHz, while keeping the sampling frequency at 10kHz.

When I increase the switching frequency to ~50kHz, but I keep the sampling frequency at 10kHz, I notice that the system goes unstable. This is my big quesiton: Why is this? I would think that, as long as I keep my sampling frequency at 10kHz, the system should be stable.

Answer 1

[Pratyush]

Hi Richard,

I understand that increasing the switching frequency while keeping a constant sampling frequency destabilizes your system.

This instability is could be caused due to the Nyquist criterion. This means that the sampling frequency must be at least twice the frequency of the signal being sampled. To address this, increasing the sampling frequency to at least twice the new switching frequency can help maintain stability and accurately capture the system's behavior.

You can refer to the following link to know more about Nyquist criterion.

https://www.electrical4u.com/nyquist-stability-criterion/#google_vignette