Is This LT8390 Design Stable? I have LTPowerCad plot attatched

Does not look stable to me.

A stable system should look something like this

Unfortunately, there will always be deviation from LTPowerCAD as compared to actual bench test because of the many factors in the real world like tolerances of different components and accuracy of the supply and output. The best way is still to do the bode plot in the actual board and adjust accordingly.

Regards,

Jhun

Thank you. I notice that the 'example' design has a load transient that looks like the first two screenshots below. Then my edited design load transient is below that. Would that be considered stable now? Or is it still off? What do you mean by finding the bode plot for the board, I have never designed a dc dc converter before so I wasn't sure how the appropriate values could be adjusted after the board is manufactured? If you could, could you please list some steps or a method to finding the bode plot and then the load transient for a manufactured pcb?

example's Buck:

example's Boost:

My edited design Buck load transient:

My edited design Boost load transient:

:

Refer to the documentation below. It should help you: Understand Power Supply Loop Stability and Loop Compensation—Part 1: Basic Concepts and Tools | Analog Devices

ITH pin, or Vc pin in this IC, have 3 components that you can manipulate in order to have a stable design. You need to understand how each of the components work to design an effective loop compensation to make your converter stable. Read about loop compensations by Henry Zhang to gain more insights.

Regards,

Jhun

Okay thank you that is a really helpful document. I had a question though is my updated load transients okay? It looks like it follows the shape of what it should be but changes much more, perhaps that is the ripple. For my buck load transients it looks like the voltage is above the expected plot. Instead of being centered at zero it is centered at around 100mV. Is that okay? Would I be okay to go ahead and make the PCB and then test with a network analyzer? I was curious if I should make the control loop adjustable if possible so I could tune it after the board is made?

william22 said:

I had a question though is my updated load transients okay?

It is better than before.

william22 said:

Would I be okay to go ahead and make the PCB and then test with a network analyzer?

Yes. Make sure to follow the PCB guidelines or refer to the demoboard if available and copy as close as possible to minimize complexity of design.

william22 said:

I was curious if I should make the control loop adjustable if possible so I could tune it after the board is made?

You can use an external compensation board to make it easier to adjust. Microsoft Word - Lab board 013A QSG.doc (analog.com) LB013

Regards,

Jhun

Thank you so much, definitely seems like I need to physically test it like you said. I haven't been able to find a way to purchase a 013 Lab board, is that something that designers will typically just do the PCB design themselves for personal use or is there one available for purchase? Due to time constraints, I don't think I will be able to design an external eval board, but I think I want to in the future now.

Thank you so much, definitely seems like I need to physically test it like you said. I haven't been able to find a way to purchase a 013 Lab board, is that something that designers will typically just do the PCB design themselves for personal use or is there one available for purchase? Due to time constraints, I don't think I will be able to design an external eval board, but I think I want to in the future now.