Buck Converter Control Loop Design

[jackbob]

I am designing a buck switching converter, the project is for a class so please don't suggest a purpose-built controller IC. I am designing the controller using a type 3 compensation op-amp circuit and built the circuit in LTspice. The simulation behaved as expected however there was a very large overshoot on the converter output during startup. I tried tweaking as many parameters as I could while maintaining an acceptable gain and phase margin but could not get rid of the overshoot. I then thought of how the compensation circuit worked and thought I could add a small RC circuit to delay the reference voltage seen by the op-amp during startup. The small addition of the RC circuit delays the reference voltage very briefly as the capacitor charges up through the resistor during startup. Once the capacitor is charged, it behaves like an open circuit as if it is not there. This delay of the reference voltage prevents the op-amp compensator circuit from generating a large error during startup.

I have included screenshots of both circuits before and after adding the RC delay boxed in red on the second circuit. I also included pictures of the startup transients for both converters to show the difference.

My question is will adding this RC circuit to the reference voltage pin change the transfer function or stability of the circuit? Is there a possibility adding this could somehow cause the converter to enter an unstable state? It seems to operate just as expected but I want to convince myself this will not contribute to instability of the system.

[jackbob]

In the industry (chips), they don't bother to dampen it as that would drastically reduce response time. The trick they did is to ramp the reference voltage over a "soft start" period.

Try adding an RC ramp on the reference, and report back.

Is that not what adding the RC circuit to the reference voltage does? It's not a ramp but the same idea, it delays the reference voltage. I'm pretty sure I did exactly what you just explained.

[jackbob]

Your ramp is not enough. Calculate the charging current of your output cap at the given ramp rate assuming the converter has perfect control loop, and ask yourself if that charging current is okay for the chosen L and C.

As I mentioned this is for a school assignment and no maximum component current parameters were given. I realize if this were a real circuit, transient currents that high would likely not go well with components this small. I will play around with the delay to see if I can reduce the current further to be more realistic for components in a converter of this size.

I am planning to possibly tweak this design slightly with common component values and actually build this circuit to analyze and learn from.

Another trick is to say goodbye to compensator and PWM. Just do a ramp compensated hysteresis based control, or add a bit extra spice to make it a COT converter. Those are known for being almost unconditionally stable.

When I implement a buck converter, I use those two control topologies the most as the loop is basically tuning-free.

I have to use a compensator and PWM for the assignment but I will look into this control topology you mentioned as it sounds very interesting. I plan to specialize in power electronics and I am always excited to learn more.

[jackbob]

 

Your ramp is not enough. Calculate the charging current of your output cap at the given ramp rate assuming the converter has perfect control loop, and ask yourself if that charging current is okay for the chosen L and C.

Within a couple minutes I was able to reduce the transient current a significant amount by increasing the delay as you said. Thank you for your help and suggestions! I appreciate it.