Buck-boost INPUT voltage regulation stability (MPPT)

Question

I am designing buck-boost converter with Maximum Power Point Tracker. The idea is to build feedback loop that regulates input voltage by changing output current. Output will be connected to Li-Ion battery. Constant voltage output and MCU driven MPPT algorithm will be implemented in the future, the main objective now is to build constant INPUT voltage converter.

I am doing simulations in LTspice but all designs that I try are unstable. I attach one of the designs which after some fluctuations starts to regulate input voltage but there are only few degrees of phase margin.

Could someone suggest the right feedback loop design that would be more stable than that? Would it be enough just to add Type III compensation network instead of Type I or maybe there is completely smarter way to do the feedback loop?

In the end, after designing stable input voltage regulation loop, the idea is to multiply input voltage and current with MCU and with perturb and observe method find the MPP. MCU will regulate input voltage by changing resistance in digital potentiometer which will be connected instead of bottom resistor in voltage divider of the feedback loop.

Answer 1

This is a slightly unusual use of this device according to the data sheet but, I can see it operating with a little more care in component selection. See figure 9 in the data sheet for the loop compensation network applied across what is R3 in your diagram. I suggest that you do use the type II compensation scheme because of the addition of the op-amp in the feedback path.

I would also suggest that you use the run pin so that the initial starting transient is also avoided. You could use a capacitor on this pin to create a soft starting circuit.

Answer 2

What for do you need the analog feedback loop?

As far as I understand from your question, you plan to use MCU to control the device anyway. It is much more feasible to implement the whole control in MCU rather than to combine analog and digital controls.

I would suggest the following design:

• generate control sequence for your converter either by MCU or by some simple logic, if MCU is not fast enough;
• your power stage (switches and inductor) form "DC-DC transformer" with voltage ratio defined by the duty factor of control signal;
• this design is inherently stable as it does not need feedback loop;
• then you vary the duty factor to find maximum power point.