DC-DC converter error amplifier output

Question

I am new to DC-DC converter electronics. I was wondering what the output of the error amplifier/compensator looks like.

Is the output of the compensator being driven to the Vref value set on the non-inverting pin of the op-amp and if so, how?

Answer 1

The error amplifier theoretically delivers a clean dc voltage representative of the error between the desired output voltage set by \$V_{ref}\$ and what is actually measured. Its role is to drive the control variable, here the duty ratio \$D\$, for regulation purposes. In your example, the error voltage \$v_{comp}\$ (or \$v_{err}\$ in the below picture) is permanently compared to a sawtooth and adjusts the duty ratio between 0 and 100%:

In practice the error voltage is affected by some high-frequency noise, a ripple, which is coming from the switching action of the converter. This ripple has to be minimized so that the system does not react on its own noise. This is where the compensator is considered as an active filter which not only shapes the loop gain response (sets a crossover frequency \$f_c\$ and ensures phase/gain margins) but also makes sure the output ripple does not significantly go through and affect the control variable. This characteristic is usually obtained by quickly rolling off the compensator gain beyond crossover.

The below simulation shows a voltage buck converter with a fixed load during the start-up sequence. In absence of output voltage, the error amplifier rails up and pushes the duty ratio to the maximum possible value. However, a cycle-by-cycle current limit can truncate pulses length for safety reasons and effectively reduce the duty ratio. For instance, the first pulses may correspond to a 90% duty ratio but quickly shrink to 10% because the inductor current has reached the maximum limit you've set. When the output voltage reaches the regulation level, the compensator brakes and reduces the duty ratio until steady-state operation is obtained:

As you can see, there is a little bit of ripple straddling the dc bias but its value is insignificant compared to the continuous value which matters for control purposes. Finally, you can see how the error voltage will vigorously react (assuming a fast control system of course) when a perturbation like an output load step is applied. You can download this ready-made template from my webpage and exercise it on the free SIMPLIS demo, Elements.