Peak current-mode controller

Question

I am studying a peak current-mode controller, and here is the schematic without slope compensation:

Before introducing RSENSE to the design, I understood that if feedback voltage increases, it is compared with a constant VREF to determine to do what with switches. According to the information from the site, the switches are reset when R_i · i_L = V_EA, indicating that i_L can't exceed I_EA.

I am encountering difficulties in understanding:

1. In many designs, RSENSE is selected in a way that i_L never reaches I_EA. For instance, when I simulate the circuit using LTspice and choose a smaller value for RSENSE, it always operates without issues. However, problems arise when I opt for a larger resistor value, leading to current limiting through the inductor. I am confused about what actually triggers the switches to turn ON/OFF if I_EA is picked to be very high. What is going to limit i_L then?

2. The schematic (figure 1) suggests that the switches are reset when R_i · i_L > V_EA. However, I am puzzled by datasheets such as the one for the LT3890, which consistently mention a fixed sensing voltage, in this case, 50 mV. I am uncertain about the role of this fixed sensing voltage in the comparator. My understanding is limited to the idea that if the sensed voltage exceeds 50 mV, the switches turn off, but I am unsure how this concept aligns with the schematic above.

Answer 1

1. Yes. Shunt resistance is a design parameter.

Given that EA is a real op-amp with bounded output voltage (saturation limits), we can select it, and the current sense (resistor, sense amp, offset, and comparator) such that peak current only up to some given design limit can be delivered.

Since the switches and inductor will be selected for a given current range, this allows us to fully bound one aspect of converter operation, greatly improving reliability in the face of uncertain input or output conditions (e.g. output overload or short circuit).

Varying the resistor arbitrarily, is simply varying part of that design limit. So you are making your design higher or lower powered, but independently of other parameter changes -- which as you have seen, doesn't make a whole lot of sense, or seem to accomplish much at a fixed operating point. But, even at a fixed operating point, inverter gain and therefore EA compensation is changing, so you can eventually reach a point where gross behavior changes (i.e. instability).

2. All instances use the term "maximum current sense".

That is, the threshold is variable internally, up to a maximum of this. This is equivalent to saying: the EA output is 0-50mV, in terms of current sense amp input voltage.