Dynamic control of DC-DC converter current limit

Question

I am interested in having dynamic control of the output current limit circuit of a DC-DC converter controller (the LTC3777 buck-boost controller in my case). This IC has a built-in current limit using an external high-side shunt resistor, which is monitored directly by the IC.

From looking at the datasheet, it appears that the two sides of the shunt resistor go directly to the inputs of a differential amplifier inside the chip, which is then tied into the rest of the feedback loop. What I would like to do is somehow modify that differential signal from the shunt resistor, so I can dynamically control the current limit setting using a microcontroller equipped with either with a DAC or digital POT.

My initial idea is to insert a differential amp (or current sense amp) in between the shunt and controller IC, and then adjust the gain of the amplifier to effectively control the current limit value. However, I'm concerned about this extra element introducing instabilities into the feedback loop due to the slight propagation delay said amplifier would introduce to the circuit.

Is the concern valid? Or are there other, more robust/simple approaches to do this (aside from controlling the current limit with a separate, independent circuit)?

Answer 1

the typical circuit in the application note has 100 ohm resistors between the sense resistor and the chip.

if you steal current from the Iavgsns pin the chip will see a lower voltage there and perceive a higher current.

You could use PWM or a DAC with a resistor to do this.

Answer 2

You can control the peak current limit by the current sense resistor .Imagine if the sense resistor was a rheostat then you could vary the current .Current sense resistors on practical SMPS chips do not waste much power because the peak voltage drop is set up to be low .I have set up discrete stuff at 600mV peak .I have not used your expensive chip but the basic idea is the same .What I have done many times is use a mosfet instead of a fixed sense resistor .The mosfet is cheap because the power wasted is low and DS volts are low and the on resistance needs are not too low .The mosfet is a voltage controlled resistor .When there is a limited range of current control needed I place a fixed resistor in parallel with the DS of the mosfet.When the max current needed can not be too high I place some fixed resistance in series with the source of the mosfet. I then level shifted the DAC output to control the gate source of the added fet .

Answer 3

You can make it theoretically. LTC3773 has opamp and you put a another opamp between IC and resistor. Opamps have high input impedance. You obtain too low current sharing in the circuit and the current resistor has negligible voltage lose.

You try this with LTSpice. You can observe voltage levels of Isnsp-Isnsn pins with opamp and without opamp.

But reality is completely different. PCB design very important. You have to very short length kelvin connections and good design rules. Because, Current resistor differential voltage has too low and noise can be effect largely.

Also, IC-mosfets-opamp should not too high heat etc. If IC and opamp have too heat, they change input impedance. As a result, current sharing increase between IC and opamp hence IC will not work as intended.