Hello AAC Community,
I would like to design a feedback control system for a DC/DC converter and need some hints for guiding me in the right direction as I’ve never done something like this before.
Background and what I know:
I am designing a step-up push-pull DC/DC converter feeding a particular load, and have already done some research into the kinds of integrated gate drive PWM controllers available on the market. Examples include the LM5030 and the LM25037 which I'm looking into. I have also read a bit about opto-isolated feedback circuits involving a reference regulator (such as the TL431, or differential op amp with offset reference) and an optocoupler to feed back into a “COMP” pin on the controller, which sets a DC voltage to compare against a triangle carrier to adjust gate PWM. The integration of opto-isolated output voltage feedback, with a current sense resistor on the primary source, and one of these controllers seems to provide a full picture of successfully designing such a push-pull converter.
Where I need your advice:
How do I go about designing a good feedback compensation network? Assuming a type-II isolated error amplifier, it appears to me as though I would need to derive a dynamic model of a push-pull converter (with an full bridge inverter as a load!) in order to determine the zeros (RHZP) and poles of the system, and then design the Rcomp and Ccomp compensation network to compensate the particular transfer function in question. Is my thinking correct or am I missing something? What is standard design practice for SMPS engineers in such a situation? The only modeling software I have available is TINA by TI, LTSpice, and of course Excel. Unfortunately I don’t have things like MATLAB or PLECS (although I can try to get it). I haven’t derived a transfer function for a push-pull converter yet, but am considering it if it would be necessary. Surprisingly not much “good” info online as far as dynamic modeling for push-pull converters are concerned; although I can imagine these dynamic state-space equations are a mess.
Even more so, how would I determine what are acceptable phase and gain margins on a bode plot? Further, how would you recommend testing this in real-life? I have an HP 33120A function generator, a MSO-X 3104A oscilloscope, and some power supplies kicking around to work with.
P.S. My load is a full bridge converter that takes the output from the push-pull and SPWM’s it through an LC filter onto a mostly resistive load with some capacitance.
Thank you,
I would like to design a feedback control system for a DC/DC converter and need some hints for guiding me in the right direction as I’ve never done something like this before.
Background and what I know:
I am designing a step-up push-pull DC/DC converter feeding a particular load, and have already done some research into the kinds of integrated gate drive PWM controllers available on the market. Examples include the LM5030 and the LM25037 which I'm looking into. I have also read a bit about opto-isolated feedback circuits involving a reference regulator (such as the TL431, or differential op amp with offset reference) and an optocoupler to feed back into a “COMP” pin on the controller, which sets a DC voltage to compare against a triangle carrier to adjust gate PWM. The integration of opto-isolated output voltage feedback, with a current sense resistor on the primary source, and one of these controllers seems to provide a full picture of successfully designing such a push-pull converter.
Where I need your advice:
How do I go about designing a good feedback compensation network? Assuming a type-II isolated error amplifier, it appears to me as though I would need to derive a dynamic model of a push-pull converter (with an full bridge inverter as a load!) in order to determine the zeros (RHZP) and poles of the system, and then design the Rcomp and Ccomp compensation network to compensate the particular transfer function in question. Is my thinking correct or am I missing something? What is standard design practice for SMPS engineers in such a situation? The only modeling software I have available is TINA by TI, LTSpice, and of course Excel. Unfortunately I don’t have things like MATLAB or PLECS (although I can try to get it). I haven’t derived a transfer function for a push-pull converter yet, but am considering it if it would be necessary. Surprisingly not much “good” info online as far as dynamic modeling for push-pull converters are concerned; although I can imagine these dynamic state-space equations are a mess.
Even more so, how would I determine what are acceptable phase and gain margins on a bode plot? Further, how would you recommend testing this in real-life? I have an HP 33120A function generator, a MSO-X 3104A oscilloscope, and some power supplies kicking around to work with.
P.S. My load is a full bridge converter that takes the output from the push-pull and SPWM’s it through an LC filter onto a mostly resistive load with some capacitance.
Thank you,
