I have made a schematic for an average current mode controller for car use. It has been designed around an old UC2526, the current loop is compensated around the error amplifier in the PWM chip while the outer voltage loop is compensated around a TL431/opto coupler combo. The transformer has been designed to allow a constant output of +-45V for the entire 10-16V input range, and the overcurrent protection is designed to kick in at around 110A peak on the primary side.
I dont know if I leave the current transformer enough dead-time to self-reset, is a duty cycle of 0.9 too much for this to happen?
Furthermore, is the feedback design sane? the main idea has been to use a little too low amplification in the current loop (22 as opposed to the slope-matching 29.2) to allow for unknowns, and also to let the current loop cut out high enough to allow the voltage loop to see the current loop as constant over its entire operating bandwidth.
I have biased the opto coupler and the current feedback loop around Vref/2 to allow the opto coupler to feel comfortable, is this the way to do it?
Here is the schematic
Please provide input on this as I otherwise will start to make a prototype in a couple of days. I will add thermal guarding and fuses in the real implementation later (also extra LC-filters to separate the switching amplifier channels it will feed)
I dont know if I leave the current transformer enough dead-time to self-reset, is a duty cycle of 0.9 too much for this to happen?
Furthermore, is the feedback design sane? the main idea has been to use a little too low amplification in the current loop (22 as opposed to the slope-matching 29.2) to allow for unknowns, and also to let the current loop cut out high enough to allow the voltage loop to see the current loop as constant over its entire operating bandwidth.
I have biased the opto coupler and the current feedback loop around Vref/2 to allow the opto coupler to feel comfortable, is this the way to do it?
Here is the schematic
Please provide input on this as I otherwise will start to make a prototype in a couple of days. I will add thermal guarding and fuses in the real implementation later (also extra LC-filters to separate the switching amplifier channels it will feed)