Cascaded Switching Regulator Stability

[mirlo]

The input impedance of any moderately efficient switching voltage regulator is negative, because the input current decreases as the input voltage increases.

This creates a problem in power distribution systems that use cascades of regulators. The first regulator has to handle a negative impedance load. Since the output impedance of a regulator is inductive (it goes up with frequency as the control loop dies out) and there is an output capacitor on it, it is easy to get an oscillator: inductor and capacitor and negative resistor in parallel.

What can be done to solve this problem, assuming that there is a necessity to regulate the input of the switching regulators because of requirements of other things connected to their input rail?

-- mirlo

[jackinnj]

higher bandwidth error amplifier -- and of course the switcher is going to want to see a minimal current load (or else the values of the "magnetics" -- both in dollars and physical units, become unrealistic and expensive).

Linear Tech, Maxim and National have application notes specifically addressing the topic. I am sure that TI does as well.