Q: What is a topology?
A: A topology is an idealized power conversion circuit. It consists only of power handling inductors (L), capacitors (C), Transistors (Q), and Diodes (P). Pulsewidth modulated square-wave drives are assumed at all transistor gates. Often, a load resistor (R) is included, as is a source voltage (Vg), or, occasionally, a source current generator (Ig).

Q: Why is this done?
A: Much of the functioning of a power conversion stage depends on the topology used in it. Most of the important characteristics of a power converter are topology dependent. These performance characteristics are not easily changed once the topology is chosen. Unlike op-amp circuits, where another stage is easily added if performance leaves something to be desired, power stages come in small stubborn lumps.

Q: Is topological analysis practical, or just a bunch of esoteric theory?
A: See Jerry Foutz's web page for a good discussion of using topological concepts in the real world.

Topology in the Real World


Jerry has worked for many years with the US Navy, and knows what he is talking about!

Q: What does Optimum Topology mean?
A: Originally, it probably meant that the cicuit so named out-performed all others known at the time. Since then, it has come to be understood that optimality is more profound than this.

Q: Like how??
A: In Physics, there are two ways of determining the motion of a system in Newtonian Mechanics. One is to simply write down the forces and masses, and solve the equation.

The other is called the method of least energy, where all systems are found to follow the path of least energy. Thus the motion of a system can be found by minimizing the energy at every point.

By analogy, I believe that the optimality of the Bostbuck topologies represents a sort of performance maximum. Thus the Cuk Converter is, in some sense, the answer to the problem of power conversion; just as 4 is the answer to the problem x+1=5.

It is based on this kind of reasoning that we are skeptical that a better converter will ever be found, at least within the ground rules of switched-mode power conversion!

Q: Which are?
A: Lumped element, switched-mode, non-negligible power processed, feedback controlled, stand-alone conversion, behavior defined by the usual 4 of the 5 terms of Maxwell's equations. This rules out waveguide and antenna effects.

By usual, I mean the KVL/KCL circuit theory used to analyze linear and digital waveform processing circuits. The theory is a low frequency, lumped element approximation of all possible electromagnetic behavior.

Q: Which term is left out?
A: curl B = dE/dt