Where are Ćuk converters

Question

According to "the art of electronics" (p 651) and other internet sources, the ćuk switch converter has the most desirable properties: continuous input and output current waveforms, boosting and bucking, and "zero-output ripple" current. There are plenty of switch converters sold in any store like eBay (e.g.:this one), but I found nowhere a ćuk converter at such a low price (and even hardly at high price). Is there a problem with them that makes them unpopular? Why aren't they sold ?

Answer 1

There are a few reasons.

1. Polarity

   Ćuk converters must always have opposite input and output polarity. This makes them somewhat inflexible and would prevent them from being used in any application that simply wanted to reduce/buck or raise/boost the input to output voltage. The only way a Ćuk converter could do what the buck converter you linked does is by completely isolating the input from the output. This significantly increases the complexity, cost, and size of the already more complex Ćuk converter, while having a slightly negative impact on performance. Once all of that is factored in, the theoretical advantage of such a topology is not nearly as compelling.

2. Control

   Ćuk converters are very difficult to control and keep stable over load ranges, and most importantly, cannot be controlled using a PWM controller. Ćuk converters are 4th order systems and compensating them difficult, controlling them even more so. The buck converter you linked uses a fairly jellybean part, one of the UC2/384*current mode PWM controller chips. These parts are cheap, established, and can be used to build a buck, boost, flyback, forward, and other topologies, as can any other PWM controller chip. A PWM controller cannot control 4th order topologies like the Ćuk or SEPIC. As PWM control is simple and works, simply by that quality, Ćuk converters will always be much less widespread than other, simpler topologies.

3. Regulation

   The above two points certainly contribute, but this is definitely the primary reason Ćuk converters have never seen any sort of large-scale commercial adoption: transient response and regulation. I love looking at switching topologies both from a practical and theoretical standpoint, and I get that r Ćuk converters seemingly appear to be the perfection or logical conclusion of the other topologies. Any performance improvements applicable to the simpler topologies can just as eLargely due to a Ćuk topology. It really just seems like the thing we should be using, doesn't it?

   And we should be using it, and would be, if it were only as simple as being a great switching topology. The most important thing being overlooked here is when it comes to actual products, converters, devices, that people need and buy, they aren't buying a switching topology. They aren't even buying a switching converter specifically, it just happens that switching converters are a good fit for their needs. They're buying a voltage regulator. Or regulated power supply if you prefer. They want a certain voltage and just that voltage. Sure, there are adjustable converters out there, or even arguably most are adjustable, but you're still wanting a fixed output voltage. You can just change what it is. But once you've set it to, say, 7V, you want it to stay at 7V and not randomly deviate 2V from that unless you told it to.

   Voltage regulation and transient response go hand in hand, and are really the same thing. As a load changes, to maintain a fixed output voltage, the converter must respond appropriately, but no converter can do this instantaneously. There will always be a small amount of voltage sag depending on how quickly and how much the load changes. The faster the transient response, the better the regulation. This applies to the input as well. The input may vary, and to stay in regulation, the converter again must respond appropriately.

   And while the Ćuk is a frankly downright beautiful switching topology, it's not a very good voltage regulator. It is slow, and will always be slower than simpler topologies due to its higher order nature. No matter how good the transient response of a Ćuk is, it will be significantly poorer than simpler, lower order topologies. You can understand this like the settling time of a filter. The higher the order, the longer the settling time. Compound this with the other difficulties in controlling them, and what you find is that for decent amount of power, a Ćuk converter has such poor line and load regulation, as well as transient response, that you have to use it as preregulator only, and use linear post-regulation on the output to get a voltage stable enough to compete on the market. And at that point, you're burning up all the advantage as heat in a linear regulator. While depressing, it's the reality of things. Well, it's depressing to me, I'm a bit of a Ćuk fan.

   Sure, there are times one doesn't need voltage regulation, or can tolerate loose regulation, and the Ćuk topology is inferior to others in that space. Nonregulated converters are ratiometric converters (they behave akin to a linear transformer, they multiply the input by a ratio to get the output. a 2:1 converter simply yields half the input voltage at the output, without regulating it). In the ratiometric space, SAC (sine amplitude converters) absolutely dominate. They are virtually EMI free, have efficiencies greater than 98%, and have staggering power density (current commercially available units achieve 2870W/in3. Yes, you read right, almost 3kW per cubic inch. ~140W/in3 is considered quite good for regular DC/DC converters. Multiphase high end ones can achieve maybe 400-500W/in3). A Ćuk is still a hard-switched topology, and softswitched ones will always be superior, to hardswitched ones, so the Ćuk is out of its element here unfortunately.

4. Patents

   Another minor reason that contributed to the absolute dearth of Ćuk topologies in use today (but remember, the earlier reasons would have, and will continue to prevent it from ever seeing widespread use - this reason simply made it even more so) is that the topology was patented, and future improvements on it have also been patented. Dr. Slobodan Ćuk's original patents expired in the early 90s I believe, but subsequent ones by other parties have not.