Most Versatile Voltage Converter you never heard of! The (S)EPIC Converter

I love DC voltage converters like these and so should you. I mean you can get them inexpensively from sites like Amazon or Ebay and they are oftentimes crucial when it comes to powering your electronics projects.

For example, a Buck Converter takes a higher input voltage and converts it efficiently into a stable lower one. While a Boost Converter takes a lower input voltage and converts it into a higher one and if you got a variable input voltage that changes over time then you can simply get yourself a Buck-Boost converter which like the names implies can work as a buck or boost converter and thus delivers a stable output voltage no matter what you got on the input. That is why Buck-Boost converters are in my opinion the most versatile ones. But here is a little secret you might not know and that is that pretty much all small commercial buck-boost converters you find on the internet that come with two coils are in fact so-called SEPIC converters aka Single Ended Primary Inductor Converters.

I know that term sounds complicated but in a nutshell, it is just a special voltage converter circuit topology that actually does not require two coils to function properly. It could theoretically also work with one physical coil that comes with two windings but you never see such designs on Amazon or similar.

So is using one coil instead of two a secret hack that lets this converter work even better? Well; we will find out in this article in which I will not only show you why the SEPIC converter is so versatile and epic but also try out the coil modification and even build a DIY version that you might want to build at home.

Let’s get started! To kick things off I got myself three kinds of commercial SEPIC converters which all came with slight variations concerning their used controllers, inductor values, and so on. If you are asking yourself right now why I am so sure that those are truly SEPIC converters and not like maybe an inverting Buck-Boost converter then let me tell you that I actually desoldered the components from one of them in order to reverse engineer the PCB and thus create this schematic for it. And I think everyone can agree that it is pretty much identical to the SEPIC schematics you can find on the internet.

With that out of the way let’s firstly answer the question of how this thing works and don’t be scared it is not rocket science because there is really only one switch that can either be opened or closed. So let’s imagine the switch was just open and now it is getting closed. But what is noteworthy is that while the switch was open the capacitors here in the middle and at the end got charged up.

Now with the switch being closed, current flows and rises almost linear through inductor 1 because it is building up its magnetic field. While that is going on the middle capacitor discharges itself through inductor 2 whose current flow also rises linearly because it is also building up a magnetic field. And while those two inductors are basically storing energy, the output capacitor supplies the load with current and thus keeps the output voltage stable.

This goes on for a specific amount of time which is of course defined by the frequency and duty cycle of a square wave signal applied to the MOSFET that pretty much always acts as the switching element in such a circuit. And as soon as the signal changes from high to low, the MOSFET turns off and the inductors pump the energy from their magnetic fields in the form of current through the circuit to the load while of course charging up the output capacitor and capacitor in the middle.

Then the control signal once again changes to high and this whole cycle repeats. Now by closing the switch for a longer time and thus increasing the duty cycle, we can basically vary how much energy we pump to the output side and therefore create a variable voltage on the output. But please don’t think that you can easily create super high voltage high power outputs this way. Because as you would expect the amount of energy you can transfer depends on a whole lot more than just the duty cycle like the inductor, capacitor, and diode parameters as well as the switching frequency.

This SEPIC converter topology is so popular because it does not come with inverted outputs like certain other buck-boost converters, it comes with good efficiency, its output can be completely shut off, and its short circuit behavior does not result in complete destruction of the converter unlike some others because the input and output sides are coupled by a capacitor and finally its MOSFET position and overall control feedback operation are the same as with a boost converter which means you can simply take an off the shelve boost converter control IC and easily build this thing.

Now that we understand how the converter works and why it is pretty EPIC; let’s get back to the one inductor instead of two inductors question and to answer that we can simply have a look at the voltage drops across the 2 inductors. As it turns out the voltages across them are the same in both switching positions which means you could wind both coils onto the same core, call it a couple inductor and the circuit should still function the same while apparently reducing AC current losses meaning this modification should increase the efficiency.

By the way, I had to double check this fact as well because this is no longer something you can easily spot. But anyway to practically check whether coupled inductors make a noticeable difference I first had to determine the efficiency of one of my commercial SEPIC converters. So at an input voltage of 12V and an output voltage of 5V it came with an efficiency between 72 to 62% which is OK I guess.

Next, I desoldered its inductors and measured them to find out that they came with all of these beautiful values at a frequency of 52kHz which is, by the way, the frequency this specific controller IC uses. Now at this point I tried finding a commercial coupled inductor with similar values but oh boy that was certainly not an easy task and thus I eventually gave up on that. So what I did instead was firstly winding my own coupled inductor around a toroid core who was not designed for this but in the end featured really promising values.

My second idea was to simply reuse this 100uh power inductor that you can get for cheap from the internet. As you can see one winding consist of 4 wires so by separating them and finding out which one is which we can create two strands and thus get the 4 wires we need for the coupled inductor design. So I soldered those wires to the PCB according to how the schematic demands it shown by those two little points.

After the first power-up, I was very happy to find out that the theory was in fact correct because everything still worked like usual. And after once again measuring a couple of input and output power values I was also happy to find out that the efficiency did in fact increase by around 2 – 4%.

Now using the other DIY coil initially didn’t work quite right but eventually, it decided to cooperate, and with it, I achieved an even higher efficiency. So does that mean that using couple inductors is a secret hack? Well, not really because I cheated by creating way bigger DIY coils in comparison to the commercial coils. But if we trust other experts who created this graph then we can see that there is definitely a slight improvement with coupled inductors.

The only problem is that this difference is so little that most manufacturers do not bother because coupled inductors are simply more expensive. And that basically solves my initial question meaning you can get yourself such a commercial SEPIC converter without hesitation. And If you are not sure which one to get then I would recommend this one since its build on MOSFET comes with very low resistance and thus it guarantees the highest efficiency compared to the SEPIC converter we modified before and the one which is built around the XL6009 controller IC.

Now while looking at those efficiencies I even thought about taking on the challenge of creating my own DIY SEPIC converter that would be more efficient but as it turns out building a high-frequency voltage converter on a preboard results in a ton of noise which makes it pretty much unusable even while only drawing 1A of current.

But on the other hand, designing a circuit around the TL494 PWM controller, IC was really educational and fun. So here is my final schematic in case someone wants to give it a shot or improve the design because as you can see it does in fact work creating a variable output voltage between 2.5 and 20V DC. And with that being said you should now definitely be familiar with the SEPIC converter and thus hopefully learned something new.

Leave a Comment