I have a SEPIC supply with coupled inductors. The inductors have phase dots. What do these "phase dots" mean? Is it important to have the phase dot in the same place?
If we're thinking about same thing, then dots show relation between coils. Here's an article about that.
If I get the dot relation correctly (and I'm not sure that I do), if current is going into the dot on one side of the inductor and the other dot is on the same side of the inductor, that means that current is going into the dot on the other side too.
How important it is to have dots in the same place is up to design of the circuit. In some cases it may not be relevant, but often it is important to take it into account.
The dots just indicate the polarity of the windings on the schematic.
Currents entering on the dotted ends of the windings will produce magnetic flux in the same direction, whereas if you have current entering one dotted end, and leaving another dotted end, the currents will produce opposing flux.
If you're looking at something like a power supply schematic, the dots show you the ends of the coils that have the same phase angle.
Normally we apply AC voltage on the (input) primary coil of the transformer and we have an (approximately) resistive load on each (output) secondary coil of the transformer.
In that case, the dotted end of every coil will reach the positive peak voltage relative to the non-dotted end of that coil at (approximately) the same time.
Also at that same instant current will be flowing into the dotted end of the primary coil and current will be flowing out of the dotted end of each secondary coil.
The dots describe which way each coil was wound. If I take a CAT5 cable and wind it around a ferrite core, each wire-end at one end of the CAT5 cable is dotted. Each wire-end at the other end of the cable is not-dotted.
I honestly don't know if dot orientation makes any difference for a SEPIC converter.
I know that the dots are important in the very similar "Coupled Inductor Cuk Converter" and "Integrated Magnetics Cuk Converter" ( The Four Topologies ). If one accidentally swaps the two ends of a coil in that converter, you get increased (worse) ripple on the input or output or both.
I think I can help.
The dot indicates the outside wire of the inductor, particularly for inductors that are wound on a ferrite spool or bobbin. This can be useful for minimizing unwanted emissions. For example, if you are designing a switching converter (buck, boost, or SEPIC), one end of the inductor(s) is normally tied to a DC power rail while the other end is switched up and down. If you tie the "dot" end of the inductor to the DC rail, the inner windings of the inductor will be doing most of the voltage swinging, and the outer layers will help shield them from radiating.
In a SEPIC converter, inductor coupling is optional. The only way to couple the inductors is to wind them on the same core. If you're buying single inductors off-the-shelf, your design is UNcoupled and you want to make sure the inductors do not interfere with one another. For uncoupled inductors, tie the dotted end of the series inductor to the input voltage and tie the dotted end of the parallel inductor to ground. That will minimize emissions and unwanted coupling between the two inductors.
Film-type capacitors are marked the same way. Film capacitors have no voltage polarity restrictions (unlike electrolytics and tantalums), but they still have a polarity mark on them. Film capacitors are made by wrapping foil and insulating film (or metallized film) like a roll of toilet paper, and the dot indicates the outer layer of that winding. Tie the dotted end to DC, and the capacitor is partially shelf-shielding.
Wurth indicates the dot is the start of the winding (which makes it the inside winding) and therefore the dot should be placed at the switching node for the reasons Marc mentioned above. I don't know if there's a convention for this, but if there isn't, I would ask the manufacturer what the dot indicates when used with switchers.