Transformers in high frequency switched mode converters are commonly made from ferite or (alloys of) powdered iron. Examples of ferrite materials for this task include EPCOS/TDK N27, N87 and N95 (https://www.tdk-electronics.tdk.com/en/ferrites); Mag Inc. F material (https://www.mag-inc.com/) and Feroxcube 3C90 and 3C95 (https://ferroxcube.com/) among many others. Powdered iron and alloys of iron include Sendust and newer materials such as HighFlux both of which are used by several producers including MicroMetals (https://micrometals.com/).
When we design a transformer (excluding coupled inductors in which energy is stored in the magnetic field for an appreciable time before being released) the received wisdom is that no gap is required because there is no desire for energy storage – this idea goes hand in hand with the idea that energy is “stored in the gap”. However, there are some good arguments for including a gap.
First Argument – Degree of Freedom in choice of Inductance(s) and Core Flux Density
Choosing the number of turns is generally about setting the flux density. This may seem counterintuitive as the number of turns and the inductance are bound together by the AL value of the core. However in a so-called “magnetically limited” design The flux density is chosen based on the losses that the designer intends to sustain. These are, in turn, driven either by the loss budget for the converter (the efficiency required) or by the thermal constraints of the design (the temperature rise we can accept on this part). So the number of turns is not really about setting the inductance of the winding, it’s about controlling the flux density in the core.
If the ferrite material is fixed and the core cross sectional area (Ae) is fixed (i.e. we have chosen a core size/shape and a ferrite material) then the inductance of the winding is effectively decided as well. This situation is not unlikely as calculating the number of turns generally occurs after these choices (core and flux density) have been made. So, if we hold everything fixed then the flux density and the inductance of the winding go up and down together with the number of turns. This is not desirable as the designer generally wants to set the inductance of the winding as well as the flux density and the turns ratio.
The designer resolves this difficulty by adding a degree of freedom in their design, by adding a gap. The size of the gap is chosen such that the inductance and the flux density targets are both satisfied simultaneously. It should be remembered that the gap(s) will have fringing effects so the gap must be modified accordingly to account for the effects of the fringing flux which act to increase the inductance of a winding.
Second Argument – To control the spread of AL values and so improve manufacturability
The effect of adding a gap is to cause the reluctance of the magnetic path to be dominated by the gap reluctance. Even a small gap will quickly provide significant reluctance compared to the core because ferrites tend to have high permeability. Provided the gap reluctance is approximately equal to or greater than the core reluctance, the core reluctance will not play a significant part in determining the effective AL (or the effective permeability if you prefer) of the total magnetic pathway. We can control the gap with reasonably good tolerances by having the cores machined to size (generally 50 um is the minimum gap to consider viable, double that is easy. The surface roughness between two cores which have been ground (machined) but not polished is about 20 um or so. Since the gap reluctance dominates the total reluctance, the inductance formed by a certain number of turns varies less with changes in the core reluctance (which is driven by batch to batch changes in the permeability of the ferrite). Therefore inserting a small gap provides a method to ensure that the design inductance has a closer tolerance specification over a batch of parts. This is desirable as converters should operate closer to the design ideal. The spread of AL in ungapped cores can routinely be as much as 20%.
Design Support from our resident Magnetics Designers
Bear Instruments provides attractive magnetic component solutions including full design documentation, production of samples, validation and reporting as well as mass production and distribution via our selected partners. Call or email to find out how we can push your product development forward.
