Eight Common Design Mistakes of High Frequency Magnetic Components in Switching Power Supply
1) Filling the window of magnetic core-optimized design
Many power supply designers think that in the design of high-frequency magnetic components, the best design can be obtained by filling the core window, but it is not. In the design of many high-frequency transformers and inductors, we can find that adding one or more layers of windings, or using enameled wires with larger wire diameters, not only can't get the optimal effect, but will increase the total winding loss because of the proximity effect in winding.
Therefore, in the design of high-frequency magnetic components, it doesn't matter even if the winding doesn't completely wrap the iron core window, but only wraps 25% of the window area. You don't have to try to fill the whole window area.
This misconception is mainly influenced by the design of power frequency magnetic components. In the design of power frequency transformer, the integrity of core and winding is emphasized, so there is no gap between core and winding, and the winding is generally designed to fill the whole window, thus ensuring its mechanical stability. However, the design of high-frequency magnetic components does not have this requirement.
2) "iron loss = copper loss"-optimized transformer design
Many power designers, even in many reference books for magnetic component design, list "iron loss = copper loss" as one of the criteria for optimal design of high-frequency transformers, but it is not. In the design of high-frequency transformer, the difference between iron loss and copper loss can be large, and sometimes the difference can even reach an order of magnitude, but this does not mean that the high-frequency transformer is not well designed.
This misconception is also influenced by the design of power frequency transformer. Power frequency transformers often occupy a large area because of the large number of windings, so from the perspective of thermal stability and thermal uniformity, the empirical design rule of "iron loss = copper loss" is obtained.
However, for high-frequency transformers, this rule of thumb does not hold. In the design of switching power supply high-frequency transformer, there are many factors to determine the optimal design, and "iron loss = copper loss" is actually the least concerned aspect.
3) Magnetized inductance with leakage inductance = 1%
After designing the magnetic components, many power supply designers often explain the leakage inductance requirements when submitting the relevant technical requirements to the transformer manufacturers. Many technical sheets are marked with similar technical requirements such as "magnetizing inductance with leakage inductance = 1%" or "magnetizing inductance with leakage inductance < 2%". In fact, this kind of writing or design standard is very unprofessional.
The power supply designer should set a numerical limit on the acceptable leakage inductance according to the normal working requirements of the circuit. In the process of making transformer, the leakage inductance should be reduced as much as possible without deteriorating other parameters of transformer (such as turn-to-turn capacitance), instead of giving the proportional relationship between leakage inductance and magnetizing inductance as a technical requirement.
Because the relationship between leakage inductance and magnetization inductance varies greatly with the presence or absence of air gap in transformer. When there is no air gap, the leakage inductance may be less than 0.1% of the magnetizing inductance, while when there is an air gap, even if the transformer windings are closely coupled, the proportional relationship between the leakage inductance and the magnetizing inductance may reach 10%.
Therefore, the proportional relationship between leakage inductance and magnetizing inductance should not be provided to the manufacturer of magnetic components as the design index of transformer. Otherwise, it will show that you don't understand the leakage knowledge or really care about the actual leakage value. The correct way is to specify the absolute value of acceptable leakage inductance. Of course, a certain proportion can be added or subtracted, and the typical value of this proportion is 20%.
4) Leakage inductance is related to magnetic core permeability.
Some power supply designers believe that adding a magnetic core to the windings will make the windings more closely coupled and reduce the leakage inductance between windings; Some power supply designers think that the magnetic core will be coupled with the field between windings after adding the magnetic core to the windings, which can increase the leakage inductance.
In fact, in the design of switching power supply, the leakage inductance of two coaxial winding transformers has nothing to do with the existence of magnetic cores. This result may be incomprehensible, because a material with a relative permeability of several thousand has little effect on leakage inductance when it is close to the coil.
The measured results of hundreds of transformers show that the change of leakage inductance is basically not more than 10% with or without magnetic core, and many changes are only about 2%.
5) The optimal value of current density of transformer winding is 2A/mm ~ 3.1A/mm.
Many power supply designers often regard the current density in the winding as the standard of optimal design when designing high frequency magnetic components.
In fact, the optimal design has nothing to do with the winding current density. What really matters is how much loss there is in the winding and whether the heat dissipation measures are enough to ensure the temperature rise within the allowable range.
We can imagine two extreme cases of heat dissipation measures in switching power supply. When liquid immersion and vacuum are used for heat dissipation respectively, the corresponding current density in winding will be quite different.
In the actual development of switching power supply, we don't care about the current density, but only how hot the wire package is. Is the temperature rise acceptable?
This erroneous concept is that designers simplify the number of variables and thus simplify the calculation process in order to avoid tedious repeated trial and error, but this simplification does not explain the application conditions.
6), primary winding loss = secondary winding loss "-optimized transformer design.
Many power supply designers believe that the optimized transformer design corresponds to that the primary winding loss of the transformer is equal to the secondary winding loss. Even in many design books of magnetic components, this is regarded as a standard for optimal design. In fact, this is not a standard for optimal design.
In some cases, the iron loss and copper loss of transformer may be similar. But it doesn't matter much if there is a big difference between the primary winding loss and the secondary winding loss.
It must be emphasized again that what we are concerned about in the design of high-frequency magnetic components is how hot the winding is under the heat dissipation mode used. Primary winding loss = secondary winding loss is only an empirical rule in the design of power frequency transformer.
7) If the winding diameter is less than the penetration depth, the high-frequency loss will be very small.
Just because the winding diameter is less than the penetration depth does not mean that there is no great high frequency loss. If there are many layers in the transformer winding, even if the wire diameter is much thinner than the penetration depth, it may cause great high frequency loss because of strong proximity effect.
Therefore, when considering the winding loss, we should not only judge the loss from the thickness of enameled wire, but also comprehensively consider the arrangement of the whole winding structure, including winding mode, winding layers and winding thickness.
8) The open-circuit resonance frequency of transformer in forward circuit must be much higher than the switching frequency.
Many power supply designers think that the open-circuit resonance frequency of transformer must be much higher than the switching frequency of converter when designing and testing transformer. In fact, the open-circuit resonance frequency of transformer has nothing to do with the switching frequency.
We can imagine the limit case: for an ideal magnetic core, its inductance is infinite, but there will also be a relatively small turn-to-turn capacitance, and its resonance frequency is approximately zero, which is much smaller than the switching frequency.
What is really related to the circuit is the short-circuit resonance frequency of the transformer. In general, the short-circuit resonance frequency of transformer should be more than two orders of magnitude of switching frequency.
