Engineers: How to improve power efficiency

If you want to design a good transformer, you first need to choose a good transformer. The choice of transformer is subject to many factors, which I have mentioned many times in many articles before, and I will repeat it here again.

First, you need to calculate the Ap value of the transformer. There are many related posts on the forum about the calculation method. You can search for it. I will not go into details here. After getting the Ap value, we need to preliminarily select the transformer according to the structural dimensions of the power supply, including the height, width and length of the transformer. When the overall height of the power supply is limited, you need to consider a flat transformer, and a horizontal transformer is the first choice. Common horizontal transformers include EE series, EC series, ER series, EF series and EFD series transformers; if it is an ultra-thin adapter and LED fluorescent lamp built-in power supply, you can consider a flat transformer.

Secondly, when choosing a transformer, we need to choose different transformers according to the parameters and focus of the circuit . For example, in a flyback power supply, we hope that the leakage inductance is as small as possible, because the leakage inductance will affect the voltage and current stress of the power device, and also have an important impact on EMC. Then we look for transformers that are good for leakage inductance control, such as PQ type, RM type, and ERL type transformers, and with a reasonable winding method, the leakage inductance can be controlled below 3%. For another example, in an LLC power supply, we hope to use the leakage inductance of the transformer as the resonant inductor, so we need to deliberately increase the leakage inductance, and it is ideal to use a slotted skeleton for winding.

Once again, when choosing a transformer, you need to consider cost and versatility. Cost is not only a concern for every business owner, but also the most entangled issue for our R&D engineers . Unless it is a few military-grade or high-end power supplies that do not consider cost, we must find a balance between performance parameters and costs when designing. Do not deliberately pursue a certain parameter and ignore the cost impact. Sometimes even if the cost of each transformer increases by a few cents, if it is mass-produced, it is an expense that cannot be ignored. Unless due to commercial considerations, you hope that your products will not be copied by other manufacturers, and generally do not consider private molds or unconventional transformer cores and skeletons, because when mass-producing, the supply channels and cycles will be greatly restricted, and universal cores have a lot of options in terms of price, supply channels and cycles.

Picture display:

When choosing a transformer, you must also consider EMC performance in order to comply with safety standards:

First of all, we need to consider the winding width of the transformer bobbin. In order to meet the creepage distance requirements in safety regulations, a 3mm retaining wall is generally added to the side of the winding, which reduces the available winding width of the transformer bobbin. If no retaining wall is added, triple insulated wire will be required, and the outer diameter of the triple insulated wire is generally 0.2mm larger than the internal copper wire diameter. Therefore, for the same window area, the number of winding turns is reduced.

Secondly, the slot depth of the transformer frame must be considered. Sometimes, for EMC reasons, it is necessary to add a shielding layer inside the transformer. Some are wound with thin wires, and some are wound with copper foil. These windings will undoubtedly increase the number of winding layers, which means that the slot depth that can be used to wind other windings of the transformer is reduced.

When selecting a transformer, the influence of the winding assembly process must also be considered.

Many engineers do not consider the assembly process when designing transformers. This often happens: after the transformer is calculated, the parameters are sent to the transformer factory for sampling; then, the transformer factory engineer calls and says that the winding is too tight, which makes it difficult to assemble and is not conducive to mass production; finally, the transformer parameters have to be modified; this will undoubtedly delay the progress of the project. Therefore, at the beginning of the design, we must consider the error of the transformer core window, as well as the winding process, the thickness of the insulating TAPE and other factors, which will affect the assembly of the transformer; we should give full consideration to these factors when calculating, leaving a certain margin.

The above discussed the issues to be considered in selecting the transformer's core frame. Now let's talk about the transformer's winding method and precautions.

Ordinary layered winding method:

In a general single output power supply , the transformer is divided into three windings, the primary winding Np, the secondary winding Ns, and the auxiliary power winding Nb; when the ordinary layered winding method is used, the winding order is: Np--Ns--Nb. Of course, there are also some that use the winding method of Nb--Ns--Np, but it is not commonly used. You can think about the reason first, and I will analyze it in a few days.

This winding method has a simple process, is easy to control various parameters of the magnetic core, has good consistency, and has low winding cost. It is suitable for mass production, but the leakage inductance is slightly larger. Therefore, it is suitable for low-power occasions that are not sensitive to leakage inductance. This winding method is generally used in power supplies with a power of less than 10W.

Sandwich wrap

The sandwich winding method has a long reputation. Almost everyone who makes power supplies knows this winding method, but there are not many people who have done in-depth research on the sandwich winding method. I believe that many people have eaten sandwiches, which are two layers of bread with a layer of butter in between. As the name suggests, the sandwich winding method is a winding method where two layers are sandwiched with one layer. Due to the different windings sandwiched in the middle, the sandwich winding method is divided into two types: primary sandwiched with secondary, and secondary sandwiched with primary.

Let's first look at the first method, the primary-secondary winding method (also called primary average winding method).

As shown in the figure above, the order is Np/2, Ns, Np/2, Nb. This winding method has the advantage of large quantity.

Since the effective coupling area between the primary and secondary is increased, the leakage inductance of the transformer can be greatly reduced, and the benefits of reducing leakage inductance are obvious: the voltage spike caused by leakage inductance will be reduced, which will reduce the voltage stress of MOSFET . At the same time, the common mode interference current caused by MOSFET and heat sink can also be reduced, thereby improving EMI.

Since a secondary winding is added in the middle of the primary, the interlayer distributed capacitance of the transformer primary is reduced . The reduction of interlayer capacitance will reduce parasitic oscillation in the circuit , and can also reduce the voltage and current stress of MOSFET and secondary rectifier tube, and improve EMI.

The second method is the secondary clamp primary winding method (also called secondary average winding method).

As shown in the figure above, the order is Ns/2, Np, Ns/2, Nb. When the output is low voltage and high current , this winding method is generally used. It has two advantages:

1. It can effectively reduce the temperature rise caused by copper loss: Since the output is low voltage and high current, the copper loss is more sensitive to the length of the wire. The Ns/2 wound on the inside can effectively reduce the winding length, thereby reducing the copper loss and heat generation of this Ns/2 winding. Although the outer Ns/2 winding is relatively long, it is basically on the outer layer of the transformer, and the heat dissipation is good, so the temperature will not be too high.

2. It can reduce the high-frequency interference from the primary coupling to the transformer core. Since the primary is far away from the core and the secondary voltage is low, the high-frequency interference caused is small.

This is the Vds waveform of the MOSFET at full load with 220V input .

This is the Vds waveform of the MOSFET at 260V input.

Next, let's discuss in depth the impact of this sandwich winding on EMI.

First, let’s look at how to wind the primary clamped by the secondary.

We know that the transformer primary has a higher voltage , so it has more windings, usually more than 2 layers, sometimes even 4-5 layers, which brings a distributed parameter to the transformer --- interlayer capacitance . I believe everyone is clear about the formation principle, so I will not explain it in detail.

When the MOSFET is turned off, the leakage inductance of the transformer, the junction capacitance of the MOSFET and the interlayer capacitance of the transformer will vibrate with an amplitude of tens or even more than one hundred V, which is not allowed for MOSFET and EMI. Therefore, we increase the RCD absorption to suppress this oscillation to protect the MOSFET and improve EMI.

The figure above is the Vds waveform of the flyback power MOSFET.

From this perspective, the sandwich winding method can improve EMI to a certain extent. From another perspective, the sandwich winding method does increase the primary-secondary coupling area, reduce the leakage inductance, and at the same time increase the primary-secondary coupling capacitance; when the switch tube is repeatedly switched, the capacitance will also be repeatedly charged and discharged, which means that it will cause oscillation. This oscillation is proportional to the switching frequency and will have an adverse effect on EMI. ( Please indicate the source for reprinting from the power supply network)