Faraday Screens in Switching-Mode Power Supplies (SMPS)

INTRODUCTION

One of the most difficult problems in switching-mode power supply design is to reduce the common-mode conducted RFI current to acceptable limits. This conducted electrical noise problem is mainly caused by parasitic electrostatic and electromagnetic coupling between the various switching elements and the ground plane. (The ground plane can be the chassis, cabinet, or ground return wire, depending on the type of unit.) The designer should examine the layout, identify the areas where such problems may exist, and introduce at the design stage the correct screening methods, It is very difficult to correct for poor RFI design practices at a later stage.

Likely problem areas are shown in Img. .1, a diagram of the typical problem areas for parasitic coupling in a flyback SMPS. Suitable locations for Faraday screens are shown.

In most applications, Faraday screens will be required where high-frequency, high-voltage switching waveforms can capacitively couple to the ground plane or secondary outputs. Typical positions would be where switching transistors and rectifier diodes are mounted on heat sinks which are in contact with the main chassis. Further, where components or wires carry large switching currents, noise can be coupled by both magnetic and capacitive coupling. Other likely problem areas are output rectifiers; output chassis-mounted capacitors; and capacitive coupling between the primary, secondary, and core of the main switching transformer and any other drive or control transformers.

FARADAY SCREENS AS APPLIED TO SWITCHING DEVICES

When components are mounted on heat sinks which are to be thermally linked to the chassis, the normal way of eliminating undesirable capacitive coupling is to place an electrostatic screen between the offending component and the heat sink.

This screen, normally copper, must be insulated from both the heat sink and the transistor or diode, so that it picks up the capacitively coupled ac currents and returns them to a convenient "star" point on the input circuit. For the primary components, the "star" point will usually be the common negative DC supply line, close to the switching device. For secondary components, the "star" point will normally be the common return to the transformer. IMG. .1 demonstrates the principle.

IMG. .1 Return paths for capacitively coupled Faraday screen currents in primary and secondary circuits.

One example of a TO3 transistor Faraday screen is shown in Img. .2. The primary switching transistor, with its high voltage and high-frequency switching waveform, would couple a significant noise current through the capacitance between the transistor case and the main chassis unless a screen is fitted between them. In the mounting arrangement shown in Img. .2, the copper screen will return this parasitic noise current to the input circuit, thus completing the current loop without introducing current into the ground plane. The screen will not inject any significant current through the capacitance to the heat sink, because it has a relatively small high-frequency ac voltage relative to the chassis or ground plane.

The designer may identify other areas where problems can occur; in that event, similar screening should be used.

IMG. .2 Insulated Faraday screen, positioned between TO3 switching transistor and heat sink.

TRANSFORMER FARADAY SCREENS AND SAFETY SCREENS

To prevent circulation of RF currents between the primary and secondary windings or between the primary and the grounded safety screen, the main switching transformer will usually have at least one RFI Faraday screen in the primary winding. In some applications, an additional safety screen will be required between the primary and secondary windings. There are major differences in construction, location, and connection between the Faraday RFI screens and the safety screens. Safety regulations require that the safety screens be returned to the ground plane or chassis, whereas RFI screens will normally be returned to the input or output circuits. The EMI screens and connections may be made of very lightweight copper, as they carry very little current. However, for safety reasons, the safety screen must be rated for a current of at least three times the supply fuse rating.

IMG. .3 shows the typical arrangement of safety and RFI screens in a switching-mode transformer for "off-line" use. In the fully screened application shown, the two RFI screens will be adjacent to the primary and secondary windings, and the safety screen will be between the two RFI screens. If secondary RFI screens are not required, the safety screen will be between the primary RFI screen and any output windings. As a further insulation precaution, the primary RFI screen may be DC isolated from the input power lines by a series capacitor. (A value of 0.01 µF at the rated isolation voltage is usually sufficient.)

IMG. .3 Fully screened transformer, showing positions and connections of primary and secondary Faraday screens, with an additional primary-to-secondary safety screen.

The RFI screen shown on the secondary side is fitted only when maximum noise rejection is required or when output voltages are high. This screen would be returned to the common output line. Transformer screens should be fitted only when essential, as the increased buildup and winding height increase the leakage inductance and degrade the performance.

To prevent the high-frequency screen return currents (which can be considerable during the switching transient) from coupling to the secondary by normal transformer action, the screen connections should be made to the center of the screen, rather than one end. In this way, the capacitively coupled screen return currents flow in opposite directions around each half of the screen, cancelling any inductive coupling effects. Remember, the ends of the screen must be insulated to prevent a closed turn.

FARADAY SCREENS ON OUTPUT COMPONENTS

For high-voltage outputs, RFI screens may be fitted between the output rectifiers and their heat sinks. If the secondary voltages are small, say 12 V or less, the secondary transformer RFI screen and rectifier screens should not be required.

The need for Faraday screens on output rectifier diodes can sometimes be eliminated by making the diode heat sink dead to RF voltages by putting the output filter choke in the return line. Typical examples are shown in Img. .4a and b.

If the diode and transistor heat sinks are completely isolated from the chassis (for example, mounted on the pcb), Faraday screens are unlikely to be required on these components.

REDUCING RADIATED EMI IN GAPPED TRANSFORMER CORES

Ferrite flyback transformers and high-frequency inductors will usually have a relatively large air gap in the magnetic path, to define the inductance or to prevent saturation.

Considerable energy can be stored in the magnetic field associated with this air gap.

Unless the transformer or choke is screened, an electromagnetic field (EMI) will be radiated from the gap, and this can cause interference to the supply itself or to local equipment. Further, this radiated field may exceed the radiated EM1 limits.

The largest field radiation will occur with cores that have a gap in the outer limbs or a gap that's equally distributed across the pole pieces. This radiation may be reduced by a factor of 6 dB or more by concentrating the air gap in the center pole only. With totally enclosed pot cores, the reduction in radiation by using only a center pole gap would be much greater. However, for off-line applications, the pot core is not often used because the creepage distance requirements at the higher voltages usually can't be satisfied.

Concentrating the air gap in the center pole alone increases the temperature rise and reduces efficiency. This increased loss is probably due to magnetic fringe effects at the edge of the pole pieces in the center of the winding. The disturbance of the magnetic field within the windings results in additional skin and eddy-current losses, and a further reduction in efficiency of up to 2%. Also, the increased losses in the region of the gap can cause a hot spot and premature failure of the insulation in this area.

In cores which are gapped in the outer legs, the addition of a copper screen around the outside of the transformer gives a considerable reduction in radiation. IMG. .5 shows a typical example.

IMG. .4 A method of reducing parasitic RFI currents in chassis-mounted output diode heat sinks by fitting the output choke in the common return line. (a) Push-pull applications; (b) single-ended outputs.

IMG. .5 Copper screen fitted to a switching transformer, to reduce RFI and EMI radiation. (Note: Screen goes around the outside of the core legs.)

This screen should be a totally closed loop around the outside of the transformer, over the outer limbs and windings, and centered on the air gap. The width of the screen should be approximately 30% of the width of the bobbin and should be in the same plane as the windings. To be effective, it must have minimum resistance; a copper screen with a thickness of at least 0.010 in is recommended.

It would appear that this screen is effective because of both eddy-current losses and the action of the closed loop. The current induced in the closed loop will generate a back MMF to oppose radiation. In flyback transformers, the screen should not be more than 30% of the bobbin width, as problems of core saturation have been observed with wide screens. Although the screen is normally used for cores which are gapped in the outer legs, it will be effective for transformers with a gap in either the center pole or the outer legs. In either case, there will be a reduction in magnetic radiation of up to 12 dB.

However the application of a transformer screen results in lower transformer efficiency. This is due to the additional power losses in the screen, caused by eddy current heating effects. If the air gap is in the outer poles, the power loss in the screen may amount to as much as 1% of the rated output power, depending on the size of the air gap and the power rating of the unit. For applications in which the air gap is in the center pole only, there will be little further increase in power loss from fitting a screen. However, the overall transformer efficiency is about the same in both cases, as the center pole gap increases the losses in the transformer windings by about the same amount.

It would seem that effective magnetic screening of the transformer can be applied only at the expense of additional power losses. Consequently, such screening should be used only where essential. In many cases, the power supply or host equipment will have a metal enclosure so that EM1 requirements will be met without the need for extra transformer screening. When open-frame switching units are used in video display terminals, screening of the transformer will usually be required to prevent interference with the display by magnetic coupling to the CRT beam. The additional heat generated by the outer copper screen may be conducted away using a heat sink or a thermal shunt from the screen to the chassis. IMG. .5 shows a typical example of a copper EM1 screen as applied to an E core transformer with air gaps in the outer legs.

QUIZ:

• Why are Faraday screens so effective in reducing common-mode interference in high-voltage switching devices and transformers?
• What is a line impedance stabilization network (LISN)?
• What is the difference between common-mode and series-mode line filter inductors?
• What is the difference between a Faraday screen and a safety screen in a switching transformer?

Next: Questions/Answers about Switching Power-Supply Topology

Prev: Electromagnetic Interference (EMI) in Switching-Mode Power Supplies

top of page   Home