Faraday Shields

by William D. Kimmel, P.E.

and Daryl D. Gerke, P.E.

Kimmel Gerke Associates, Ltd.

Any shield that blocks electric fields is, in fact, a Faraday Shield, also known as a Faraday Cage. The basic concept of an electrostatic shield is that a metal member intercepts the electric field, which is trivial at audio frequencies. The Faraday Shield is most commonly seen as the shield of an enclosure or cable, but the term is most commonly applied to low frequency electric field shielding.

Let’s take a look at the shielding basics, some of the common applications and then talk about the necessary conditions for effective shields.

Basic Principles

As Faraday showed almost two centuries ago, an electric field does not penetrate a metallic enclosure – if there is a net charge on the surface, the electric field will be entirely external to the enclosure, with no field within. That principal is valid for electrostatics and even low frequencies, but is strained at higher frequencies, where the shield becomes less than a perfect conductor, and fields penetrate or are generated by openings in the enclosure.

But for low frequencies (dimensions of opening less than 1/20 wavelength), the electrostatic shield is quite effective. That means that audio frequencies are almost trivial to shield. At higher frequencies, the effectiveness of a partial shield degrades, but they can still be useful.

Note that while the shield will block electric fields, it will not impede low frequency magnetic fields. This is a good news/bad news situation. The good news is that the electric field in a transformer can be shielded without blocking the magnetic field necessary for transformer function. The bad news is that magnetic fields are hard to shield, but that’s another story.

Some Common Shields

Let’s look at a textbook example, the switching power supply, as shown in Figure 1. The noisy node connecting the transformer to the switch couples capacitively to the heat sink and to the secondary of the transformer. Both of these paths can be intercepted with a Faraday Shield as shown. The undesired currents are diverted to a safer path. Note that the shield is not connected to the noisy end - that would just increase the coupling path, making things worse.

The shielded transformer is better known in 60Hz power applications. The interwinding capacitance of a power transformer becomes essentially a capacitive coupling network, rendering it useless at frequencies above about 1MHz. Insertion of a shield between the primary and secondary will extend the useful frequency range to 1MHz and beyond.

Although widely cited in EMC books, the practice is avoided as much as possible. You need to lay down a winding (usually the primary), then lay down a shield layer (any conductive material will do), then lay down the remaining winding (usually the secondary). In the case of large 60Hz power transformers, it is merely cumbersome. In the case of smaller on-board transformers, such as with switching converters, fabricating a transformer shield is not amenable to factory automation.

One alternate that we have recently seen is to use an additional tertiary winding as the shield (CE Annual Reference Guide 2005, De Leo, et al). In this example, you simply lay down three sets of windings, with the center layer being the shield – preferable to an inner shield. Test data indicates good performance, especially in the 5 to 30MHz range. We haven’t tried this one, but the solution is much easier to implement than a separate shield – we think it is worth a try.

While these are the two most common cases, there are other applications as well. Figure 2 shows several typical circuit board shields, a chip shield and a connector shield. The chip shield blocks the E-field directly from the die. This problem is particularly noticeable when there are metallic members in close proximity to the chip. Of particular concern is where a nearby member is a mediocre enclosure shield – the currents capacitively couple to the cover, and the return current path energizes the enclosure seams on the way back to the chip. In order to be effective, the chip shield needs to be grounded to circuit ground at four corners, as a minimum. The goal is to return the currents back to the place of origin, which is the chip itself.

The connector shield serves to block capacitive coupling to the connector pins. A common situation is where the filter elements are placed on the circuit board, but stray internal electric field coupling to the connector pins bypasses the filter elements. In this case, the grounding of the shield is dependent on the nature of the problem. Best termination of the shield depends on the source. Field coupling directly from a nearby chip is best terminated to circuit ground to provide a low impedance return path to the chip. If the coupling is from locations not on the driving circuit board, the shield will need to be connected to enclosure ground.

Shield Requirements

Now that we have seen some common applications of the Faraday Shield, we need to look at how to correctly implement the shield. There are several conditions that need to be fulfilled.

First, the shield must be placed to intercept the noisy part of the component – its purpose is to block the field.

Second, the shield must be terminated to the proper location – since the shield is serving as a receptor of capacitive currents, you want to steer the currents away from the undesired path and toward a desirable path, usually back to the source of the currents. Specifically, do not make the connection at, or near, the noisy node or you will make things worse. Imagine putting an arrow on the shield and pointing it to the proper direction. If you have a noisy and a quiet terminal on the component (e.g., a series inductor or a source/drain on an IGBT), you terminate it to the quiet side – connecting the shield to ground is often a good choice, but not always desirable or feasible.

Third, the shield needs to be terminated with as low an impedance as possible. Its purpose is to provide a low impedance shunt, and that won’t work if you have even a small inductive impedance in the path. Chip shields need to be grounded at all four corners as a minimum, preferably more often. Heat sinks will work as shields if they are conductive and well grounded.

Fourth, the longest path length of the shield and termination should be held to 1/20 of a wavelength of the highest frequency of interest. Once you get above this length, antenna effects start to become significant, degrading the effectiveness of the shield.

Fifth, if you are shielding an inductor or transformer, you need a gap in the shield lest you create a shorted turn.

Summary

Faraday Shields can be very effective for providing local isolation. Used on the circuit board, they can provide effective shielding at a fraction of the cost of shielding an entire board or even an entire enclosure.

While it is not difficult to achieve effective shielding, done incorrectly will degrade rather than improve your performance. Make sure the shield covers the desired coupling path, terminate the shield to a quiet node with a low impedance path, and avoid the “shorted turn” situation.