Shielding in Hi-Fi Equipment
December 1955 Radio & Television News

December 1955 Radio & TV News

[Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio & Television News, published 1919-1959. All copyrights hereby acknowledged.

Recently, I posted from this same 1955 issue of Radio & Television News magazine the article "Mac's Service Shop: Magnetic Shielding." Maybe it was a coincidence that it was printed in the same issue as this article entitled, "Shielding in Hi-Fi Equipment," but probably not. Discussed here are all three fundamental kinds of fields from which equipment and/or components might require shielding to prevent cross-coupling of signals: Magnetostatic, electrostatic, and electromagnetic. An electromagnetic field is present whenever either a magnetic or electric field is varying since one begets the other. Different types of shielding materials are required for blocking a magnetic field versus an electric field, where material permeability and conductivity, respectively, determines effectiveness. In the era when this was written, most electronic equipment consisted of leaded (not Pb'd) components interconnected inside a metal chassis via point-to-point wiring with all combinations of relative orientations so that opportunities for inductive coupling were legion. Interconnecting power supply and signal wires ran helter-skelter throughout the maze, further complicating the problem. Add to that the high voltages used by vacuum tubes and impedance matching and voltage transformers, and it's a wonder sometimes that anything worked on low level received signals.

Shielding in Hi-Fi Equipment

By W. Philbrook

Many of the facts covered herein will be helpful when shielding your equipment for best performance.

At first sight, shielding seems to be a pretty simple little topic, but often it does not behave quite the way it is expected to, and even some of the more expert engineers get into difficulties trying to find out why. So there is no need to apologize for introducing a further article on this deceptively simple subject.

One still meets such questions as: Do you ground both ends of a shielded lead? Does a magnetic shield need grounding? Should the material for a magnetic shield have high or low resistance? How important is the permeability of the material? These are only some of the questions that one encounters. There are other aspects that are not clearly understood - in particular, why it is that a high grade transformer, advertised as having a high degree of shielding which checks under test conditions, seems to have inferior shielding under practical circuit conditions? How is this discrepancy explained, and can we do something about it?

Let's start by making sure that we have a clear understanding of what constitutes each of the three basic varieties of shielding.

Magnetic Shield

A magnetic shield is made of magnetic material. The essential property of the material is that it should have a very low hysteresis loss. The purpose of the shield is to capture the magnetic field and lead it around the object to be screened, which is usually a transformer, without affecting it magnetically.

High permeability is a good thing, but more important is the fact that the hysteresis should be low. The permeability ensures that the path for the magnetic field is effectively short-circuited around the object inside the shield. Where the magnetic field is steady, due to pure d.c., the higher the permeability the smaller will be the field inside the shield.

But d.c. fields are not usually the cause of worry. It is more important to make sure that an a.c. magnetic field, such as one radiating from a power transformer or choke, does not get into an input or interstage transformer. This means that a fluctuating field must not pass through the shield. When the field fluctuates, it is important that the magnetic condition of the shield should closely follow the fluctuations of the magnetic fields.

Hysteresis means that the magnetic condition in the shield is delayed behind the magnetizing force causing it, and this means that there will be a difference between the magnetizing force and the short-circuiting effect produced by the magnetic shield. This difference will reappear as a leakage field inside the shield, so it doesn't matter how high the permeability of the material is, if it shows appreciable hysteresis, it will become a poor shield. So a primary requirement is a magnetic material with extremely low hysteresis.

While on the subject of magnetic field we can answer the question as to whether the material should have a high or low resistance. Since eddy-current losses are similar in nature to hysteresis losses in producing a delay in the magnetic field set up, they will have the same effect of deteriorating the quality of the shield. This means that a magnetic shield should be of a high-resistance, low-hysteresis-loss alloy.

The thickness of the shield will have an optimum value too, for any given frequency. Making the shield thicker will decrease the flux density in the material of the shield and so reduce hysteresis loss. But, at the same time, it will increase the path section available for eddy currents, and so increase the component of eddy-current loss in the shield. At some thickness, for any specified frequency, there will be an optimum which will provide a maximum reduction in field due to the magnetic shield.

Magnetic shields for input and inter-stage transformers are usually made of Mumetal or a similar material. An important feature for their satisfactory operation is that any lids or joints in the shield should be a good close fit so as to provide a good magnetic contact. Fig. 1A shows how a magnetic field is led around the shielded space by a magnetic shield, while Fig. 1B shows the effect of a poor joint at some point in the shield: the reluctance at the joint causes some of the field carried around to be re-radiated on the inside of the shield.

To answer another of the questions asked at the beginning of this article: does a magnetic shield need grounding? The answer to this question is: No. A ground connected to a magnetic shield has no effect upon its magnetic shielding properties. However it often happens that a subsidiary effect of a magnetic field is to provide static shielding, which will be discussed later. For this purpose grounding is absolutely necessary and hence it may be advantageous to ground a magnetic shield so that it provides static shielding as a subsidiary effect.

It is important in a magnetic shield that there should be no holes or that any necessary holes in the shield should be as small as possible.

Also, if the Mumetal has to be drilled, or worked on in any manner, after its pressing, it should be re-annealed after work, so as to operate at the lowest possible hysteresis loss.

Mumetal and similar materials are not suitable against very strong magnetic fields, because they saturate at a fairly low flux density. Therefore shields of these materials are only suitable in magnetic fields where the saturation density of the metal is not approached.

Magnetic shields are more effective against the lower frequencies, their greatest effectiveness being against a d.c. field, which is virtually zero frequency.

Electromagnetic Shields

Electromagnetic shielding keeps a magnetic field out by the principle of electromagnetic induction. It depends on the variation in magnetic field, rather than on eliminating the magnetic field itself. Consequently it is inherently more effective at higher frequencies than at low frequencies and is completely ineffective against d.c. fields.

Fig. 2 illustrates the principle. At Fig. 2A the original interfering field is shown with dotted arrows, the currents induced by the electromagnetic field are shown by the solid arrows, while the fields due to these induced currents, opposing the original field inside and aiding it outside, are shown by hollow arrows.

The resulting field around such an electromagnetic shield is shown in Fig. 2B. The shield is shown open-ended to demonstrate the manner in which the shield works. In practice, electromagnetic shields may be complete cylinders, with the ends filled in, in which case they will be equally effective in eliminating fields in any direction. The circular band shown will only be effective in eliminating fields along the axis of the cylinder.

For this kind of shield it is important that any lids or joints should make good electrical contact. There must be no gaps of any kind in the shield and it should be constructed of a low resistance material such as copper or aluminum. Sheet tinned iron does not make an effective electromagnetic shield because the iron will not make a good magnetic shield and its effect on the current in the tin will interfere with its operation as an electromagnetic shield.

Electrostatic Shielding

This kind of shield does not concern itself with magnetic fields, but with electric fields. It is intended to keep electric fields out or in as the case may be.

A good ground is essential to the operation of an electrostatic shield, although this was not vitally necessary to either of the other types. The purpose of an electrostatic shield is to interpose a grounded shield between two interacting potentials that might radiate from one to another. Fig. 3 shows the way in which an electrostatic shield intercepts an electric field.

In an electrostatic shield, provided the material is basically a conducting material and not an insulator, it is not important for it to have particularly low resistance. Tinned sheet iron will serve as well as any other material for this purpose, provided it is not also required to serve as a magnetic or electromagnetic shield.

Applications

Having differentiated between the various kinds of shield, we can now see how they may be applied. The first thing to consider is the kind of field against which shielding is required.

If it is basically a magnetic field, due to a power transformer, a choke, or a motor, then a magnetic or electromagnetic shield, or a combination of both, will be necessary to eliminate hum pickup effects.

If, however, we are concerned with a high impedance circuit, in which static fields due to power line voltages around the place can cause trouble, then electrostatic shielding is required.

Where a transformer coil is involved, such as an input or an inter-stage transformer, magnetic or electromagnetic shielding is invariably required. If one of the windings is high impedance, then electrostatic shielding may also be necessary to protect the high impedance winding against static pickup.

In circuit wiring, the kind of shielding needed will depend upon the impedance of the circuit.

Low-impedance circuits, where an interacting magnetic field may induce relatively large currents, require some kind of shielding to eliminate this effect and, if a shielded lead is used, its primary purpose is to eliminate the induction of current in the circuit, rather than to eliminate the effect of static potentials.

In high-impedance circuits it is static potentials, alternating or direct, that have to be guarded against, and this requires electrostatic shielding.

To return now to the questions asked at the beginning of the article.

Do you ground both ends at a shielded lead? If the shielded lead is intended solely for protection against electrostatic field, it doesn't matter how many times it is grounded. But it is also important that the shielded lead should not produce induction in the lead it is shielding. If there is any difference of potential between the points at which it is grounded, there will be a current flowing in the shield due to this difference of potential and this current will produce an induced current in the lead it is shielding. For this reason it is dangerous to ground both ends of a shielded lead.

There may be a difference of potential between the ground points to which the two ends are connected, and if this should occur, the shield will be effective against electric fields, but at the same time, it will be responsible for injecting, through electromagnetic induction, another source of interference which may not be present before both ends are grounded.

Distorted Magnetic Fields

Now we come to the sixty-four dollar question: the one about why measurements on effective shielding do not line up with practical circuit performance.

Large degrees of shielding are provided by concentric arrangements of either successive magnetic shields or alternate magnetic and electromagnetic shields, as shown in Fig. 4. In their best formation each one should be a complete shield symmetrically spaced from its neighbors as shown in Fig. 4A.

However, practical construction and the economics of production make this rather expensive to produce and, consequently a compromise such as that shown in Fig. 4B is used, where a succession of nesting cylinders is used and usually spaced so as to retain symmetry.

If three such cylinders are used together and each one provides 30 db of shielding by itself, the over-all shielding should add up to the region of 90 db, assuming there is no interaction between one shield and the next. This arrangement may work quite successfully provided the field against which it is shielding is uniform.

Fortunately for test purposes the specified field is a uniform one, provided by a number of turns of wire on a framework a foot square, with the shielding under test placed at the center of the framework. Here the field makes a close approximation to uniformity and is of accurately predictable field strength.

With a carefully balanced out construction of the simplified variety the shielding may be quite effective in whichever direction the unit is oriented. However, if the field is non-uniform, which is far more the usual condition, as for instance where the field is radiating from a power transformer, shown in Fig. 5B, the symmetrical distribution of the field through the different shields comprising the nest will no longer follow, and the asymmetry will tend to exaggerate the transmission of the field through the combination of shields.

Measurements have shown that under these circumstances sometimes a combination of shields intended to provide a 30 db-per-stage reduction will provide less than 30 db for the whole nest, although one stage by itself may still show as much as a 30 db reduction. This is due to the fact that the outermost shield, being asymmetrical, distorts the original field further from its original distorted condition and the successive shields never succeed in getting the field straightened out again.

Combined Core and Shield

An interesting construction providing quite a useful degree of shielding is shown in Fig. 6. Here, instead of using either of the more usual types of core construction, the core is built up of "F" laminations so as to provide a single air gap in the center leg. With the usual construction, either from two different sets of "E" laminations, or from "E" and "I" laminations, the reluctance, through all three legs of the magnetic structure is approximately equal, and so any magnetic field reaching the laminations will divide itself so as to pass a component, of from 1/3 to 1/2 of the total, through the coil. Using the "F" laminations, there is a low reluctance path each way around the coil, of approximately equal reluctance, and a high reluctance path through the center of the coil. Consequently any field reaching the core assembly will be diverted around the coil.

This construction can be as effective as a complete shield. If this assembly is then symmetrically mounted in a single further shield, the total shielding is as effective as two of the more normal types of shield, and has the added advantage that it does not discriminate against the practical type of non-linear field. In other words, it gives us results under practical circumstances as good as the test figures show, which the other constructions often do not.

Posted September 17, 2020