This is a response to Keith Armstrong's letter, "CE Mark Is Not Approval" (Compliance Engineering, September/October 2001). As an engineer responsible for getting many products tested to pass CE requirements, I find his letter illustrates many of the things that make CE marking so confusing.

When he talks about using the self-declaration method, he states that manufacturers can decide what testing to do and who they get to do the testing, or they can decide to do no testing at all as long as they apply all the relevant standards in full. He talks about self-declaration being a legal statement that says, "If my product were tested to the relevant EMC standards it would pass," but then he states that CE marking does not mean that the product had been tested, or even if it were, that it in fact passed. Statements such as these drive engineers nuts and are the reason the CE mark is so confusing.

If I need to get my product UL listed, for example, I call Underwriters Laboratories and speak to an engineer who tells me exactly what standards apply and what tests I need to pass. And, believe me, the tests are done. There is a big difference in saying you have designed something to meet a standard and actually meeting it. Just talk to anyone about the first time he or she tried to get a product UL listed! It would be nice to just say, "I have met all the standards, but I never ran any tests."

On the contrary, if I need to get a product CE marked, who do I call? Is there a CE company? Anyone know the phone number? There are many consultants with as many opinions, but there is no definitive authority that can provide something in writing. If such an entity exists, however, please let me know. I would be very happy to get to the truth.

Armstrong replies:

It was not my intention to create confusion, but to prevent it! I apologize if I wasn't successful. For instance, I didn't mean at all to suggest that CE compliance means "designed to meet." It is usually a very brave (or foolhardy) person who declares that a product would meet EMC tests purely on the basis of design, although it is not impossible to do this on some products.

Products supplied in the European Union (EU) are generally required to meet the relevant standards, and there can be substantial penalties for products found not to comply, especially for those that cause or suffer from interference problems. Most products, of course, need some testing in order for manufacturers to have enough confidence in their EMC compliance, but manufacturers with enough expertise can save time and money on the testing they do.

There is no reason why a manufacturer shouldn't take an EU-bound product to a test lab to determine what standards apply, do all the tests, identify necessary remedial actions, and produce a "pass" test report for use as the basis for CE compliance. All that would remain would be to have a company representative, such as a vice president of engineering, sign the company's EU declaration of conformity, affix the CE mark, and sell the product in the EU. Using this approach, however, could be a waste of time and money.

For EU CE compliance, small- and medium-sized companies tend to rely on external test labs. It is important to note that they can choose their own lab and aren't restricted to selecting from a few expensive ones. Nothing prohibits these companies from doing all or some of the testing themselves—if they feel competent enough—and some even rent EMC test equipment rather than buy it.

Larger companies can often save time and money by setting up their own test laboratories. To provide an extra level of confidence, some companies have their in-house labs accredited by their national bodies. Incidentally, start-up companies with limited capital and short product launch schedules can benefit from the lower compliance costs and variety of options associated with the CE compliance approach.

The easiest way to ensure confidence in a test lab is to choose one accredited by a government-appointed accreditation body. Lists of these bodies and accredited labs are available with a bit of Web searching, starting with the EU's home pages for the relevant directives. Since the EU–U.S. mutual recognition agreement [MRA] has matured, there are now two U.S. accreditation bodies that can provide a list of accredited EMC test labs. A useful list of U.S. and EU test labs is included in Decision 3/2000 of the EU–U.S. MRA.

The European Commission chose the "manufacturer's declaration of compliance" approach to regulatory compliance both to break down technical barriers to trade between the many countries in Europe and to minimize the effect on manufacturers' costs. Their chosen approach is, therefore, a great method for the globalization of trade. The "type testing and approval by nationally appointed laboratory" approach that manufacturers are used to actually inhibits global trade. Most manufacturers are probably aware that FCC has already shifted its EMC requirements to an approach very similar to that required by the EMC Directive.

It is imperative to understand CE marking to compete most effectively in Europe (and increasingly in global markets as they begin to follow the same approach). Here are a few key points to remember about compliance with the EMC Directive (and the Low Voltage Directive [LVD] too):

I'm sure this last point will give rise to howls of protest. Many company representatives miss this important issue, as do many test labs. It means that a manufacturer needs to be confident that its product will not cause interference due to its emissions or be unreliable due to inadequate immunity in its intended normal operating environment, even if the resulting EMC requirements go beyond what is covered by the harmonized EMC standards.

Similarly for compliance with the LVD: meeting a harmonized safety standard may not be enough to meet the LVD's essential requirements for a product to be "safe." I can provide examples of such circumstances to anyone interested. It is necessary to fully understand the EU market's regulatory requirements to improve competitiveness—and to reduce the financial risks of operating in that market.

I just finished reading one of Niels Jonassen's articles on your Web site. In the article, he uses the term permittivity. He does not define the term, and it is critical to understand its meaning to get anything out of the article. What is it?

Jonassen replies:

The concept of permittivity (usually labeled e) is defined in Maxwell's first equation (also known as Gauss's sentence or theorem for an electric field). The equation states that the surface integral over any closed surface S of the normal component of the electric field strength E equals the enclosed charge divided by the permittivity, which can be expressed as

or illustrated by Figure 1.

Figure 1. The concept of permittivity.

The basic definition leads to the well-known formula for the field strength E at a distance r from a point charge q:

The permittivity e is thus a parameter describing the influence of a material on an electric field. For vacuum (and, in practice, atmospheric air), the permittivity has the value

For many common, insulative materials like Teflon, Plexiglas, and so forth, e is about 2–10 times greater and is therefore often written as

where e_r is called the relative permittivity or dielectric constant. As an example, e_r = 2.1 for Teflon.

Figure 2. A parallel-plate capacitor.

But let's look at a practical example. Figure 2 shows a parallel-plate capacitor charged with a charge q to a voltage V_v (v for vacuum or air). The capacitance is thus

If the space between the plates is filled by a dielectric (an insulator), as in Figure 3, the voltage drops to V_d. The voltage drops by a factor equal to the relative permittivity e_r of the material; in the case of Teflon, by a factor of 2.1

Figure 3. A parallel-plate capacitor filled by a dielectric.

In this situation, the capacitance C_d can thus be written

Therefore, the effect of an insulator with relative permittivity e_r has been to reduce the field by a factor of e_r, and increase the capacitance by the same factor.