“NCIS” aficionados should recognize the Gibbs-ism in the title. The phrase came to mind during a recent harmonic standards group meeting. It focused on the intention of an international standards group to extend the requirements for the measurement of harmonics into the 10–150 kilohertz (kHz) region, which most of the regulatory standards consider a “no-man’s zone.” After some debate, someone suggested, perhaps, we should find out if there really are problems out there with signals in this region before we write the standard. What a novel idea—checking if there’s a problem first. Ya think?

Most people in the power quality industry know of harmonics as “signals that are integer multiples of the fundamental frequency.” The two most common fundamental power frequencies are 60 hertz (Hz) in the Americas (and half of Japan) and 50 Hz in most of the rest of the world (including the other half of Japan). For reasons that I won’t get into, the 50th harmonic (3,000 Hz) became a typical maximum measurement on 60 Hz systems. In the International Electrotechnical Commission realm, the primary region of interest until recently was up to the 40th harmonic or 2,000 Hz. Requirements for measurement and limitations on emissions have since been pushed up into the 2 kHz to 9 kHz region.

Until recently, it was uncommon to find harmonics of any significant magnitude above the 19th harmonic. As I discussed in previous articles, the typical harmonics that one finds in a power system are based on how the load uses the electricity, since the generators tend to make harmonic-free fundamental-frequency-only voltage. The number of paths of conduction, p, or pulses, is part of the h = n p +/-1 equation, with n being integers from 1, 2, 3, on up. So for a single-phase nonlinear load, such as a computer power supply, the typical harmonics will be the 3rd, 5th, 7th and 9th, found in smaller and smaller magnitudes as harmonic number gets larger.

For a three-phase, full-wave rectified load, such as in many adjustable speed drives (ASDs), the number of pulses is 6 (3 phases 2 half cycles where conduction takes place), so harmonics are the 11th, 13th, 17th, 19th, 23rd, 25th and so on. Newer ASDs use higher number of pulses (12- or 24-pulse converters are becoming more common), which have the advantage of drawing less harmonic current than lower pulse converter. However, the harmonics for a 24-pulse converter start at 23, 25, 47, 49, 71, 73 and so on, which puts signals in the range where no harmonics have gone before. Hence, the interest in standards concerning emissions above the 2 kHz value. But why the 10–150 kHz region? And are these really harmonics in the aforementioned definition of such? Does anyone care?

Starting with the latter question, there are a few standards with conducted emissions limits in this region (CISPR 16, band A), with MIL-STD-461 Requirements for the Control of Electromagnetic Interference Characteristics of Subsystems and Equipment, while the more popular commercial electromagnetic interference specs (such as FCC Part 15 Subpart B and IEC 61000-2-x for CE mark regulations), start at 150 kHz and go up to 30 megahertz (CISPR 16, band B) or higher.

Indirectly, by these standards looking above and below this region, equipment manufacturers found a sweet spot for the switching frequencies of the power supplies (hence, the first question). The rectifiers in the front end of the power supply that convert the alternating current (AC) to direct current (DC) produce the aforementioned harmonics. However, this DC value is converted into other DC values or back into AC, depending on the equipment’s power supply needs. Conversion involves “switchers,” which are transistor-type circuits connected to a transformer that “chop” the DC into an AC waveform at a much higher rate than typical harmonic frequencies.

In general, higher frequencies require smaller electromagnetic components. Smaller is generally also less expensive. Switching at too high of a frequency would result in measurable emissions where there are restrictions. Figure 1 shows one such culprit, an electronic ballast from a fluorescent lighting fixture. Significant levels of 46-kHz signals were found at one facility, which were traced back to the ballasts. There also are power line carrier and other asynchronous signals operating in that area.

As for the second question, these signals mostly are not harmonics that are integer multiples of the fundamental frequency. The clock circuitry that runs the switching frequency can be independent of the power frequency, as in the above example. Though the signals could technically be called interharmonics, they are better categorized as noise.

Back to the original premise. Are the problems out there from such noise signals significant enough to require rules and regulations to govern 10–150 kHz? There are rules and regulations for harmonics, and there are clearly documented problems. Problems also occur in the regulated areas above 150 kHz, though they aren’t common.

So what do you think? Is this truly a problem needing regulation or another rule for rules’ sake?


BINGHAM, a contributing editor for power quality, can be reached at 732.287.3680.