You know you have harmonics, because you've seen the graphical display on your harmonics analyzer. The big question is: "How do I get rid of them?" The answers lie not in magic, but in applying fundamental principles and possibly a few specialized cures.

In the third and final part of our series, we conclude our discussion with specific countermeasures to combat harmonic distortion. In Part 1 (March 1998 issue) and Part 2 (May 1998 issue), we learned that much of the panic surrounding the issue of harmonics is largely due to misunderstanding what the shape of a waveform means to a building distribution system.

Harmonics are not always going to present measurable problems. However, they can and do create havoc in many situations. Some possible problems include overheating of distribution transformers, adverse effects on electronic equipment, and system resonance with power factor correction banks. Potential sources include computers, lighting ballasts, copiers, and variable frequency drives.

In this installment, we'll look at countermeasures you can employ when installing variable frequency drives. But implement these countermeasures with care. Harmonics do not always create problems. Thus, spending time and money to carry out all of the key countermeasures is not always a good investment. If you know harmonics are causing problems, then you need to apply the countermeasures in a methodical fashion. If you can't detect problems from harmonics, scale back your countermeasure efforts.

Guidelines that address harmonics share a common goal: to maintain the quality of electrical power at the point of common coupling (PCC). You meet such a goal if you limit the harmonics-induced voltage distortion at the point of common coupling. So, what is this "point of common coupling?" It's the interface between sources and loads. Most experts select the primary of a feeder transformer as the PCC, and measure there for distortion.

Do you have a problem resulting from harmonics? Do estimates show future facility expansions may push voltage distortion limits over 5%? If so, consider using the following countermeasures to reduce current distortion. This will have a dramatic impact on voltage distortion levels.

Variable Frequency Drive (VFD). In the United States, most variable frequency drives (VFDs) go into commercial (HVAC) applications. Consequently, this is the fastest growing market for VFD manufacturers. In this market, the cumulative average growth is just over 9%. People install VFDs to reduce total power consumption. These VFDs reduce the load on the feeder transformer (see Fig. 1, original article). This reduced load reduces the heat in the transformer core. In 95% of applications (assuming you follow the 80% transformer loading rule), voltage distortion will not exceed specified standards.

Furthermore, if you retrofit drives that incorporate SCRs (in the power conversion section) with drives that incorporate diodes (to rectify the incoming voltage waveform), you eliminate voltage notching. Accomplishing this does not require additional hardware. Fig. 2 (original article) shows the waveform to expect if you measure the current waveform at the input terminals of the VFD.

A Fourier analysis may show current distortion levels between 20% and 110%, with resultant voltage distortion levels of 1.5% to 4%. These values depend on the system characteristics, including amount of impedance and preexisting distortion levels. If the transformer impedance is high, the current distortion will be relatively low (soft system). If the transformer impedance is low, the current distortion will be relatively high (stiff system). See Rule of Thumb #9 in the sidebar.

Variable Frequency Drive with DC link choke. Most VFD manufacturers incorporate a DC link choke (Fig. 3, original article) into some of their VFD models. In some VFD models it's standard; in others it may be an option. A DC link choke is simply an inductor in the ripple filter circuit, ahead of the DC bus capacitors. The added inductance limits the rate of change of line current relative to time (di/dt) into the capacitors. This results in lower peak currents. Fig. 4 shows the waveform to expect if you measure the current waveform at the input terminals of a VFD with a DC link choke.

A DC link choke reduces current distortion typically by 40% to 60%. This may not have a quantifiable impact on the measured voltage distortion. In the specification process, it's typical to substitute a DC link choke for a three-phase input reactor. Consequently, designers use this same technology in other equipment to control harmonics. See Rule of Thumb #10.

Variable Frequency Drive with three-phase input reactor. It's common to specify an input reactor when ordering a VFD (Fig. 5, original article). The reasons include high frequency noise filtering, increased protection from surges (that may result from power glitches or lightning strikes), and harmonics reductions. You can see a typical input waveform in Fig. 6(original article). Engineers size reactors according to percent impedance and the resulting voltage drop after drawing full current through the reactor. Typical values are 3% to 5%, with 3% reactors being more common. A reactor simply adds impedance to the system. This impedance slows the di/dt of the current flow into the capacitors of the drive. The result is a lower current distortion level. The effect is the same as having a high impedance transformer.

A 3% input reactor will reduce the current distortion by about the same percentage as a DC link choke (40% to 60%). Be aware you can easily reach the point of diminishing returns regarding impedance and current distortion levels. Laboratory testing reveals, after 7% input impedance, the distortion levels do not drop much. An input reactor may be two to three times the cost of a DC link choke, and may be overkill for just harmonics mitigation. See Rule of Thumb #11.

A few VFD designs capitalize on the flexibility of using two 6-pulse rectifiers in parallel (Fig. 7, on page 44, original article). This virtually eliminates power harmonics through phase manipulation. Typically, phase manipulation is a result of converting power through a specifically designed transformer called a 12-pulse transformer. This transformer is different from an autotransformer or a phase shift transformer because the impedances balance properly on the secondary side. This type of transformer effectively eliminates the 5th and 7th harmonics in the core of the transformer. Since these harmonics contribute to approximately 90% of the harmonic spectrum, using this method allows very high-density packaging of drives to transformer size. The result is pretty close to a pure sine wave (Fig. 8, original article) measured at the input terminals of the transformer. See Rule of Thumb #12.

The 12-pulse input transformer allows you to reduce current distortion levels by 90% and meet strict power quality specifications, such as IEEE 519-92. Most installations do not need cancelation of harmonics. With the transformer not fully loaded, you should have enough capacity to provide clean power to all equipment.

Passive and active harmonic filtering. Sometimes, it's possible to design a filter that will eliminate specific harmonics. It's also possible to implement a device that will actively sample the incoming waveform and inject harmonics to cancel detrimental harmonics. We call the first method passive harmonic filtering and the second active harmonic filtering .

Most trap filters employ a design that eliminates the 5th harmonic. Trap filters do this by providing a low impedance path for that harmonic. Suppose you have a concern about the 5th harmonic. We know its frequency is 300 Hz on 60-cycle power. So, you can install a filter that will trap all 300 Hz frequencies. This option is usually expensive when compared with other countermeasures. Active filters are more complicated than this, and therefore are usually even more expensive.

Other countermeasures. Other countermeasures include transformers with higher K-factor ratings, other filtering techniques, and phase shift transformers for powering different VFDs. Each of these carries a tradeoff. For example, the higher K-factor transformers can actually worsen your power quality problems if you apply them incorrectly.

In addition to harmonics-specific countermeasures, pay attention to your grounding, bonding, conductor sizing, conductor routing, and transformer loading. Implementing these basics of power distribution per the NEC can go a long way toward resolving harmonics problems. Fig. 10 (original article) compares four major countermeasure techniques.

Specifications. What guidelines should you apply if you have too much harmonic distortion? Rule of Thumb No. 13 can help you out. Here's what the typical contractor encounters:

Input Harmonic Distortion: The VFD data shall include calculations proving that the primary side of the input isolation transformer or AC line chokes draws no more than 4% fifth harmonic current of the rated full load running fundamental current. The VFD manufacturer shall make any and all modifications to the drive necessary to guarantee meeting this requirement.

In one field study (done after the installation of many VFDs), THD of the current (ITHD) was 110% at the input terminals of the VFD. If an engineer specified the most effective countermeasure in the industry before checking the voltage distortion, that engineer could have needlessly divested the company of thousands of dollars. In this situation, the total voltage distortion (VTHD) measured 2.2%. Another field study shows before appropriate countermeasures, ITHD was 21% and VTHD was 6.1%. Both examples focus on the pulsed current waveform in an attempt to control the quality of the power system. This may not be the most prudent approach. A better specification might be:

VFD shall not increase the voltage distortion above 5% at the input terminals of the VFD or line filters. If calculations show otherwise, the manufacturer shall make all modifications to the drive necessary to meet this requirement.