Skip navigation
Shedding light on power quality for energy efficient bulbs

Shedding light on power quality for energy efficient bulbs

Energy efficient lights can cause power quality problems on ac lines. Be prepared for these power line effects to get more scrutiny.


Power measurements using Fluke metering equipment,

Wikipedia page on lumens,

CLF bulbs Energy Star specifications,

One aspect of energy efficient lighting that is rarely discussed is the way in which it will affect the utility power system. With this in mind, we decided to measure the new modern light bulbs. We tested incandescent light bulbs (ICBs), compact fluorescent light bulbs (CFLs), and light emitting diode bulbs (LEDs). This is not an exhaustive test in that it covers only a few representative samples available from large retail chains. Nevertheless, we feel the results are a good snapshot of how present-day lighting products behave.

The efficiency of the light bulb can be measured by how much light the bulb emits as a function of the power it consumes. This may seem simple and straightforward, but several caveats apply to how the bulb emits light and to its power consumption. The emitted light should be in a range of wavelengths that the human eye can see. Light emitted in the infrared range would be of no use for conventional illumination.

First a few basics. The common measure of light strength or intensity is lumens. The power a light bulb consumes should be divided into two parts, power and harmonics. The power is a function of the real power in watts, the total power in volt-amperes and the reactive (imaginary) power in volt-amperes reactive. Power factor is a ratio of the power used to do work, typically in kilowatts, divided by the total power in kilovolt-amperes. The power utility must generate, transmit, and transform the total power. The customer usually only gets billed for kilowatts used over a period of time measured in kilowatt-hours (kWh).

Power line harmonics are the other piece of the power puzzle. The harmonics are the whole number multiples (one, two, three, and so on) of the fundamental frequency that the utility delivers, 60 Hz in the U.S. Harmonics are caused by the rapid switching of modern power supplies that convert the 120-V ac supply into some dc voltage. The measure of harmonic distortion is the percent Total Harmonic Distortion (%THD) for voltage and percent Total Demand Distortion (%TDD) for current.

The Institute of Electrical and Electronic Engineers’ (IEEE) 519-1992 “Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems” standard for Current Distortion Limits, Table 10.3, lists 15.0% for the 2nd through 10th harmonic and 20.0% TDD for small loads where the short circuit current divided by the load current is greater than 1,000. (The IEEE term TDD will be used to differentiate between voltage total harmonic distortion [THD] and current total demand distortion [TDD]). The IEEE 519-1992 voltage distortion limits in Table 11.1 for Bus voltages less than 69 kV are 3.0% for individual voltage harmonics and 5.0% THD.

The accompanying figure shows the voltage and current waveform for an old GE 60-W ICB. The voltage supplied by our local utility (Avon, Co.) is the larger of the sine wave traces; the current is the smaller trace. Note that the voltage and current are in phase with reasonable sine waves. The waveforms have a voltage distortion of only 2.4%.

The ICB current causes only 2.3% TDD. The power used by the ICB is 61.14 W, 61.15 VA, with only 0.367 VAR. Of the 61.15 VA that the utility must generate, transmit and transform, the customer with a 60-W ICB is billed for 61.14 W. The power factor (PF) is 61.14 W/ 61.15 VA = 0.9998364 or basically 1. The displacement power factor (DPF) is the real power in watts at the fundamental (60 Hz) divided by the total power in volt-amperes at the fundamental.

Now consider the voltage and current waveform for the Wal-Mart Great Value 13-W CFL, and how its current waveform differs from that of the ICB. The voltage supplied by the local utility is still 2.4%THD but the current that the CFL uses has a 114.5% TDD with 77.7% on the 3rd (60 Hz × 3 = 180 Hz) and 47.5% on the 5th (60 Hz × 5 = 300 Hz) harmonic. The Wal-Mart Great Value 13-W CFL uses 11.45 W, 12.12 VA, and 3.950 VAR with a power factor taking the harmonics into account of 0.61.

Continue on next page

The utility must supply 12.12 VA to operate the CFL but the customer is only charged for 11.45 W. In addition to supplying the difference between the total power and the power to do work, the utility must supply 3.950 VAR and contend with the harmonic current circulating on the power system. There will be more on this harmonic current and the effect on the utility power system later.

A comparison of many different kinds of bulbs shows that the Wal-Mart 13W Great Value CFL purchased in a six pack for $5 is truly a great value for price, lumens, wattage, and is marginally better in harmonic distortion than the other CFLs and LED.

The Energy Star specification for a CFL dictates that its power factor must exceed 0.5. Suppose power factor is compared with school grades, where 100 to 90 is an A, 89 to 80 is a B, and so on. A power factor of just 0.5 would be considered an F minus. Most utilities would like their customers to have a power factor equal to or exceeding 0.9. There is no mention of the IEEE 519-1992, “Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems,” in the Energy Star specification. The IEEE standard for Current Distortion Limits, Table 10.3, is 15.0% for the 2nd through 10th harmonic and 20.0% TDD for small loads where the short circuit current divided by the load current exceeds 1,000.

One might think that CFLs use so little power and current that their poor power factor and large current harmonics will have little effect on the overall power grid. This is NOT our recent experience. One must bear in mind that CFLs are not the only energy-efficient loads with poor power factors and high harmonics. Energy-efficient washers, dryers, and refrigerators all have variable speed drives and onboard computers. All of them also have poor power factors and high harmonics. Ditto for computer systems and electronic ballasts in commercial fluorescent lighting,

On this score, the only good energy-efficient loads would be electric hot water heaters and fan-less resistive space heaters. Back on the utility power lines, the harmonic distortion shows up in the 5th, 7th and 11th harmonic because the utility distributes single-phase 120/240 V from single-phase transformers connected to each of the three phases in home subdivisions. Commercial buildings use one 120/208-V three-phase transformer.

As an example, the accompanying figure shows all three phase currents in a college student center building. The loads consist of electronic ballast fluorescent lighting, computer equipment, and variable speed ac drives on the heating and ventilating units. The current total demand distortion was 19% TDD, mostly on the 5th harmonic.

Clearly the 5th, 7th and 11th harmonic distortions will become more prominent on the power grid as residences and commercial buildings follow government mandates for energy-efficient electrical equipment. These harmonics interact with the utility system with adverse effects. In particular, interactions with high-voltage capacitors commonly used on the power grid can be destructive. These capacitors help maintain the power line voltage in the presence of heavy loads.

The main problem is that the power system resonates or oscillates with these 5th, 7th and 11th harmonic frequencies. The resonant frequency fR of the power line is approximately:

where L and C are high-voltage regulators and capacitors respectively. The natural resonant frequency of the power system drops as the utility adds more high-voltage regulators and capacitors to stabilize the main’s voltage with severely varying loads. Eventually, the natural frequency approaches the resonant frequency and is excited by the 5th, 7th, and 11th harmonic current generated by the rich harmonic loads.

For example, in one case a normal utility voltage waveform changes dramatically with the introduction of harmonic current from a variable-speed dc ski lift drive. There were two high voltage capacitor banks, 1,200 kVAR and 1,800 kVAR, within a mile of a ski resort gondola system. These lowered the resonant frequency of the lines into the region of the variable speed lift drive’s harmonic current.

Continue on next page

The “turn-on” of the light bulbs and the effect of the resultant current spike is another issue. Compare this with the turn-on of the CFL. Note the sharp spike in the current that even creates a spike in the voltage waveform. Note on the traces for an LED, there is more noise on the current and voltage waveform at turn-on.

The CFL is great for the consumer with its low wattage, but bad for the utility with its higher volt-amperage, harmonic distortion and low power factor. Just because a CFL or LED is expensive does not mean it is good on all levels. We think the Wal-Mart Great Value bulb is indeed a great value. The poor power factor, high harmonic distortion, high inrush current and high temperature of the power supply in the base of CFLs and LEDs are an indication their design has yet to be perfected.

Now a bit of editorializing: With the low power factor spec and absence of a harmonic distortion limit, Energy Star should be referred to as Energy Challenged, to be politically correct. The new energy saving lights, appliances, computers, and entertainment systems will be detrimental to utility power quality.

Undoubtedly, the DoE eventually will wake up to how energy- efficient lighting (and energy-efficient appliances, for that matter) are degrading the utility grid. Manufacturers that take preemptive action on power line quality and PFC issues will be ahead of the curve.

Making the measurements

The equipment used for testing electrical properties included a Fluke 434 Power Analyzer with 40-A 1 mV/A current probes. A Wavetech ELS2 current sensor with X1 and X10 current loops and an AEMC Model MN103 1 mV/mA current probe was used for more accurate current measurements. The Fluke 434 measurements were compared to those of a Kill A Watt model P4400 where applicable. A Fluke 199 200 MHz 2.5G S/sec Scopemeter captured the turn-on current and voltage waveforms. Temperature measurements were via a Raytek IR gun model RAYRPM30L3U.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.