Is it a serious problem if the light intensity varies in your workplace? It all depends on the kind of work you are doing and the way in which the light changes. Intensity fluctuations of less than 1% are quite noticeable. The amplitude, frequency of change, and rate of occurrence determine the effect of flicker.
For example, the human eye is most sensitive to light fluctuation at about an 8-Hz rate. But, even if light varies over an interval of perhaps a few seconds, it can become very annoying. For a worker in a munitions plant or a machine operator, such an annoyance could become a distraction and very soon a safety issue. Flicker can cause convulsions in persons prone to epilepsy.
Flicker testing evaluates the importance of fluctuating light levels by mimicking human perception. Standard IEC 61000-4-15 summarizes the research that supports the IEC 61000-3-3 flicker standard and establishes a measurement method. A flicker meter:
- Records voltage fluctuations.
- Converts voltage changes to an estimate of light variation from an incandescent bulb.
- Weights this estimate according to frequency to account for human perception.
- Determines an instantaneous flicker perceptibility reading (Pinst).
- Derives a short-term flicker indication (Pst) over a 10-min period.
- Derives a long-term flicker indication (Plt) over a 2-h period.
As a guide, the threshold of flicker perception at about 8 Hz for half the population is only 0.3% relative modulation. This means that accurate, high-resolution voltage measurement is a prerequisite to good flicker measurements. The range of threshold levels across a wide frequency up to about 15 Hz is collectively described by Pst = 1 or the threshold of irritability.
Figure 1 (see the October 2001 issue of Evaluation Engineering) plots maximum allowable voltage change (dlim) vs. frequency of change for Pst = 1, where Pst is defined as d/dlim, d being the actual voltage change.1 From Figure 1, a 1% voltage change results in Pst = 1 at about 20 changes/min or 0.16 Hz; two changes are required per cycle. The minimum point is near 1,000 changes/min or 500/60 = 8.3 Hz. At this frequency, dlim is close to 0.3.
According to Larry Conrad, chairman of the IEEE P1453 Flicker Task Force, “We have quite a bit of experience in the United States that shows almost no customer complaints with Pst less than 1.0. By the time Pst reaches 1.2, we can count on getting complaints, and they increase rapidly after that.”
Many power-distribution networks actually behave very well most of the time. They have a low impedance that provides a stable voltage despite varying load currents.
Of course, a network’s wiring has physical limitations. The large conductors have small but finite amounts of resistance and inductance. Should there be a large load change on the network, electric arc furnaces being a particularly disruptive example, the rms voltage available to other local electricity users will change.
A circuit diagram for the IEC 60725 European Reference Impedance is shown in Figure 2.2 The circuit values are based on work done in the 1970s for 230-V, 50-Hz networks. Since that time, much overhead cabling in Europe has been replaced by underground wiring. The effect has been to reduce the reactive value somewhat, although the resistive component “appears to be about correct.”2
When a suspect product is tested for excessive flicker, it is powered from a zero-impedance supply in series with the reference impedance. If the normal wall socket is used, the AC source impedance may be comparable to the reference impedance. In this case, the measured voltage variation would be twice its actual value.
Obviously, this aspect of flicker testing is subject to error. In fact, IEC 61000-3-3 suggests that the voltage fluctuation be measured on both sides of the reference impedance to verify the assumption of near-zero source impedance. Some test equipment, such as Credence Technologies’ AC2000, avoids the potential problem by measuring current fluctuation and calculating the corresponding drop that would have occurred across the reference impedance.
To Measure Flicker, Use a Flicker Meter
Figure 3 (see the October 2001 issue of Evaluation Engineering) shows the various areas, designated by blocks, that are required to implement a flicker meter according to IEC 61000-4-15.3 Block 1 establishes the reference level against which voltage fluctuations are measured.
IEC 61000-2-8 discusses the effects of using fixed or sliding voltage references. A sliding rms voltage reference is continuously calculated over a specified interval to represent the value of voltage at a point in the network just before a change occurs. Using this type of reference, it’s a simple matter to establish some percentage tolerance band outside of which a voltage deviation has occurred. A fixed reference is awkward to use because the supply voltage may be exhibiting long-term fluctuation as well as short-term flicker.
In block 2, the modulation caused by flicker is separated from the 50-Hz or 60-Hz frequency of the AC supply. The squaring multiplier simulates the variation in light output from an incandescent bulb in response to voltage fluctuation.
Block 3 accounts for human perceptibility. The first subblock, a Butterworth filter, limits the measured frequencies to those the eye can observe. The second subblock shapes the frequency response envelope to correspond to the eye’s sensitivity to flicker, peaking at 8.8 Hz.
Voltech noted in Reference 3 that the filter shape cannot be produced using a conventional filter design. This point is important because it underscores the difficulty of measuring flicker by means other than a proper flicker meter.
As Tom Mahr, general manager of Voltech Instruments, pointed out, all flicker meters are not created equal. “Flicker analysis is a very complex algorithm to properly implement in accordance with IEC 61000-4-15, the flicker-meter design specification.
“The standard defines a series of reference waveforms for type-testing a flicker meter as well as some relatively simple qualifier waveforms for test purposes. By using only the qualifier waveforms, a noncompliant flicker meter can be made to appear to comply,” he explained. “To ensure compliance to the Pst measurement standard, the meter must comply with the full table of both sine and square wave inputs specified in IEC 61000-4-15.”
Figure 4 (see the October 2001 issue of Evaluation Engineering) shows the response of the Voltech PM3000ACE-002 Meter together with the IEC 61000-4-15 sine and rectangular limits.
The final section of block 3 in Figure 3 selects an appropriate measurement range. Because there could be a very wide range of perceptibility values, selection is necessary. Alternatively, a logarithmic classifier could be implemented in a flicker-meter design that would not require range switching.
Block 4 in Figure 3 combines a squaring multiplier and a first-order sliding filter to simulate the brain’s ability to identify change. Finally, block 5 performs the statistical analysis required to assess the probability that the measured flicker would be irritating or actually hazardous.
Flicker is referred to a reference rms voltage level, and this level is calculated each half-cycle of the voltage waveform. For the Agilent Technologies HP 6800 System, “Each rms/instantaneous flicker record contains rms data and instantaneous flicker data acquired at a rate of one record per second. Each rms voltage record contains 120 data points at 60 Hz, integrated over half-cycle periods. Each instantaneous flicker record contains 120 data points at 60 Hz, sampled once every 8.33 ms.”4 In addition to the instantaneous values, an integration period of 1, 5, 10, or 15 min can be selected for the evaluation of Pst.
Three parameters are associated with the rms voltage level. dmax is the relative voltage change between maximum and minimum rms values relative to the nominal voltage. The present limit for dmax is 4%. dc is defined as the difference between two adjacent steady-state voltages relative to the nominal voltage. The limit for dc was 3% but has been increased to 3.3% in the newest edition of IEC 61000-3-3. dt specifies the period during which the rms voltage exceeds the nominal voltage by more than 3.3%.
Compliance with the specification has been complicated by lack of a definition for steady-state voltage. For example, Agilent chose ±0.15% of nominal, inferring from the 5% measurement-error allowance and dc 3% limit that this was the intention of EN 61000-3-3. A level must not change by more than ±0.15% of the nominal for at least 1 s to be considered a steady-state value by the HP 6800.
J. M. Woodgate, a member of IEC SC77A/WG1 and WG2, the working groups responsible for IEC 61000-3-2 and -3 and their EN clones, commented on the less than rigorous language of the specification. After work had begun to replace IEC 555-3 with IEC 61000-3-3 in about 1988, “It was realized that the effect of the European EMC Directive would be to make these standards quasilegal documents instead of the voluntary standards they superseded. This meant that the precision of the language should have been orders of magnitude higher, but it wasn’t. Some of us tried to improve the drafts in the early ‘90s, through the national standards committees, but it was very difficult to have much influence after the work had been in progress for some years.”5
Feedback from flicker-meter users has resulted in some recent changes to the standard. IEC 61000-3-3 can be interpreted to apply to any equipment with an initial inrush current exceeding the 4% limit on dmax. For this reason, an amendment to the standard will authorize the use of IEC 61000-3-11 for some types of products that cannot meet the dmax limit of IEC 61000-3-3 for equipment with rated current =16A. IEC 61000-3-11 applies to equipment with £75-A rated current and subject to conditional connection to the public distribution network. The utility company must approve connection.
In addition, the dmax limit has been raised to 7% for some types of devices not likely to cause annoyance. Mr. Woodgate cited an example of high-power garden tools. They normally do not cause flicker because they are only used during daylight hours.
A standard based on IEC 61000-4-15 but relevant to 120-V circuits is under development by the IEEE P1453 Flicker Task Force. In the United States, IEEE 519-1992 and IEEE 141-1995 recommended practice documents have dealt with flicker, but they are based on very old original research. According to a white paper written by the task force, the actual practice of flicker measurement in the United States varies greatly among electricity suppliers and models of flicker meters.6
The task force’s work is centered on establishing an IEEE standard that will unify flicker measurement as well as IEC 61000-4-15 has done. In addition, the group wants to incorporate recent research into the effect that different types of lighting equipment has on flicker perception.
For example, most of the early research was done using 60-W, 230-V incandescent bulbs. Fluorescent bulbs flicker much less than incandescent bulbs for the same degree of voltage fluctuation. On the other hand, power-line harmonics can cause fluorescent lamp flicker. The 230-V incandescent European lamps have thinner filaments than equivalent power 120-V lamps. This means that more flicker would be observed from the 230-V lamp for the same voltage change.
Mr. Conrad, the P1453 Flicker Task Force chairman, made clear that his group is attempting to get a 120-V, 60-Hz specification included in IEC 61000-4-15. The IEEE then would essentially adopt IEC 61000-4-15 as the flicker-meter standard. However, the IEEE has no plans to adopt product test requirements such as IEC 61000-3-3 and IEC 61000-3-11.
So, How Do I Choose a Meter?
After determining that all the meters you are considering conform to IEC 61000-4-15, the choice becomes complicated. Flicker measurement in accordance with EN 61000-3-3, although mandated since Jan. 1, 2001, as a prerequisite to the CE Marking, usually is provided as part of another instrument.
For example, the TMX Series from Pacific Power Source is termed a one-box solution for harmonic and flicker measurements. On the other hand, Mr. Conrad cautioned that ds (steady-state voltage) and dmax measurements are not part of IEC 61000-4-15 but rather are required to meet IEC 61000-3-3. A flicker meter that correctly implements IEC 61000-4-15 may not measure these quantities.
You will need to decide how extensive the main features of the instrument need to be. For example, are you interested in full compliance or precompliance test capabilities? Herman van Eijkelenburg, a product marketing manager at California Instruments, provided additional insight. “Most power analyzers that offer flicker measurements do not display real-time information during the test. This means the user does not know the outcome of a flicker test until a 2-h test time has elapsed.
“Since most flicker meters do not provide a record of the data collected, but only a final result, there is no practical means to determine if the flicker results obtained are correct. The California Instruments Compliance Test System stores all the test data to a PC hard disk,” he continued. “The data can be used to replay the flicker test at a user-specified speed to help look for changes that caused the flicker levels to exceed allowable levels.”
Another important consideration is the capability to upgrade the instrument as more changes are made to the IEC 61000-3 series of specifications. Many flicker meters incorporate the latest specifications and test limits into their software as well as a custom or advanced mode of operation. In this mode, the flicker-meter settings can be changed either to conform to a revised specification or help determine the cause of a test failure.
But, before you consider buying a meter, make sure you really need to test for flicker. Do your products require the CE Marking? Even for those that do, if you can demonstrate that your product’s rated current is very low, and that its power-on inrush also is small, flicker probably is not an issue. Document your argument in the product’s technical file and start selling it.
References
1. Slade, Dr. P.D., “Harmonics and Flicker—The Low-Frequency End of the EMC Spectrum,” UK EMC Journal, www.emc-journal.co.uk/archive1.
2. Harmonic Analysis per CENELEC Amendment-14 to EN/IEC61000-3-2: CENELEC Amendment-14 Explained, California Instruments, 2001.
3. Power Analysis Reference Sheet 011—IEC 868 Flicker Response, Voltech Instruments, VPN:86-163 Issue 1, March 1998.
4. HP Users Guide for 14761A Harmonic and Flicker Emissions Test, Part No. 5962-0831, 1998. rndcentral.tm.agilent.com/rndcentral2/cgi-
bin/epsg.cgi?action=rndcontent&content_id=15&category_id=
59&638314727
5. Woodgate, J.M., EMC-Low Frequency Conducted Emissions: The Truth About IEC61000-3-2 and -3 and Their EN Clones, www.conformity-update.com/iec-61000-000908.htm
6. Halpin, S.M. et al, Voltage and Lamp Flicker Issues: Should the IEEE Adopt the IEC Approach?, IEC/TC77A/WG2/TF1 Flicker Task Force white paper, grouper.ieee.org/groups/1453/drpaper.html
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents © 2001 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
October 2001