There is a lot of talk about the power factor of LED lamps, which has an important influence on the electrical behavior. There is quite a bit of chatter about how serious it is with the influence of LED lamps on the power grid. But even people who measure it sometimes don't realize what exactly they are measuring. This article describes a practical situation at an importer of LED lamps who was looking for answers to questions about the how and why of the power factor issue.
It will not have escaped your notice: the LED lamp is on the rise, for various reasons. Manufacturers screen with, among other things, light output, light quality and lifespan. But we are also seeing a price war, especially through the massive influx of Chinese-made lamps. Not necessarily bad, but what do we get? The power and light output that are on the home theatre power manager box, that is what every user looks at. But the power factor is something that few people are guided by when purchasing.
The power factor (PF) is a property of any load in an AC circuit. The PF represents the ratio between the actual power – used to create light and (undesired) heat – and the apparent power – determined by multiplying the RMS values of current and voltage. What remains is the reactive power. This flows back into the net unused.
The higher the PF, the more linear the load. A PF of 1, the maximum value, is a linear load, so a resistor. In that case, both the voltage and the current are a sine wave, and both are in phase. This is ideal, because then all the current is used to also generate power in the load. The other extreme is a PF of 0. Then we speak of an ideal coil or capacitor. Still a sinusoidal current, but now shifted ninety degrees. So current flows, but it is not used in the load. However, this current causes loss in the wiring, which translates into heat. Depending on the drive topology, correction circuits can significantly increase the PF. However, this does come at a cost.
LED lamps are in fact nothing more than small power supplies, and without measures switched power supplies absorb the current with peaks. Because these current peaks have a non-linear relationship to the voltage, this generates higher frequencies (harmonics) that are out of phase with the 50 Hz AC voltage. This is not an actual power, so the PF will be less than 1.
Grey area
A good importer takes responsibility and validates his delivery, as well as looking for standards and regulations. Do all these lamps meet the standard? And what standard? And how do we actually measure this? With these questions, an importer was confronted with a delivery of 4.5 W LED lamps from China, which were specified at a PF of 0.85.
Measuring the PF is usually done with a power analyzer. However, our lamp gives quite different PF values on different meters: with an AC power supply with built-in analyzer, the PF came to a neat 0.89, but with a Zes Zimmer power analyzer and an AC power supply that generates a stabilized and normalized mains voltage, we arrived at a PF of only 0.38! Why the differences? And what is the reality?
Power analyzers come in many shapes and sizes and they give widely varying results under complex loads. Often the measuring equipment is inaccurate at low currents because the shunt is too small. Many power analyzers therefore have difficulty with low LED currents. In addition, the AC supply may become unstable due to capacitive loading. RF ripple on the current can cause unpredictable measurement results if the antialiasing filter is not properly matched to the sample rate of the measuring instrument. And the bandwidth of the measuring instrument may be too large. So when choosing a measuring instrument, several aspects have to be taken into account.
To investigate the differences, we used a high-definition oscilloscope, the Teledyne Lecroy HDO6000. This device features a 12-bit AD converter, allowing high-resolution signals to be viewed. But it is especially important that this oscilloscope has a higher absolute accuracy than the common 8-bit oscilloscopes that are currently available, so that power analysis is also possible. Measurements with the HDO6000 revealed that the stream of this particular LED contains an extreme amount of high-frequency signals (see Figure 1).
Of course a power analyzer is much more accurate than even this HDO6000, but an oscilloscope does provide insight into what exactly is happening with the current, and more importantly: we can perform simulations on the behavior of the power factor when we start filtering the current. For example, we see that the PF improves greatly when the current passes through a low-pass filter. At 2.5 kHz we see the PF already moving towards 0.8 (Figure 2). Let it be the case that it is generally accepted to limit the bandwidth when determining power (and therefore also PF) to the fiftieth harmonic (2.5 kHz). There is a gray area between 2.5 and 9 kHz, but everything above it is EMC.
What does the standard say? The standard that applies in this situation is EN 61000-3-2, for limits on mains harmonics. In class C for lighting, limits above 25 watts only apply. Most LED lamps are below this threshold, so no requirements apply. There is still an Eco-design guideline (Dim2), which requires a minimum PF of 0.4 for our 4.5 W lamp. With its filtered PF of 0.8, it passes the test with flying colors.
Extreme case
So is there no problem at all? Anyway. In the field of the EMC directive (EN 55015) it is a completely different story. Further measurements revealed a frequency of 200 kHz with a very large amplitude. You would rather not place such an LED lamp next to a radio receiver.
It thus appears to be perfectly possible to determine the power factor with a power analyzer with limited bandwidth. The standards in this area don't go any further anyway. But then we shouldn't want to see what else happens.
A broadband measuring instrument does expose the potential issues in other areas that could be even more challenging, such as our lamp, which almost certainly does not meet the EMC standard. The reality depends on the purpose of the measurement.
So what about the PF of LED bulbs? Is it really all that serious? Yes, and no. Our example is an extreme case. Importers of lamps from China in particular have to deal with this type of situation. It is very important for them to know exactly what is being marketed in their name. In this case, measuring is clearly knowing.
But we should also not make the problem bigger than it is. An LED has only a small influence on the total consumption of electrical appliances. As long as LED lamps are not used simultaneously in large numbers, we are in fact talking about very small reactive currents.
Post a Comment