NPL System for the Calibration of Flickermeters
If the EUT fails to meet the flicker requirements of the standard, this will prevent the appliance from gaining access to important markets. This can lead to costly redesign and delays in marketing a new product. In such cases, in particular if the failure is 'borderline', attention is given to the accuracy of the Flickermeter. This need has led NPL to develop and offer a calibration service for Flickermeters.
The Flickermeter output depends on the depth of modulation and rate of change of the mains type signal applied to its input. In order to calibrate the Flickermeter it is necessary to characterize such a signal and compare the theoretical response of an ideal Flickermeter to the response of the device being calibrated.
There are two methods used at NPL for calibrating Flickermeters. The first of these, which uses basic squarewave modulation, is described here. The second uses a reference Flickermeter, which is an NPL implementation of the IEC 61000-4-15 Flickermeter. This is used to calibrate Flickermeters under more general conditions with more complex waveforms. This second method is described in Calibration of Flickermeters Using More Complex Modulation Signals.
A simple fluctuating mains-level voltage waveform can be obtained by amplitude modulating using a squarewave as shown in Figure 4. This gives rise to two distinct RMS levels which occur at a given repetition rate (changes per minute).
Figure 4: The Squarewave modulation of the mains voltage as used to calibrate a Flickermeter
For the special case of squarewave modulation, IEC61000-3-3 contains a graph to relate the changes per minute, to the depth of variation in voltage (dV), such that the value of Flicker severity is unity. By choosing points from this theoretical response graph, the Flickermeter can be calibrated over a range of modulation rates and depth. IEC 61000-4-15 tabulates six points from the graph for assessment purposes.
Generation and Analysis of the Calibration Signal
The modulated calibration signal can be produced by a commercially available waveform synthesizer with an amplitude modulation (AM) input. The synthesizer must be of sufficient quality to produce a stable amplitude and low distortion mains frequency signal. A second waveform generator is then used to produce a square wave which is applied to the AM input. It is desirable to phase lock the two waveform generator time bases to ensure the modulation frequency has a fixed frequency and phase relationship to the mains carrier frequency. The system is shown in Figure 5.
Figure 5: System for calibration of Flickermeters
The signal is then amplified using a laboratory quality power amplifier to give the required mains voltage level for application to the Flickermeter being calibrated.
Having chosen a point on the Flickermeter curve and set the required repeat rate on the squarewave generator, the calibration signal amplitude is then independently measured and adjusted to the required voltage depth by changing the amplitude on the squarewave generator.
In order to measure the calibration signal the voltage is stepped down to 1 V rms using an inductive voltage divider (IVD). This signal is applied to the input of a calibrated analogue to digital converter (ADC) of NPL design. The data analysis method used depends on the repeat rate of the modulation.
Use of the Squarewave Calibration Signal to Calibrate the Flickermeter
Having set-up the calibration signal to the required depth and rate, the Flickermeter is used to make a 10 minute Pst measurement. For the points used on the theoretical curve, this result should give a value of one. The allowed tolerance for Flickermeters as defined in IEC 61000-4-15 is ± 5 % of reading. The type-b uncertainty on the measured value of Pst is typically 0.01 Pst units.
At a given repeat rate, the Flickermeter should have a linear amplitude response, such that if the modulation is increased, the Pst reading should increase by the same factor. Tests are carried out to ensure that the device being calibrated is within the allowed tolerance up to a Pst of five.
A number of pitfalls exist regarding the generation of the signal. These include dc level, rise-time of the squarewave through the system, mains beat problems and timing accuracy. In general, these can be readily assessed and either corrected or accounted for in the measurement uncertainties.