The NPL Reference Flickermeter is used to calibrate commercial flickermeters by comparing readings obtained when a range of waveforms are applied.

The waveforms are intended to test that the flickermeter under test correctly implements the specifications given in IEC 61000-4-15: 2003 ^[1]. The waveforms are diverse and difficult to characterise. The only way to determine the correct flicker readings when these waveforms are applied is to measure flicker directly using a reference flickermeter. A reference flickermeter has been developed at NPL and the details of its design are given here.

The correct performance of the Reference Flickermeter has been verified using waveforms having known or calculable responses. The results of this are given in Verification of NPL Flickermeter

A description of the waveforms used for calibrations is given in the NPL Flicker Waveform Library. For a complete description of the methods used at NPL to calibrate flickermeters, please see NPL Flickermeter Calibration Service.

IEC Flickermeter Design Specifications

The IEC 61000–4–15 flickermeter design is broken down into five blocks:

Block 1

The applied voltage signal is stepped down using an appropriate transducer and sampled using an analogue to digital converter at a sampling rate of 50 kHz.

As explained in the ‘Pre-test’ section below, the flickermeter relies on a Scaling Reference Voltage value to scale the flicker readings. This reference voltage is calculated by taking the half-cycle RMS voltage values over a one-minute period (including the pre-test period) and averaging. This one-minute sliding average value is updated every half-cycle of the applied voltage waveform.

Block 2

The output from block 1 is then squared.  This block is intended to simulate the behaviour of an incandescent lamp to the input voltage signal.

Block 3

Block 3 consists of a cascade of three filters, which are designed to simulate the response of the eye to input from an incandescent lamp ^[2].  The first is a first order high pass filter with a 3 dB cut-off frequency of 0.05 Hz.  The second is a 6th order low pass Butterworth filter with a 3 dB cut-off frequency of 35 Hz, which is implemented as three cascades of second order filters. The third filter has a transfer function defined in IEC 61000-4-15, which is implemented as two cascades of second order filters.

Block 4

The output of block 3 is squared and then filtered using a first order low-pass filter with a time constant of 0.3 s.  This filter is designed to represent the storage effect in the brain [2].  The output of block 4 is normalised to give a maximum value of 1 during a 10 minute flicker test, when a sine wave modulated 50 Hz input signal, with a modulation frequency of 8.8 Hz and a modulation depth of 0.25 %, is applied.

For a full description of the design and implementation of the filters in Blocks 3 and 4 please see Flickermeter filter design and implementation.

Block 5

The final PST reading is evaluated in this block.  A Cumulative Probability Function (CPF) is evaluated by classifying the output of block 4 into a number of bins, which have a logarithmic spacing.  This CPF is used to obtain the flicker levels exceeded for various percentages of the test time.  The final PST reading is calculated using the equation given in IEC 61000-4-15: 2003.

For a complete description of the implementation of the classification routine, please see Block 5 classification routine.

Pre-test

A Flickermeter uses filters (described in Blocks 3 and 4 above) to determine its results and relies on a Scaling Reference Voltage value (described in Block 1 above), which is used to scale the flicker readings. The method that a flickermeter should use to process initial conditions, such as pre-test time, initialisation of filters and initial determination of a reference voltage, is not specified in IEC 61000‑4‑15: 2003.

The filters of Blocks 3 and 4 must be initialised, or charged, so that they are in a steady state condition, prior to beginning a flicker test to avoid excessively high flicker readings. IEC 61000-4-15: 2003 does not specify a pre-test time to allow for this. If a modulated voltage is applied to the filters it will take time for them to reach a stable condition. An example of this for 8.8 Hz sinusoidal modulation is given in Figure 1, the filter charging characteristic can be seen.

 

Figure 1 – Typical flicker meter output (8.8 Hz, sinusoidal modulation)

It can be seen that the filters reach an approximately steady state condition after around 20 seconds and hence this is chosen as the pre-test time (a longer pre-test is needed for higher accuracy verification, for example when comparing simulated responses to theoretical responses). During this pre-test all flicker readings are rejected and not used in the classification of Block 5.

The pre-test is also used in the determination of the Scaling Reference Voltage described in Block 1 above. The Reference Voltage is calculated every half-cycle of the applied voltage by averaging over all the half-cycle values up to one minute. I.e. for the first minute of the test all the half-cycle RMS values are used, but after this all the values from the past minute are used. The value is hence updated every half-cycle.

As explained, the implementation of a pre-test period is unspecified in IEC 61000-4-15: 2003, but can have a significant effect on the eventual flicker readings.

References

[1]    IEC 61000–4–15, Electromagnetic Compatibility (EMC) – Testing and measurement techniques – Flickermeter – Functional and design specifications, Published by The International Electrotechnical Commission, Amendment 1, 2003.

[2]    Calculating a New Reference Point for the IEC Standard Flickermeter, W. Mombauer, European Trans. On Electrical Power, Vol. 8, Part 6, December 1998, pp 429–436.