CFD as an abbreviation stands for Compact Flicker Degree.
The CFD is a unit of measure light flicker in percent, which is the Pythagorean Sum of all frequency-dependent weighted frequency components of the AC part of a light signal. The frequency-dependent weighting is proportional to the frequency-dependent perception sensitivity of the human being.
The CFD is the world's first sensible measurement method to measure light modulation which, unlike all other methods, takes into account the amplitudes of all occurring frequencies with regard to the influence on humans and expresses them in a single percentage value.
The results of about 1000 real measurements can be found on the page test
The CFD method is disclosed and not bound by any patent.
Der Lichtpeter has been offering a CFD survey service since August 2016.
In order to evaluate the Compact Flicker Degree CFD in practice, some examples of light signatures from each of the 5 categories are shown below. This is a description so that you can imagine how
the corresponding light signal affects humans.
The CFD of 0% (flicker free) is not shown, in a diagram only a straight line would be seen which corresponds to natural daylight.
The images shown are produced only by artificial light:
This signature corresponds to that of a dimmed incandescent lamp. It is virtually imperceptible to humans, thus low flicker and less than 1% flicker-free. Suitable for general use and a signature under 1% is suitable for filming.
"imperceptible", deep green.
Rarely perceptible to humans.
Suitable for general use (also public places).
Possibly perceptible and after prolonged exposure people may experience discomfort and increased eye strain at work.
Less suitable for general use and much less suitable for longer working. Suitable for street lighting and underground garages.
Old fluorescent tubes with magnetic ballast fall into this category.
Probably perceived by more than 50% of the population, first stroboscopic effects (phase-wise no light, dark). After prolonged exposure discomfort and headache. According to DIN EN 12464 such light is classified as hazardous for work and must be avoided:
"strongly affected", orange.
Perceptible by more than 75% of the population, strong stroboscopic effects (periodically no light, long periods of darkness). Prolonged exposure will lead to
impaired physical condition (headache, feeling of general malaise) and there is risk of epileptic seizure. According to DIN EN 12464 such light is classified as hazardous for work and must be avoided:
"extremely affected", red.
Extreme stroboscopic effects, each movement is only recognizable in individual sequences. Photographs of a moving object will multiply the object on the picture in different places. For discotheques a desired effect. Such light is classified as hazardous according to DIN EN 12464 and must be avoided:
"extremely affected", red.
After discussions with manufacturers and building biologists, it is now agreed that a luminous flux can be regarded as flicker-free if the CFD is less than 1%.
Stroboscopic light (CFD above approx. 50%) is to be avoided at workplaces according to DIN EN 12464-1. This can not be otherwise good. Illuminants that do not meet this standard should actually already have a limitation on usability: "Not suitable for working". Here I clearly see a gap in the EU labeling regulation.
The compact flicker degree CFD takes into account the following characteristics of flickering light:
– alternating amplitude related to its DC portion,
– base frequency,
– waveform (all frequency components),
– light-dark adaptation (contrast from darkest to brightest),
– stroboscopic effect (portions of complete darkness), (phantom array effect),
– Human perception threshold depending on amplitudes and flicker frequencies,
– Easy-to-read measured value for marking on the product packaging.
– Operating modes (dimmed, not dimmed, 12VAC).
In order to determine the CFD, a measurement by means of hardware is required, followed by the calculation.
Categorization for general use is made according to the traffic light system (e.g. for labeling on the product packaging).
The following is recommended for setting up the hardware:
– Use of a V(λ) photo diode for suppressing the infrared light component.
(So the smaller relevant visible light component for incandescent lamps is measured).
– Transimpedance amplifier with variable transimpedances from 10 kΩ to 1 MΩ
to optimal use of the vertical measuring range.
– Anti-aliasing low-pass filter to maintain the sampling theorem, dep. on the sampling frequency.
– ADC sampling frequency: min. 20 kHz (for sufficient presentation and calculation).
– ADC sampling amplitude of 12 bits for a sufficiently vertical resolution.
– Setting the variable transimpedance amplifier so that the peak voltage is approx. 1 V.
– ADC acquisition of at least 1 second duration.
– This results in at least 20000 samples per measurement.
The laboratory system developed and operated by Der Lichtpeter samples the signal at 500 kHz. This way even the highest flicker frequency is detected. The sampling time of 1 second resolves the frequency range into 1 Hz.
The CFD (frequency-based CFDFB) can be calculated with a powerful computer in several steps.
The calculation is shown here in principle. In practice (in the measuring system) other factors are considered such as signal noise and quantization errors.
A discrete Fourier transform is applied to the measured values as:
The result Xk is the frequency domain representation of the N samples designated as xn.
The complex representation of Xk is converted into amplitude values Ak :
These are related to the DC portion (normalized) as:
The frequency-dependent light flicker perceptibility threshold for humans as a function of frequency is obtained from a variety of studies. In addition to the consciously perceptible flicker, the CFD specific curve also takes into account the flicker unconsciously percepted in the peripheral field of vision, the stroboscopic effect and the phantom array effect and , which is lets the observer perceive one object as several objects of the same kind.
The frequency-dependent light flicker perceptibility threshold corresponds approximately to this graph:
The normalized values are scaled to percent and weighted with MW:
Then from the weighted proportions in percent the Pythagorean Sum is calculated which results in the frequency-based CFDFB, which is generally communicated with CFD.
For dimmable illuminants, the CFD value during operation with dimmers is determined in addition to the determination of the CFD without dimming. The dimmer is adjusted so that the average light emission is 25% of the maximum emission without dimmer (a critically assumed value). Care must be taken here to ensure that the dimmer is a universal dimmer or a leading edge phase dimmer (because it is traditionally used in retrofit applications). It is important that it is operated correctly with its minimum load by switching a corresponding power resistor (60 W recommended) parallel to the tested device, which can also be a normal incandescent lamp. The dimmed operation sets higher demands on the illuminant (a lower voltage is applied to the light source for a shorter time), so it is to be expected that the CFD is higher in dimming mode.
The higher (poorer) of both CFD values then forms the final value.
CFD = max(CFDnorm, CFDdimm)
For devices under test operated at 12V and specified for alternating voltage (based on the classical use of 12V halogen lamps), a corresponding iron or ring core transformer is used to obtain a distortion-free sinusoidal waveform of the supply voltage. A switching power supply with a harmonic component is out of the question for use in the test because the influence of the switching power supply distorts the measuring result unmanageably.
: Cree, Inc.: 2014: Flicker happens. But does it have to?
http://www.cree.com/~/media/Files/Cree/LED Components and Modules/XLamp/White Papers/Flicker.pdf
: Lighting Research Center; Rensselaer Polytechnic Institute; May 2012: ASSIST recommends: Flicker Parameters for Reducing Stroboscopic Effects from Solid-state Lighting Systems, Volume 11, Issue 1
: Lighting Research Center; Rensselaer Polytechnic Institute; Oct 2014: ASSIST recommends: Application Considerations Related to stroboscopic Effects from Light Source Flicker, Volume 11, Issue 2
: Lighting Research Center; Rensselaer Polytechnic Institute; Jan 2014: ASSIST: Recommended metric for assessing the direct perception of light source flicker, Volume 11, Issue 3
: IEEE Standards Association (IEEE-SA); 2010: Summary of Scope IEEE PAR1789
: IEEE Standards Association (IEEE-SA); 26.02.2010: A Review of the Literature on Light Flicker: Ergonomics, Biological Attributes, Potential Health Effects, and Methods in Which Some LED Lighting May Introduce Flicker
: Once Inc., Plymouth; 07.07.2015: Minnesota Based ONCE intends to Challenge IEEE 1789 Standard and the Process of Standard Development
: Solid State Lighting Design; 13.07.2015: The IEEE 1789 Standard on Light Flicker a Bit Perplexing
: Michael Poplawski & Naomi J. Miller; 2013: Flicker in Solid-State Lighting: Measurement Techniques, and Proposed Reporting and Application Criteria
http://www.lichtundgesundheit.de/Lichtundgesundheit/Blog/Eintrage/2013/5/31_Auf_wie_vielen_Augen_darf_man_blind_sein_files/Poplawski and Miller CIE Flicker Paper 2013 shorter-1.pdf
: Naomi J. Miller, Brad Lehman; May 2015: FLICKER: Understanding the New IEEE Recommended Practice
CFD calculation software