WO2012171531A1

WO2012171531A1 - Led fault diagnostic method and system

Info

Publication number: WO2012171531A1
Application number: PCT/DK2012/050193
Authority: WO
Inventors: Anders Rand Andersen
Original assignee: Emazys Technologies Aps
Priority date: 2011-06-17
Filing date: 2012-06-03
Publication date: 2012-12-20
Also published as: DK201100460A; DK177516B1

Abstract

There is provided a method for fault diagnosis on an LED armature in which electrical potentials are checked within the LED armature using impedance spectroscopy. Specifically the LED armature is excited by both a DC bias and an AC voltage over a wide frequency range, and the impedance of the LED armature is measured as a function of the frequency response.

Description

LED fault diagnostic method and system

FIELD OF THE INVENTION

This invention relates to a fault diagnostic method and system of an LED lamp system having an LED matrix comprising a plurality of interconnected LEDs.

BACKGROUND OF THE INVENTION

Usually, a light armature has a control device for monitoring current that, for example, indicates, in a case of controlled filament lamps, whether a filament lamp must be changed. This is problematic when the controlled lamps are LED-based lamps which because of their rate of response, reliability, and design possibilities. Furthermore it appears that LED lamps are increasingly replacing filament lamps. In order to meet light values required by law, it is necessary to control, or drive, several LEDs simultaneously. In a known manner, serial and parallel connections are used. Control occurs via the light control armature, which monitors current consumption of the lamp for short circuits and open circuits. This results in the following problem: when using an LED lamp, the light control armature cannot recognize a failure of one individual LED or some plural of LED connected in an electronic circuit, by simply monitoring current consumption, because of relevant tolerances and, where applicable, because of double usages of individual LEDs in different light functions. To provide diagnostic data to the light control armature, it is possible to provide a diagnostic output at the LED lamp that is conveyed to the light control armature via a diagnostic wire. Such an arrangement, however, proves to be quite costly.

The wiring cost for an additional diagnostic wire for each LED lamp entails additional expenses; which is also the case for components for issuing the diagnostic signals (output driver, perhaps even a microprocessor) within the LED lamp, with these additional expenses accruing even if the diagnostic capability is not even used in a vehicle. It is therefore an object of this invention to provide a diagnostic method and system for an LED lamp armature that is particularly uncomplicated and inexpensive in construction. The most common way for LEDs to fail is the gradual lowering of light output and loss of efficiency. Sudden failures, however rare, can occur as well. Early red LEDs were notable for their short lifetime. The failure could be due to epoxy degradation; some materials of the plastic package tend to yellow when subjected to heat, causing partial absorption (and therefore loss of LED efficiency) of the affected wavelengths. Also thermal strees can give rise to sudden failures. When the epoxy resin package reaches its glass transition temperature, it starts rapidly expanding, causing mechanical stresses on the semiconductor and the bonded contact, weakening it or even tearing it off. Conversely, very low temperatures can cause cracking of the packaging. Further differentiated phosphor degeneration may be the reason for a failed LED. The different phosphors used in white LEDs tend to degrade with heat and age, but at different rates causing changes in the produced light color, for example, purple and pink LEDs often use an organic phosphor formulation which may degrade after just a few hours of operation causing a major shift in output color. Also nucleation and growth of disloscations is a known mechanism for degradation of the active region, where the radiative recombination occurs. This requires a presence of an existing defect in the crystal and is accelerated by heat, high current density, and emitted light.

AC impedance spectroscopy has been used in the prior art to test components in integrated circuits (IC). In such setups the components (transistors etc.) are tested during or after the manufacturing process, and by measuring the real and imaginary part of the complex impedance one can construct a Nyquist plot and use such graph for identifying various defects or flaws in the manufacturing of the component. Surprisingly when the object under test is not a simple IC but an entire array of light emitting diodes (LED), a so called LED armature, the application of AC impedance spectroscopy in a wide frequency interval while applying a steady state potential or a current, can led to fault detection. Furthermore the AC impedance spectroscopy is ideal as a monitoring tool that is in use during the life cycle of the LED armature.

A major issue found when using LED armature for industrial applications such as greenhouse illumination is that the LED develops heat and ultimately single LED in the armature starts to break down as the polymer materials in the LED and the mounting board tears. In the greenhouse the LED degradation process is accelerated due to high levels of moisture, which will set on corrosion processes once the polymer material is partly open to the atmosphere.

Another application of LED armature is in the form of LED rods, which can be used to grow algae in marine algae farms. In this case the LED armature is housed in a transparent tube and submerged in the sea or a dedicated water tank. The AC impedance method will in this case very quickly aid in detecting defects involving flooding of the tube that sets on corrosion in the metallic parts of the LED. Such corrosion will lead to significant changes in the electrical properties of the LED and its interconnection, hence such damage is easily measured using the AC impedance spectroscopy method.

The break down of an LED typically happens slowly as a degradation process. For instance involving the deterioration of the polymer materials, which leads to diminishing illumination from the LED in terms of intensity and further the emitted light spectra can be shifted. This effect is highly undesired when the LED is used for promoting biological photosynthesis, as this application is based on the specific emission spectra and intensity of the LED. An interesting feature that occurs when testing partially degraded LED armatures, is that the test can lead to the coupling between the alterations in the illumination as a function of the electrical parameters. This directly leads to the possibility of determining the state of degradation i.e. the system can estimate the quality of LED light as a function of electrical parameters. This means that monitoring of LED using AC impedance spectroscopy will also imply that the user can estimate boundaries for how much degradation that can be allowed, and hence maintenance of LED armatures is highly optimized.

A highly unexpected technical property is thus, that optimization and characterization of the LED based artificial light environment can be carried out without using optical measurements, because the monitoring is solely based on automatically acquired electrical data.

For gaining effective features to realize fault diagnosis in LED illuminating circuits, a new method of fault diagnosis is necessary. It is therefore an object of the present invention to provide a more reliable method for fault diagnosis of LED armatures. SUMMARY OF THE INVENTION Specifically the present invention provides a method for diagnosing failure modes within an LED system including one or more LED armatures, said method comprising the steps: i) applying a constant electric potential with the power supply in the form of a DC signal across/through the LED array, said potential being in the range of the power supply voltage V_s which is typically in the range 100 - 400 V DC for a single armature (potentially more or less depending on the specific armature and LED type)

ii) applying in addition to the DC signal of i) an AC voltage, while scanning a frequency range from 1 kHz to 1 GHz to achieve an impedance spectrum;

iii) comparing the impedance spectrum with a control impedance spectrum recorded from an intact LED system or a previously measured impedance spectrum on a subunit of the entire LED system; and

iv) detecting and identifying faults by detection of significant changes in model parameters, when numerically fitting a physical model to the measured electrical data.

The present invention provides a method for the detection of material-specific changes in interconnected LEDs armature that temporally evolves in the processed materials of the armatures and their interconnections, effected by outside and inner causes, before the occurrence of larger damages are recognizable.

In response of the inner and outside inductances, capacitances and resistances the ageing condition produce a characteristic frequency response of the impedance (impedance spectrum), for a given design of the devices. These outside and inner influences are e.g. the UV irradiation, temperature, temperature changes, the concentration of humidity and duration thereof. In particular according to the present invention the course of the impedance as function of an AC and DC in series or parallel,

• applied between the positive and the negative terminal of the LED armature,

• or armatures interconnected in matrix and/or interconnected armature strings in direct contact. In particular are also provided that the impedance for the characterization of resultant changes and for the early recognition of loss of efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows the minimum LED circuit comprising the power supply which is a DC voltage source V, a resistive element R, and the LED connected at its Anode and Cathode. Fig. 2 shows the system for analysis in a schematic form.

DETAILED DESCRIPTION OF THE INVENTION The characterization system is comprised by a power supply, a frequenzy generator and a frequency response analyser (FRA). The power supply is within this invention used as BIAS potentiostat that will supply either a constant DC voltage or current while the frequency generator is superimposing a small amplitude AC voltage V_Ac, across the LED armature. The FRA is able to measure and record the data. The data collected by the FRA is the alternating small amplitude current l_Ac, resulting from the alternating voltage. The impedance of the LED armature Ztotai is defined as the ratio between the two:

Ztotai ⁼ AC/I Ac (1) This implies that the formats that can be stored by the FRA unit are data points given as AC frequency f, phase angle Θ and total impedance IZI.

When the AC voltage is supplied in a broad frequency range we hence obtain an impedance spectrum Z_totai(f). ^and ^tne invention is specifically designed to identify and monitor changes in this physical observable Z_totai(f)-

The concept of equivalent circuits is used for failure mode identification based on analysis of Ztotai(f). Figure X?. Such a model typically employs several equivalent electric circuit components such as resistance elements R_n, inductive elements L_n and capacitive elements C_n. Specifically the model also contains an element equivalent to the array of interconnected photodiodes. This element is specific for the type of LED armature in question and depends on the impedance the single diodes used in the armature and the connection scheme.

The circuit shown in Figure 1 shows only the minimum example where a single LED is connected to a power source. Equivalent circuits used to model real armatures will have a pronounced complexity in comparison.

The model can be translated into mathematical equations based on circuit analysis, and hence a numerical fitting procedure to the impedance data Z_totai(f) can be carried out. The result is a series of model parameters, and it is specifically these model parameters that can be used to identify the origin of changes in Z_totai(f)-

In Figure 1 the minimum LED circuit is shown comprising the power supply which is a DC voltage source V, a resistive element R, and the LED connected at its Anode and Cathode.

A special feature regarding the measurement principle of the invention is that a potentiostat, either in the form of the LED power supply or in the form of an externally applied potentiostat, can set up a BIAS potential or current. This means that the impedance characterization can be carried out at different BIAS settings. A semiconductor device such as an LED will typically have abrupt changes in conductivity and impedance when the BIAS is changed within a narrow interval in the vicinity of the DC working point (V_s, I) of the device. This principle is found surprisingly use full when using the impedance method is used as a probe. In Figure 2 the main unit is device comprised by a potentistat (and or a power supply) which will feed power into the LED armatures and this keep the armature at a fixed DC potential. The AC frequency generator is able to send a small amplitude oscillating voltage (or current) signal, and the frequency response analyser is able to measure and record the resulting AC current (or voltage). Data is stored and analysed in the server unit.

Automated failure mode detection and alarm

The characterization system is within this invention automated so that the measurement units (potentiostat, frequency generator and FRA) are connected to a central computer system. The central computer system is responsible for storing measured data, and for carrying out the numerical fitting procedure. The main data stored in the central computer is hence the model parameters extracted from the fitting procedure. The automated failure mode detection principle is based on identifying whenever an extracted model parameter is significantly changed in comparison to previously recorded data. The central computer is considered a server that will thus be able to interact with the characterization system and automatically perform testing.

EXAMPLE

One way of carrying out the invention is to use a commercial function generator for the small amplitude AC signals, see Fig 2 (AC frequency generator). The DC voltage, see Fig. 2 (power supply/potentiostat) can be delivered by a voltage supply that is an integral part of the LED armature system or it can be delivered by an external voltage source. The poles from the AC and DC generators are connected to the terminals of the LED armature(s) in the voltages are supplied in a series connection. The resulting currents that will flow through the LED armature is measured using a frequency response analyzer (FRA), see Fig. 2. The FRA can be connected in a parallel connection relative to the AC and DC generators. The FRA should be a device that is able to record the electric current with a high time resolution in order to be able to track the many details of the signal. The current signal will be a complex wave depending on the condition of the individual LED^'s in the armature. The measured current signal and the AC signal can be considered as two waves that are intrinsically connected. The analysis of the current signal is thus based on a comparison with the AC signal, which is a probing signal that is not altered over time.

A number of statistical and mathematical analysis methods can be chosen. One choice is to construct an equivalent circuit that describe the LED armature with a satisfactory precision, and secondly state the mathematical system that is used for analysis in some form of mathematical software routine. The mathematical model, typically an equation that relates current and voltage, can then be fitted to the measured data. This action will then yield a set of model parameters. Ultimately, the parameters may change over time as the recorded current signal changes due to damage and failure modes in the LED armature. The very simple monitoring of a set of model parameters as a function of time is the essence of the invention.

Claims

1 . Method for diagnosing failure modes within an LED system including one or more LED armatures, said method comprising the steps:

i) applying a constant electric potential in the form of a DC signal across/through the LED armatures, said potential being in the range of 100-400 V DC, depending on the powersupply voltage V_s Volts DC;

iv) diagnosing faults by detection of significant changes in model parameters, when numerically fitting a physical model to the measured electrical data.

2. Method according to claim 1 , wherein deviations in recorded spectra in the frequency range 1 kHz - 1GHz are detected automatically by a computer program and used for further analysis of the system 3. Method according to claim 1 or 2, wherein the AC voltage has an amplitude of 0.1 -100 V AC and over a frequency range between 1 kHz and 1 GHz.