Theoretical considerations for the design of self-diagnostic circuit
for LED based street lightings
Sándor Glisics
1
, Zoltán Kovács
2
, Gábor Marosy
3
, András Poppe
4
1
glisics@eet.bme.hu,
2
kovacsz@eet.bme.hu,
3
marosy@eet.bme.hu
4
poppe@eet.bme.hu
Budapest University of Technology & Economics, Dept. of Electron Devices, PO Box, H-1521 Budapest, Hungary
ABSTRACT: Nowadays street lights employing Solid State
Lightning devices are increasingly gaining ground. This
paper deals with measurement methods of self-diagnostic
electronic circuits for street lamps. Diagnostics is based on
thermal transient measurement of LEDs. The measurement
is achieved via the change in the forward voltage of the p-n
junction. This paper shows and compares three different
measurement schemes. One of the schemes is demonstrated
with a concrete circuit realization. Shunted LED method
was selected for demonstration.
1 Introduction
Nowadays street lights employing Solid State
Lightning devices are increasingly gaining ground. These
light emitting high power diodes are mounted into the
lamp in groups of multiple units [1][2][3].
As these are relatively high power LEDs, producing
heat during their operation, an appropriate cooling must
be installed by the manufacturer. The cooling of light
emitting diodes is usually implemented using the metal
housing of the lamp to conduct the heat. As the lifetime of
an LED is highly dependant on its operating temperature,
it is therefore an important task to ensure that the
temperature stays below a critical limit. The elevated
temperature also decreases the emitted light intensity
[2][3].
Using the temperature dependence of the p-n junction
conclusions can be drawn regarding the heat transfer
paths. Using this quantity it is possible to determine
qualitative quantities of the heat conduction [8].
2 Thermal transient measurement of LEDs
During a thermal transient measurement of LEDs the
measurement of the network of heat resistances and heat
capacitances from the semiconductor to the outside world
is targeted. These values and the time dependence of them
originating from ageing can be determined using heating
and cooling curves. Therefore, based on the cooling and
heating transients it can be decided whether the thermal
resistance of any of the LED's heat transfer path changed
drastically due to transformation, based on which the
degradation of thermal heat conductors (e.g. de-
lamination) might be induced.
Forward Voltage Range
U
LED
Input
dynamic range
U
Time
Heating curve
Figure 1: LED heating curve
The heating curve on Figure 1 can be obtained by
suddenly increasing (stepping) the current flowing
through the LED. As a result of the abrupt current change,
the on-voltage of the LED abruptly increases as well at
the beginning. The following, second phase is the so-
called heating curve. In case of a cooling curve, the high
current flow of the LED is abruptly decreased. Then the
cooling curve can be measured, which is conducted
similarly as for the heating curve, but with the process
running in the opposite direction [2][3][7].
LEDs used in our design have the following
parameters: typical forward voltage is 3,7V, which has
-2mV/°C typical temperature coefficient and 9°C/W
thermal resistance junction to case [9].
3 Measurement principle
During the measurement, a current change which can
be considered as a unit step function is forced on serially
connected LEDs, while the time variance of the on-
voltage is measured. According to measurements, a
function with a rise time of few hundred microseconds
can be considered as an unit step, as it is 2 or 3
magnitudes faster than the dominant time constants [3][8].
Considering the measurement arrangement, it seems
obvious to measure the on-voltage of the LEDs directly
via an A/D converter, as shown on Figure 1.
115
2012 13th Biennial Baltic Electronics Conference (BEC2012)
Tallinn, Estonia, October 3-5, 2012
978-1-4673-2774-9/12/$31.00 ©2012 IEEE