LED projects

Infrared light head with 16 high power LEDs for long range night vision applications. The IR16TU has an interference filter and is non visible for the human eye even when you are directly spoted, no red glow! IR16TU.pdf

The light head can be combined with different LEDs but also with different lens types. Shown frozen flood. See output pattern.

LED head designed for powers from 3 up to 10W. length: 80mm, diameter: 35mm out of black anodized aluminium.
Mounting bolt M8 stainless steal. Silicon potted. Light versions: IR-cluster, white spot, flood or elliptical. Version shown is elliptical used as headlight for rowing boats being seen long distance and by wide angle.
Input voltage: 6-32V
Thermal LED control and regulation
polarity protection: yes
Remote: yes
Status LED for low battery: yes
MCU PIC16F688
DIMM frequency 350Hz

LED buck regulator PCB, 28x28 mm, with special features:
there is no LED driver IC, the PWM driving is generated with digital settings. The micro is placed on the bottom side a PIC16F1765. The intergrated comparator does the current regulation. The software inside the micro is managing everything else. The software determines directly the temperature of the LED by measuring the remaining discharge voltage over the LED when it is turned off. Cable length to the LED is no problem. The circuit has a special active turn off control of a bipolar PNP transistor (protection applied). So there is no MOSFET we use an inexpensive PBSS5560PA from NXP: 60V/5A with 2.1x2.3 mm size. Bipolar transistors can still work pretty well in special applications and sometimes even better than FETs, because they don't need a minimum drive voltage. A micro is able to dimm, can remember the dimm value, can detect the type of LED, he monitors the LED temperature, he can communicate with a PC and delivers a report. Just the best combination of digital and analog. It's more safe, traceable and it has a conclusion more quality for the customer AND the manufacturer than a normal LED IC can offer.

CXA3050 attached by foil on a block

The Cree chip has 156 single LEDs: 12 x 13.

New: LED Power driver PCB for CREE CXA3050 50W 6000 Lumen. Temperatur of the LED is directly measured and controled.

LEDs requires a very well thermal coupling to the cooling block. The array plates are made of very good thermal conductive ceramic. The plates should be bressed by clips on to the block. In between the plate and the block a quite soft but very effective thermal coupling should be made. For testing the thermal management there is an extra reference point on the plate for solder a thermal sensor directly on top. No good performing TIMs can lead to very high temperature drops over it resulting quite fast in damaging the LEDs. Only by a direct constant monitoring by Nucon software solution you can grant a non thermal overload if not using such thermal sensor on top of the LED.

The thermal managing concept is quite similar to the thermal management you have inside every PC to monitor the processors heat and control the fan speed over the BIOS of the mother board. Dust and aging of the TIM material, which can dry out , makes a monitoring as a must have and state of the art technology.

The array LEDs are offering a replacement of the plates without the need to solder. Best thermal performance of TIM is possible by foils of new designed carbon. They do not need to use any kind of glue and provide a non aging soft elastic solution without any mechanic stress. However the setup may differ in perfomance or even the TIM was forgotten to be installed. So the quality can be only measured by extra wire on top of the Tc point or more smarter controlled by Nucon LED software solution.

When using a temperature sensor mounted on the cooling block not far away of the LED it is a not very good solution, because there is a considerable delay involved. When writing software for clipping rise of temperature it may result in a poor off on control switch. If the software is too slow you may turn off already overheated the LED. When the cooling block is designed the delay of reaction must be considered.

Nucon software is reaching a temperature precision of about plus minus 2 centigrades over the full temp range,

Because the measuremnt of temperature is the LED itself there is a fast and effective regulation response always active. The system can not be cheaten.

The Nucon software solution does not only saving the extra temp sensor it makes it possible to down size the cooling block, because of it's sharper and faster direct sensing. This is a big advantage for the using inside mobile applications to save weight and cost.

CXA 3050 with clip holder and graphite sheet as TIM. The arrays do have a thermal reference point to measure Tcase by soldering a thermal sensor directly on to it (green wire).

LEDs are meanwhile in many different types and versions available. For the project it is now the most important to find the best combination of lens and LED. LEDs are mono chromatic. In case there is need for white there must be a mixing of LEDs or there is a must have for a luminescent material, converting most part of monochromatic blue into yellow. Here we do look at so called white LEDs based on luminescent material and their performance:

Avago ASMT-Jx3x @ 3W

This LED gives in a spot reflektor a superspot, but as a bad suprise only blue. The yellow light is scattered out of the center seems to be a strange behaviour?
Reason: The blue emitting chip is located at optimium focus position of the lens and is therefore perfectly bundled. The yellow light is generated in the full volume of the half sphere and can not be bundled for that simple reason. The yellow part of the LED is not acting as a filter - it is source of luminance. So as a bad result we obtain less focal light intensity compared to weaker LED powers of 3W.

This is the spot of famous known Hokus Fokus by Two Brothers. There are no chromatic errors, why? The most important reason for the good performance is that the luminance material is not distributed into the silicon dome it is just directly on the chip and the dome is clear. Hokus Fokus wins easily against the blue spot of the Avago with much less electrical power. This negative blue performance example you find much more often and it has nothing todo with a very high color temperature of the chip itself it is just the pure combination of specific LED type and optical lense. Now you may know why your bike light guides you into a blue hole.....

Since capacity can be replaced only by more capacity :-)) now there is a state of the art big CREE XMLBWT in the same reflector installed with power little bit higher than 3W. This chip has the luminance material just on the chip and has a clear dome on top, but it has a very big chip size now. The picture shows now all three LEDs in action: left side Hokus Fokus, middle the strong XML and on the right the weak blue of the Avago.
For sure those pictures are just relative comparisons and the camera as well as the human eyes adopts to a new white balance in darkness. But to measure is to compare so the pictures are having a statemanet of perfomance.

LED pattern of an ellipsoid lense dimmed down in front with the LEDTUBE. This is the special front light for our race rowing boats to be seen in a wide angle mode from the long distance!

Rowinglight.com

This picture shows the Hokus Fokus in the flood position. You can see in the middle area now a hole of intensity which is moved in a ring to the outside. The problem is, that in most cases your object of interest for fotographie and movies is in the middle located.

Here you can see the advantage of the Avago LED having the luminance material in a wider dome distributed in combination with a frosted flood lense. You obtain smooth transistions of brightness from the center to the outside. Even still areas corrsponding to a very wide angle are getting some light, which let's you orientade better in a full darkened room.

Extreme long distance spot lights are not easy to be designed. It is better to use other LED types than Cree XML when using lense diameters 26-35 mmto get all the light into the spot and not into the scattering section. On the right you can see a cluster prototype which may be used later on for MIL applications. If bigger XML chips are used you need lenses 45mm and beyond to obtain good results.

Retrofit of halogen lights systems can be done if the LED is with the chip dimension small enough to be within the limits of the lense system or of the reflector. If the halogen filament was small, there can be placed a single emitter source only and not such junk shown on the left ;-)