LED Spots for Lightening A Sales Hall – Part 2 of 4 – The Choice of LED and of the Correct Heat Sink

In the last part of this Blogreihe, I have explained the general situation and described what is to do. Also I have a series of LEDs listed coming from the light output in question and also which excluded. You can find out which I’ve decided now in part 2 of my Blogreihe:
After really many hours and nights, I decided MK-R then for the CREE. Why can you learn in this post. Also I did me some thoughts for proper cooling. Mean everything in this post series of measurements with different heat sinks and much more -!

The Choice of LED and of the Correct Heat Sink

Why I the CREE MK-R LED Prefer for This Project

  1. The LED is supplied on a star PCB and is therefore very suitable for me as hobbyists and developers to try out.
  2. The LED chip (ok are actually 4 chips connected in series) is not much bigger than a normal XML.The Board has even the same mass.
  3. The performance data have it done to me, because CREE campaigns is to create with the MK-R as the first LED, which 200 lumens per watt.(More info at the manufacturer) That but only at very low currents. It is also relatively cheap compared to the rest with only €15.

Light currents for Various Amperages

The table values at 85 ° C reference temperature of emitted light performance:

0,7 A 1 A 1,25 A
11,6 V 12 V 12,15 V
8,12 W 12 W 15,19 W
1040 l m 1350 l m 1664 l m
128 l m/W 112, 5 l m/W 109, 6 l m/W

(lm stands for lumen)
That really fascinated me. Would I keep the LED below 85 ° C, I would get a small bonus to light again. A LED is cool, the more relative light output emits it. Would I now maintain 25 ° c the LED with a water cooling, I would get up to 20% more light output at same electrical power consumption than at 85 ° C.
Now is just the question of how I can cool this but very high LED performance. There are many LED sellers advertise 12 W or even 15 W led to 9 W but that the heat sink is not much bigger than the others with 3 W in eBay. Alone it can be seen already, that because something must be wrong. After all, one must assume that about 1/6 tel of the electrical power is transformed into light. 5/6 tel in heat (at a 100 lumen per watt LED).
That would mean that I would convert electrical power into heat at 12 W MK-R solution 10 W and this heat must give environment more quickly and effectively to the. To do this, I have already a detailed FAQ post.
Here a comparison: our site
so something can there not quite agree…

My Thoughts on the Heat Sink

Many say yes, the heat sink is heavy, the better it works also. I have to disagree with this statement. The weight, so the mass of the heat sink is absolutely irrelevant to the performance of the heat sink. In static mode, i.e. in continuous operation which plays absolutely no role, but only the surface and the thermal conductivity of the heat sink. So the more surface of the heat sink has, which may come in contact with the air, the better it also cools. The thermal conductivity is equivalent. Better heat can be dissipated, the better the heat sink cools down. More mass is only a subjective good cooling, because if a heat sink has more mass, you need with the same thermal performance (thermal performance is meant here when e.g. heating 5 W converts electrical energy into pure heat) longer to warm up to him. He stays longer “cool”. First time shows how well the heat sink is really.

For the people who now don’t imagine that with heat conduction, I have here a small example: If you go in the 100 ° C warm sauna, sit on wooden benches or 100 ° C warm wooden infusion spoon also touched. Why build the banks not of metal? Quite simply, anyone who would sit on metal benches would extremely burn themselves, although they had also “only” 100 ° C. The key here is 1. In the warmth and the 2. In heat conduction. One touch wood, the heat can be passed not so close to the body. The hand abruptly cools the wood. The wood can pass but the heat from the environment not so good in the hand. Metal would pass on the energy of the warm air to the hand and therefore extremely fast heat up the hand. Wood has 1. A lower heat capacity and 2. A very low heat conduction. The lower the conduction of a material, the better suited’s as an insulator (see rock wool for the roof, or styrofoam)
An example of a heat sink with a large surface is here. In contrast, here ‘s a heat sink with less surface.
Can be seen with the bare eye, that the large heat sink has significantly more space. He has more cooling fins and this can deliver significantly more heat to the environment. It has therefore a reason why a 10 W LEDs is suitable and the other only for 3 W.

I decided a heat sink with a 90 mm diameter and 10 mm thick and a total of 52 individual cooling fins for a. That however is not enough. Yet somehow, I need a device to hang the heat sink. That’s why I thought about me to build this LED apart and to apply the heat sink on my other heat conductive glue. Thus, I have a recording with a GU 10 base, and a connection from my large heat sink to my recording.
I expect better cooling performance but didn’t, because the surface of the outer ring of the heat sink is not just great and the outer ring also on the heat sink is screwed up. The heat transfer should be therefore not exactly optimal. It is only the appearance and the installation.
I thought about yet also me what happens if I put together 2 of the large heat sinks.

So, I’ve Started a Series of Measurements.

Here the three versions:

  1. A large heat sink alone
  2. Pasted a large heat sink with a small heat sink as recording
  3. Two large heat sinks

Measured I have the whole thing with 700 m A power supply as well as 1 A power supply. So even with approx. 8 W LED performance and once with 12 W LED performance

The Measurement Results:

  1. Measurements with a heat sink:
    • Measurement of 0,7 A: LED temperature 67.6 ° C
    • Measurement at 1 A: LED temperature 75 ° C
  2. Series of measurements with the glued-on heat sink + great:
    • At 0,7 A: 62 ° C
    • At 1 A: 70 ° C
  3. Measuring range with 2 large heat sinks
    • At 0,7 A: LED temperature: 60 ° C
      • The upper heat sink on the metal measured: 46 ° C
      • Measure the bottom heat sink on the metal: 44 ° C
    • At 1 A: LED temperature: 68 ° C
      • The upper heat sink on the metal: 52 ° C
      • The lower heat sink on the metal: 50 ° C

You can already see that the small heat sink has however a good additional cooling effect and can keep up even very well with the two large heat sinks. Measured I did the whole thing with so a conventional multimeter with temperature sensor (bimetal sensor). The heat-sensitive bimetal was held directly at the LED chip and thus measure the real temperature of the chip.
It was striking that I came with none of my measurements on chip temperature of 85 ° C. That means I’d have to still a bonus of 5% at 70 ° C chip temperature of the count given luminous flux.

Visit Us Again Next Week for Part 3

That’s it now with the 2nd part of the Blogreihe. In section 3, you which driver will end up energized the led (s) and why my base solution but does not work erfahret. So you miss nothing also, If you subscribe to our RSS feed.
All posts of the series are now on line:

More from this series of articles

  • Part 1: the project whitening a sales Hall
  • Part 2: The choice of the LED and the heat sink
  • Part 3: Housing, lens and costs
  • Part 4: The shadow and the diffusion disc
  • Part 1: The order and the first considerations and plannings