MC-E, best cooling with or without starboard?

[HansV]

I just wonder if using a starboard on a Cree MC-E will give better or worse cooling that putting the emitter directly on to the heatsink?
The starboard has the advantage that the heat pad under the emitter has a soldered connection to the board. The starboard will also help to spread the heat.
On the other hand, the starboard will add up on the thermal resistance between the emitter and the flashlight.
Mounting the emitter without starboard should be easy enough. Juse bending the pins up a bit, solder wires directly on and put on some capton tape on the sides as isolation if neccesary.
I do not know what is the best solution, starboards are of cause easy to mount but that is no issue when optimazing performance.
The goal is to reduce the initial output drop after the light is turned on, before the whole flashlight is heated up, as much as possible.
I also have plans for a 5xMC-E in a MAG insert, that can give a output close to 100lumen/W at 20W drive.

 

[HarryN]

Hi, just so I understand, your goal is to run each MC-E at 4 watts, so aournd 1 amp per package?

From a lumens efficiency perspective, this is a great idea, and it makes a lot of sense for general lighting or perhaps a very floody light.

The only challenge will be that you are driving the individual die at such a low power, they will have very low surface brightness. This will result in relatively low Lux numbers (very limited throw) This would be a good light for lighting 10 - 20 meters, but probably not 50 meters.

As far as mounting the MC-E, most people underestimate just how much pressure is needed to keep the thermal transfer interface very thin at the thermal pad. If you are really good at mounting the emitters, then the star board is not as good, but the difference in the two approaches starts to be "user dependent" really fast.

 

[HansV]

I was thinking more general about the mounting. One of my flashlights is a Ultrafire WF-501B with a MC-E insert on a 14mm board directly driven form a 18650 cell. There I am running almost 4A with a fresh battery, and that works fine because I have optimiced the cooling all the way. But I was thinking that I might get even better cooling without the mounting board. I can apply a lot of preassure with this insert by screewing it tight.

What do you mean by surface brightness? Can the usefull brightenss be anywhere else? I though the lux value you get is a product of the total ligth output, spead angle and distance? I don't follow your line of though.

 

[HarryN]

Hi - The LEDs surface brightness, sort of a "lumens per mm2" has a large effect on the ultimate Lux level. Suppose you drive 1 amp through an MC-E, divided into 4 die, that is 250ma / die.

Now take the same 1 amp and drive a single emitter, say a Cree XP-E. The XP-E will emit fewer lumens, but its surface will be much brighter than the MC-E. This directly links to how far the beam will "throw" (lux)

If you have the ability to collect all of the light from both sources and collimnate them into the exact same spot area, then in theory, it comes out better to have the larger die. In practice, this is very hard to achieve, so your under driven MC-E will "TEND" to lack throw.

This is one of the reasons why many incan bulb based lights can so easily out throw many well designed LED lights. It is not the "total lumens", it is how bright that tiny filament is.

This is easy to test on your work bench. Just setup a few LEDs, incan bulbs and various optics / reflectors and see for yourself. It is an annoying but very real effect.

Coming back to your thermal path question - it is probably best if you can solder your thermal pad directly to the heat sink. Rebels can do this, but I am not sure about the MC-E.