DIY Powerfull LED Dive Light

Pizeer

Newly Enlightened
Joined
Sep 26, 2012
Messages
21
Location
Montreal
The Taskled Flex models have programmable temp shut offs. The only thing is that the driver has to be close to the LED or you have to run the LED until critical temp is reached then program the driver to shut off at whatever temp it is when LED critical temp is reached. Hope that makes sense...

The Flex drivers can be programmed for temp, voltage, turn on safety, turn off safety (safety by way of how many presses to turn on and off), etc, etc. There are about 14 menus to program. A bit overwhelming at first, but once you get used to it not bad..

Is it programmable with a PC?
 

Packhorse

Flashlight Enthusiast
Joined
Nov 29, 2007
Messages
1,912
Location
New Zealand
Does anyone has any temp reading of the Luminus SST-90 LED or 3 x CREE XM-L U2?


Not sure what you are asking.

The temperature at which they run depends on how much current you pump into them, how well the heat sink drains the heat away, and ambient temperature.

The aim is to keep them as cool as possible. If you look at the data sheets you will find specs on how hot they get before they fail and also how much efficiency they will loose as they heat up.
 

Pizeer

Newly Enlightened
Joined
Sep 26, 2012
Messages
21
Location
Montreal
Well Packhorse, I'm looking for the operating temp when I have max voltage to the LED.
I was going through Luminus SST-90 data sheet but it's kind of confusing.

I guess I have to keep the LED under it's failure temp.
I also wanted to know temp data to know what kind of thermal expension I'm going to get on the heat sink.
I need to know that for the gap I need to leave between the heat sink and the housing since I want to use copper heat sink vs Aluminum housing.
Maybe you could guide me into understanding the data sheet.

thanks
 

lucca brassi

Enlightened
Joined
Feb 1, 2008
Messages
887
Location
Kocevje , Slovenia, Europe
I guess I have to keep the LED under it's failure temp.
Lower is better don't even think about failure temperature , you should keep your led (and that mean emiter core - not star PCB) on temperature cca 20°C over ambient temperature ! Same thing as by computer procesor !

I also wanted to know temp data to know what kind of thermal expension I'm going to get on the heat sink.
If you do correct there would be no thermal expension because you will keep all enclosure on almost ambient temperature.

I need to know that for the gap I need to leave between the heat sink and the housing since I want to use copper heat sink vs Aluminum housing.
Best way to fit heatsink in to the housing is tight fit . I leave max -0,006 mm on 50mm diameter of housing . Aluminium housing ''shrinks '' and cooper insert ''expands'' and that is key for good thermal path. I use ''arctic silver grease'' as ''lubricant '' for cooper insert.

You have here all phases http://www.scubaboard.com/gallery/showgallery.php?cat=5565

PS : on new lamp I have no aluminium heatsink with driver . I just put driver instead of switch and lens is inside of body instead of in the head (50mm)
 
Last edited:

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,706
Location
Connecticut, USA
Lower is better don't even think about failure temperature , you should keep your led (and that mean emiter core - not star PCB) on temperature cca 20°C over ambient temperature ! Same thing as by computer procesor !
...

If you mean to keep the die temperature at 20C above ambient, you'll have to run the emitter well below it's maximum ratings. The SST-90, XM-L, XP-E, and XP-G all have 20C or more just inside the emitter when run at full power. Of them all, the XP-G is the best, with a nominal temp rise that has to be rounded up to get to 20C (calculates at 19.5C).

Unfortunately, a more reasoned temperature analysis requires an understanding of thermodynamics that's far beyond explaining in a post. But to answer your question, the remainder of Lucca's response is pretty reasonable. You want to know how hot the copper slug will get. If you have lots of contact area between the copper and aluminum, and lots of contact between the aluminum and the water, the copper won't be much warmer than the water. In air it's a whole different story, but since this is a dive light, we don't have to go there.

It remains only to define what constitutes 'a lot of contact area'. At under 30 watts for the SST-90 (yes, the total power input is probably more like 35 W, but some is put out as light), 1-2 square inches is probably sufficient to keep temperature rise within reasonable limits, assuming tight fits and full contact with good thermal compound.

Wait a minute!!!! I just looked up the CTE of aluminum and copper, and aluminum is 50% higher than copper. That means that as things get hot, your gap will expand, not contract. You should worry more about loosing contact than anything else. On the other hand, if you lose contact, the copper will get hot and expand, while the aluminum won't. Sounds like a self-regulating system. I think Lucca's answer is about as good as you are going to get.

I'm not sure about the 0,006mm clearance. That's beyond most hobbyists' ability to machine or to measure. If you can achieve that, I'm impressed. I'd hope that 0.05 mm is close enough
 

Packhorse

Flashlight Enthusiast
Joined
Nov 29, 2007
Messages
1,912
Location
New Zealand
Yeah....0.05mm is close enough.

Machine your heat sink so it justs fits in the head. Thats the best you can get. I wouldnt even bother with copper. Stick with aluminium.

If you are machining the head from scratch then the heat sink and the housing can be one thing.
 

Pizeer

Newly Enlightened
Joined
Sep 26, 2012
Messages
21
Location
Montreal
Yeah....0.05mm is close enough.

Machine your heat sink so it justs fits in the head. Thats the best you can get. I wouldnt even bother with copper. Stick with aluminium.

If you are machining the head from scratch then the heat sink and the housing can be one thing.

Thanks for dimension.
Yeah, I was looking at the 0.005 mm and was a bit confused. Even a CNC machine could not get that tight.
0.05 = 0.002" and that's better.
 

350xfire

Flashlight Enthusiast
Joined
Aug 14, 2008
Messages
1,229
Location
Texas
Ive bought from all of them and if ignore shipping times, I'm happy with all of them.

Yeap... and if you can live with some stuff being dead when you get it. Not often but sometimes. Oh, this goes for DX and Kai...
 

Nwdiver

Newly Enlightened
Joined
Dec 4, 2012
Messages
6
What setup did you end up deciding on? Also did you come across any reflectors that would give a nice tight beam with 3 xml?


Thanks,
James
 
Top