7 500 Lumens under water
Earlier this year I started scuba diving, only in pool first to pass the first level. This is now done and I’m able to dive up to 20m ! I already planned some practical diving in natural environment next month.
This being said, and as a photography and video enthousiast, I looked around for some powerfull lighting. When dealing with image capture, and particularly underwater, the more light there is, the better will be the picture. As I did not find a truly powerful lamp in a reasonable price range, I of course decided to create my own design, inspired on various things I found on the internet.
In this first article, we will take a look at the optical needs and the mechanical design of the lamp.
To shoot underwater, it is important to have a good lighting uniformity, and the widest beam possible, this is called a “flood” light.
Most modern white LEDs achieve, due to their construction, a lambertian emission pattern without special optics. It means that the light emission follows a cosine rule with the angle from optical axis, this gives 50% intensity at 60°, resulting in a very wide and homegeneous beam, which is exactly what I was looking for.
Particular attention needs however to be taken on the edges of the beam. Often the color on the edges is shifting a lot to a more “yellow” white due once again to the construction of the LED and the quantity of phosphor used.
The first thing was to decide which LED to use knowing that the target flux was at least 5000 lumens. Having a certain experience in LED product design, I decided to take a look at Cree portfolio, particularly their CoB (Chip on Board) LEDs.
They propose a wide range of LEDs, mainly based on their size. My choice went to this particular reference : CMA1840.
I will, in a near future, write an article about how to read an LED datasheet to estimate how much flux we can expect but for now, let’s keep in mind that I target to get at least 7500 Lumens at 2A out of this little beast. Indeed the emission surface is only 14mm in diameter !
My first limitation was on the tools I was about to use. In order to build a fixture for the LED that would allow to be water tight, to sink the heat away from the LED, and to limit the optical losses, I designed this aluminium part:
For a better understanding, this is how it will be mounted:
The O-rings will ensure that the LED is sealed from water ingress, as well as the back side (interface with the housing) in which there will be holes for head fixation and supply wires.
To build such a part in aluminium, I planned to use my CNC converted Proxxon MF-70 mill which has some limitation in the maximum size it’s able to mill : 134mm in x and 46mm in y.
Therefore, taking into account the diameter of the routing bit and some margin I decided to limit myself to 40mm max diameter.
With the size of the O-ring and the screw holes the 1840 LED was the maximum size I could put.
As stated previously, this part have different roles, apart from the sealing and mechanical attachment, its main feature is to take away the heat from the LED to dissipate it into the environment.
To have an order of magnitude of the power we need to sink, we can do an easy calculation, LED will be powered at 2A and the forward voltage will be 36V ==> 72W.
This is the electrical power, the addition of Optical and Thermal power whose ratio is depending mainly on the junction temp and the current density. But this is a good starting point as it’s kind of worst case, it’s always higher than the actual thermal power.
Luckily, the environment is… water, which will provide a very good thermal bond to any metallic part and dissipate the heat very efficiently, therefore an accurate thermal design is not mandatory.
In the aluminium part there is a pocket for the LED to be mounted on, thermal bond will be ensured by adding some thermal paste between the LED and the head. This will also ensure a good cohesion as I don’t plan to screw the LED in the heat sink (Which would be of course better for thermal bond because it will reduce paste thickness, and increase the number of metallic contact points) but I don’t think it will be an issue here.
The other side of the head will be bonded to the die cast aluminium enclosure using screws. This time with silicon thermal grease to achieve good heat exchange and ensure better sealing. At the end of the day the whole lamp body will act as heat sink.
Main limitation to this design is that power will be too high to be used outside of water. But I plan to add thermal sensors so it doesn’t exceed maximum ratings if operated in the air.
Enclosure size was chosen taking into account the battery design, that I will detail in part 2.
For now :
You can notice that the screws will be in the optical path. I did not perform optical simulation but I don’t think this will be noticeable because the emission area is quite wide compared to screw size, therefore the shadows will be very scattered.
And here is the result of the milled part, it’s a bit different because I changed some things during the manufacturing based on the available material I had. The Idea is however still the same.
In those views some sanding and polishing were still to be done.