Initial research

As I have not owned my own torch I knew very little of what a properties a good torch had.

My initial research started with browsing the web and visiting my local dive shop, Oceanic Tech. Some very nice threads and homepages were found, which showed what was on the market and what people have done by themselves.

It was pretty clear that the torch would be a LED-type. The mayor drawback for the HID and HMI are that they are far more expensive than a LED-solution. With this settled I looked at how much light the torch might need. A LED is also quite flexible in that you can use PWM to set the light at the light strength (~lumen output) that you want. Common strengths are up to 1 000 lumen. I felt that the starting point would be to source a appropriate LED and then calculate the needed amount of battery to accomplish a desired running life.

Some initial requirements:

  • Minimum 3 hours of burn time
  • A dry lid
  • Pins connecting lid with main body
  • Transportation mode for pins (can be seen here)
  • Housing made of POM (Delrin)
  • >1 000 lumen
  • Luminous emittance should be customisable
  • 8mm cable
  • Should not be able to overheat if turned on outside water
  • 18650 Li-Ion batteries
  • Withstand pressure of at least 6 bar / ~50 m
  • Between 1-2 kg out of water

The choosing of material for me was not something I would want to choose at such an early stage, but as all mayor canisters are made of this material, it seems as there has already been some research into this area. The material can also be found for a non-back breaking sum.

Some initial wants:

  • Serviceable, as few permanent seals as possible
  • Have cable glands
  • Some sort of mechanic to tell the user that the battery is running out
  • Easy to turn on and off with gloves
  • Hard to accidentally turn on and off
  • Go-Pro easy to fix to torch
  • Two customisable emittance settings
  • Some colour to make it easy to distinguish
  • Customisable emittance setting while diving
  • Standardised O-ring sizes
  • Easy to clean
  • Easy to service
  • Withstand pressure of at least 12 bar / ~110 m

Finding the LED

After some research I found a suitable LED, the Cree MK-R (datasheet), with a hefty maximum of 1665 lm and among the highest lm/W ratings (up to 200 lm/W, although this of course depends on the output) in the business.


At maximum power it draws 15.25 W (12.2 V, 1250 mA). This was available at several vendors (Elfa, Farnell, Mouser, Digikey) at around €20, but most had a surcharge for shipping from the US of €20 added to that. I ended up buying it at Lumitronix for €20 and €10 shipping. I choose the option with a small 12×12 mm pcb, I can always remove it if it turns out I don’t need it.

Finding the batteries

As I’ve chosen Li-Ion battery technology and the fact that these often come at 3.6-3.7 V I’ve concluded that the supplied voltage would be at approximately 14.4 V and as such conservatively increased the power drawn to 18 W. For three hour burn time that amount to 54 Wh or 15 Ah for 3.6 V Li-Ion batteries. Searching Deal Express for gave a initial choice of batteries mAH rating (these should of course be taken with a big grain of salt!). Calculating how much batteries would be needed gave between five and six batteries.

Some info on the 18650 form factor (source)

  • ⌀18.5±0.2 mm
  • 72.75±7.75 mm

Some calculations on how to stack the batteries (these were done in Autodesk Inventor 2013 for convenience, available for free for at least students in technological fields)

Number of Batteries (n)
Outer diameter (dy)

Battery pack diameter

The fact that the canister lamps available all are about ⌀60 mm or greater, about 200 mm long and the fact that the outer diameter only increase only 5 mm to 46 mm at n=4 made me choose to have eight stacked batteries in groups of four. This is primarily because the use of four batteries gives the desired voltage of 14.4 V. This would also give a theoretical minimum battery time of a bit over 4 hours (calculated somewhat conservatively with Sanyo batteries at 3.7 V and 2600 mAh), which is a bit more than the requirement.

Diameter of canister

With the outer diameter of the battery pack now chosen, the outer size of the canister could be chosen preliminarily. The size may change after some simulations have been made with the initial wall size. I plan on doing these with ANSYS. Well just have to see how this pans out.

Thoughts on moving forward

Lots of ideas are circulating in my head and I’m trying my best to put them into words. Some thoughts are that I want the top lid to be removable and that if this happens in the water, the batteries won’t flood. This means a seal must be in place for the inner lid and electrical sockets. If I were to bore out the whole inner core, there would be no given location for screws to apply the loading on the o-ring (I want to be able to service it, and as such I do not want a permanent seal here). I am however unclear on what is available in the way of drills at our shop and the length of drilling at roughly 200 mm could prove problematic. At this point I’m looking at latches, lenses and electronic components while I let this potential problem simmer.

Leave a Reply