4000 lumen Canister dive light

CodyS

Newly Enlightened
Joined
Jun 26, 2018
Messages
8
Evening all :)

I am posting here after some help for a light or a couple of lights I am building. I have a fabrication an joinery workshop, so fortunately I will be able to machine weld and generally fabricate everything myself. I am considering Bronze as a material to machine all the components from, because why not!

I am trying to build a Canister light to use while Scuba diving. The parameters I have outlined below:

3500-4000 Lumen
5000K
3.7V 26650 Batteries
2.5hr Runtime
Reed Switch

A few questions I have specifically, but I am really in need of general suggestions in regards to electronics.

What LED to use? A link to where to buy them preferably mounted to 1 MCPCB.

What driver to use?

How do I work out the number of batteries to use?

I want to use a reed switch to mean no further potential water ingress points, but I would also like to have it fully dimmable. I don't know if this is possible or where to go from here?

Any help would be very much appreciated.

Thanks in advance.
Kind regards
Cody
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,723
Location
Connecticut, USA
Hello, and welcome to the forum!

Boy, a bronze light would be cool. Unfortunately, I can think of two good reasons 'why not'. One is that it would be heavy. Even my aluminum lights are negatively bouyant. A bronze one would be heavy. Also, the thermal conductivity of bronze is only about 1/3 that of aluminum, meaning your LEDs would be significantly warmer, all else being equal.

Gee, I didn't know Cutter was carrying Taskled products. They are quite solid and it's hard to go wrong with them. That's quite a markup though! Direct from Taskled, they are $40 usd.

It may be a bit premature to choose the driver before you work out the batteries, but it can always be revisited later if necessary.

To calculate the batteries, work backward from the output you want, namely 4000 lm for 2.5 hours. You chose a triple XHP50, so you need about 1300 lm from each. From the datasheet, that's going to put you just under 2A for the 6V version, 1A for the 12V version. Either way it's around 12W each, 36W total. To run it for 2.5 hours, you need 2.5H x 36W = 90 W-H. If you figure your driver is 90% efficient, that means you need about 100 W-H from the batteries.

For a 26650, 5 A-H is pretty doable, and the average voltage is 3.7V, that gives you 5 * 3.7 = 17.1 W-H per cell. So you need 100/17 = 6 cells. Depending on your design philosophy, you may want to cushion that for when the cells start to age, your driver isn't quite what you thought, your dives run a little long, etc., you may want to cushion that and use 7, 8, or even more cells.

In general, at least up to 50V or so and above a few amps, it's best to keep the voltage as high as possible, and thus the current as low as possible, and use a buck-type regulator if possible, as they are inherently the most efficient. Your cell voltage will vary during discharge, from 4.1-4.2V right off charge, to 2.7-3.0V depending on how you define end of charge. With 8 cells in series, you'd want your driver to handle at least 24-32.8V, with 21.4-33.6 being better. That's quite reasonable, and in the range of several Taskled drivers.

To use a buck driver, the LED voltage needs to be below the minimum battery voltage. Cutter isn't real clear on how the board is wired, but I'm guessing since it says XML/XHP50 6V on it, the XHP50s are wired in the 6V configuration. If you got the series-wired board, you'd have 3x6V = 18V. This is a few volts below the max battery voltage, so IDEAL, if you can find the right driver. The Taskled Hyperbuck would work great, but I see it's not on his order page. You could contact him for status.

But neither the Hyperbuck nor the Hyperboost is dimmable with a switch (needs PWM or potentiometer). For that, at least from Taskled, you need to be looking at the ---flex drivers.

This is getting a bit long, so I'll discuss those drivers in another post.
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,723
Location
Connecticut, USA
So the Taskled h6Flex, b3Flex, and HBFlex would all give you several different brightness levels, controllable from the Hallsw board, they all have max input voltage of 25V. This means a max of 6 cells in series (6x4.1V = 24.6V). Then at the end of charge, you have 16.2-18V, which is not enough to drive the 18V LED. This wouldn't be a terrible problem, as the LED would begin to dim as you reach full discharge (3.0V/cell is pretty close; there's not much left before you are at 2.7V). Another option would be to run the h6Flex and the parallel board, which would want 6V, 6A. This lines up well with the max output of 6.6A for the h6Flex. In this case you'd only need 3 cells in series (min 8.1-9.0V), but you could use up to 6. This would give you the option of using 6, 8, 9, 10 or 12 cells.

Turning to the HBFlex, the only boost driver of the lot, you'd definitely want the 18V board, and absolutely no more than 4 cells in series, probably 3 is better. Why is this? The basic boost design, which includes the HBFlex, cannot stop the battery from driving the LED directly. That means that if the battery is at 16V, and you have an 18V LED board, the board will actually draw some current from the battery and will light up. It won't be really bright, but it won't be off either. The driver cannot stop this. Better 3 cells and max 12V or a bit more. Even at 12V, an 18V board will probably draw a very little current, and light very dimly. If you connect the battery at the beginning of the day and disconnect at the end of the day, this is probably not significant. But over several days or weeks, this can discharge the battery significantly.

So for what you want, from the Taskled line, I think I would choose the h6Flex, 6V (parallel) LED board, and run 6 or 9 cells in a 3S2P or 3S3P configuration.
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,723
Location
Connecticut, USA
One more thing: magnets!

I found that neodymium magnets rust in salt water. This is because they are actually made of neodymium, iron, and boron. When they rust they expand and they lose strength. It doesn't matter that they are nickel plated. That plating is not perfect, and it WILL fail.

Unless you are going to encapsulate them in a completely waterproof coating, don't use them!

What to use instead? Samarium Cobalt. My experience is that they are completely corrosion resistant in sea water. They aren't quite as strong and are a bit more expensive, but quite worth it for a dive light that you expect to last a few years. As I type this, the DIY dive light sitting on my desk has a rebuilt magnet ring because the original one was destroyed by rusting neodymium magnets. Once i found they were rusting, I could not remove them because of the expansion. I couldn't dig them out, I couldn't press them out, I couldn't even drill them out because they are ceramic, and quite hard. I suppose a carbide drill would have worked, but it was easier/cheaper to just make another ring.

Even if using samarium cobalt magnets, it might be a good idea to protect them if you do much wreck diving. Magnets will attract tiny bits of iron from the wrecks, and this can eventually make a mess too.
 

CodyS

Newly Enlightened
Joined
Jun 26, 2018
Messages
8
Thank you for your response.

I never realised that bronze had so much lower thermal conductivity in comparison to Aluminium, that answers that one... perhaps gold or silver? Ha!

Are you able to elaborate on why it is important to keep voltage high?

Seems to me your logic is sound in regards to the H6Flex, so I'll give that a go.

I would probably plan to run the light at or closer to it's capacity so might even go for 3S4P just to produce plenty of capacity.

This makes me wonder having to use so many cells, is there a bigger better option for batteries?

Would running the leds at 3A be ok for the driver?

Do you know a good source for waterproof connections/quality cable? Currently I am considering just epoxy or silicone etc and hardwiring it.

In regards to the magnets, I will seal them in by welding an aluminium case around them. Of course being conscious to not overheat them.
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,723
Location
Connecticut, USA
I did a bit more research, and found some interesting things. Here are some thermal conductivities of various materials. There are a few real surprises. This data was collected from several sources, and is not intended to be either definitive or beyond reproach. Values are W/m-K.

Aluminum
pure 237
2024-T4 121
6061-T6 155
7075-T6 122

Silver
pure 428
sterling 361

Gold 318

Copper
pure 401
commercial 350-360

Nickel 90

Brass around 100

Bronze 26-76

Titanium
pure 15.5-22.5
alloy 5.6

Steel
carbon 36-54
stainless 16-24

plastics 0.04 - 1.0

Graphene around 5000

Diamond 1000 - 40,000
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,723
Location
Connecticut, USA
Keeping voltage high: two reasons
1. Keeps current low. This is important because resistive losses are often significant, and this keeps them low.
2. Allows use of buck regulators, which are inherently the most efficient.

Batteries:
It depends on what's important to you. I like the 26650 because it packs the most energy in a single cell that's reasonable for a dive light. If you are looking for the lowest cost per W-H, it's probably in 18650s. I hear Tesla is switching to a 2170 (which we would call a 21700). If you can get your hands on those, LMK:). There are bigger cells that make huge packs easier to build, though the cells are not cheap. Batteryspace.com sells quite a variety. For the pack you want, I think you'll come to the conclusion that 18650 or 26650 is probably the best size.

The h6Flex won't run three LEDs at 3A each. It maxes out at 6.6A, which is 2.2A per LED. You simply can't get more than that out of it. Hang twenty 3A LEDs on it, it will still only output 6.6A.

For dive lighting parts, you should check out the Dive Lighting subforum in the Special Applications Lighting forum. There's a sticky there that lists sources for all kinds of parts for dive lights. There's also a ton of interesting reading if you are getting into DIY (or even commercial) dive lighting.
 

archimedes

Moderator,
Staff member
Joined
Nov 12, 2010
Messages
15,792
Location
CONUS, top left
Interesting build thread @CodyS ... but this is drifting off away from "LED" topic(s)

Would you prefer that I move this to "Dive Lighting" , or rather "Homemade / Modified Flashlights" ?
 
Last edited:

CodyS

Newly Enlightened
Joined
Jun 26, 2018
Messages
8
External Charging of Canister, Glands and cable

Hi All,

I am currently working on a canister light to use as a primary.

I'm a bit stuck on what glands and cable to buy. Any links to specific items would be very helpful.

I am wanting to copy an external charging setup a few brands use. I can't really find any examples on how to go about it? I've linked a pic as an example below.

http://www.divephotoguide.com/images/imguploader/755330050.jpg

Any tips or links appreciated.

Cheers
Cody
 

CodyS

Newly Enlightened
Joined
Jun 26, 2018
Messages
8
Sounds good. I have posted another thread to get better attention, but this one does belong in the dive lights forum. If you could delete my last post and move the thread that would be great. I'll try keep this thread up to date as a log then.
 

CodyS

Newly Enlightened
Joined
Jun 26, 2018
Messages
8
Well, thanks to a bit of help from George @Taskled, I've decided to go for a reed switch/relay circuit to isolate the pins when not charging.

Moving onto my next question. How do I charge 3S3P set of 26650 batteries? What charger do I need, is there an off the shelf 'solder and play' option that I could go for?
 

cpassuel

Newly Enlightened
Joined
Sep 6, 2014
Messages
8
You can use for example skyrc imax B6AC or B6 type of charger (preferably genuine one), it supports balancing if you battery pack has wires for balancing.
 
Top