DIY Canister Light

[victorzamora]

Hey guys,

I'm in the initial design phase of a canister light. I have access to a couple of well-equipped shops, so tools *shouldn't* be an issue. I also have my own battery charger capable of Lithium, Lead, and Nickel batteries from my R/C hobby. I also have access to a laser cutter for anything 2D smaller than 18"x24"....and most of the cutting it does is 1/4" acrylic. So I know that I can get awesome, clean cuts on 1/4" plexi. I have read more than a few dive light design guides, and have a pretty good grasp of basic electronics. I have also been looking into stacking circles inside of circles and am pretty good with my hands.

I think I have decided on LiFePO4 cells. They are a relatively new lithium technology much more durable than LiIon and LiPo, have 3.3V per cell nominal charge rates, and have MUCH flatter discharge curves. 1.2V Nickel cells charge to 1.65V and quickly discharge to around 1V. LiFePO4 cells charge to 3.6V max, quickly drop to 3.3V, and maintain. I think this will give the most consistent lighting and will be safer than most other cell chemistries. The only two "genuine" A123 cells (the original LiFePO4 brand) are 26650 and 18650, 2300mAh and 1100mAh respectively. The batteries will be in a 4S configuration for 13.2V nominal. In a 10.5" long bundle, no more than 3" around, I can get 146Wh and 178Wh respectively. If I go a little bigger, I CAN fit more. It seems like the smaller cells are a better fit but will be more expensive. 12x26650 cells or 48x18650 cells. Due to cost, I'm still looking into Nickel and other Lithium chemistries....but LiFe would be my first choice.

My only concern is WHAT to do about the lighthead. Any good tips? I have a MagLite I'm willing to sacrifice, but would rather not have to if there is a better/cheaper option. HID vs LED....I can't figure it out. Any hints? I also can't figure out how to properly watertight anything.

 

[350xfire]

Maglite is by far the most hacked and cheapest head you can probably build.

LED will be simpler to design and cheaper than HID.

What will this light be for? 48 18650 batteries will run a 1000 lumen LED for a lifetime... 8 18650 will do it for about 5 hours. So, really, what is the purpose of the light?

 

[victorzamora]

Wow, my calculations might be off then....it IS a canister light for diving. The 18650 batteries aren't 18,650mAh....that's a model number. They're 1100mAh a piece, at 3.3V a piece, so that's 3.63Wh a piece. For one reason or another, I assumed I need a 12V power source....so 4 of them would've been 13.2V, 1100mAh, 14.52Wh a piece. A 50W light could blow through that in 20ish minutes...right? I mean, is my math wrong?

Also, looking at other people's DIY canister-lights it seems like the Cree XR-E Q5 Emitter is very popular and very bright. Would I need an LED driver with that? Also, I've been looking into Maglite upgrade paths and can't find one bright enough to make me want to go through the hassle of waterproofing (without a ridiculous cost, that is). Would it be easier to just use a plug-and-play upgrade? I'm also almost completely decided upon LED over HID...I'm glad you recommended LED.

Edit:
I have now realized that 50W might be quite a bit of overkill. To possibly fix my math, that Cree XR-E Q5 is a 3.7W LED. That would be nearly 4 hours of runtime on 4S1P of 1100mAh/3.3V batteries (13.2V, 1100mAh). Also, I'm looking at TLSLights site and finding exactly what I've been looking for for a lot of the parts.

 

[350xfire]

Wow, my calculations might be off then....it IS a canister light for diving. The 18650 batteries aren't 18,650mAh....that's a model number. They're 1100mAh a piece, at 3.3V a piece, so that's 3.63Wh a piece. For one reason or another, I assumed I need a 12V power source....so 4 of them would've been 13.2V, 1100mAh, 14.52Wh a piece. A 50W light could blow through that in 20ish minutes...right? I mean, is my math wrong?

Also, looking at other people's DIY canister-lights it seems like the Cree XR-E Q5 Emitter is very popular and very bright. Would I need an LED driver with that? Also, I've been looking into Maglite upgrade paths and can't find one bright enough to make me want to go through the hassle of waterproofing (without a ridiculous cost, that is). Would it be easier to just use a plug-and-play upgrade? I'm also almost completely decided upon LED over HID...I'm glad you recommended LED.

Edit:
I have now realized that 50W might be quite a bit of overkill. To possibly fix my math, that Cree XR-E Q5 is a 3.7W LED. That would be nearly 4 hours of runtime on 4S1P of 1100mAh/3.3V batteries (13.2V, 1100mAh). Also, I'm looking at TLSLights site and finding exactly what I've been looking for for a lot of the parts.

lol... something seemed odd! OK, get a Cree XML (1000 lumens)- better than the SSC P7 and Cree MCE. Cree XR-E only about 300 lumens. Run a 7.4 volt pack (2S4P). Use the better 18650 rated around 2500 mah. You will get a run time of about 4-5 hours on high.

Yes, 50W LED is way big. You'll need a big expensive pack. That is the issue with running big lights. Batteries are becoming the weak link both in price and size.

Yes, TLSlights- your source for quality DIY parts...

Oh, and BTW, 1100mah 18650 cells don't sound all that great!

 

[victorzamora]

So, I've been looking around and was wondering.....does THIS include the emitter, driver, reflector, et al? What else would I need to get this light working if I didn't care about the lighthead?

I'm thinking for the MagLite I'll order a lens&o-rings kit, a couple of glands, couple latches, and a bunch of wiring. I'll get some CAD models drawn up once I get the latches. Once I nail down an emitter, I'll decide for sure on a battery configuration and therefore will be able to get a completed prelim very quickly. Once I get this laser cutter access issue squared away, I'll get the acrylic shipped so I can cut it out. This SHOULD be a relatively simple assembly, especially if the emitter/reflector and glass lens drop in fairly quickly into the Maglite!

 

[350xfire]

So, I've been looking around and was wondering.....does THIS include the emitter, driver, reflector, et al? What else would I need to get this light working if I didn't care about the lighthead?

I'm thinking for the MagLite I'll order a lens&o-rings kit, a couple of glands, couple latches, and a bunch of wiring. I'll get some CAD models drawn up once I get the latches. Once I nail down an emitter, I'll decide for sure on a battery configuration and therefore will be able to get a completed prelim very quickly. Once I get this laser cutter access issue squared away, I'll get the acrylic shipped so I can cut it out. This SHOULD be a relatively simple assembly, especially if the emitter/reflector and glass lens drop in fairly quickly into the Maglite!

Yes, that is a complete module. All you need is a battery and a switch and it's ready. Stick with 2 LiIon strings (2S4P has worked well for me). That assembly will need the outer lip removed to fit in a Maglite. Also, to get a 5mm lend in a Maglite you will need to remove about 2mm or threads from the head to get the bezel to sit as the original. Otherwise you will have a not-so-nice gap on the head.

 

[victorzamora]

Is it just a gap that doesn't hurt the waterproof performance of the system? I haven't disassembled my Maglite in a LONG time....so I can't remember exactly how it goes together. If it's a gap between the bezel and the head (as in it looks like it's not screwed on all the way), I might just put an O-ring on there. Again, it's been years since I've disassembled the Maglite.

 

[350xfire]

Is it just a gap that doesn't hurt the waterproof performance of the system? I haven't disassembled my Maglite in a LONG time....so I can't remember exactly how it goes together. If it's a gap between the bezel and the head (as in it looks like it's not screwed on all the way), I might just put an O-ring on there. Again, it's been years since I've disassembled the Maglite.

Yeah, I am pretty sure you can get it to seal. You'll just have that unsightly gap on theren

 

[victorzamora]

Okay...so I've gone back and forth a couple of times, but I'm trying to decide between this and this. There aren't as many specs on the second one, but it seems identical to the first one, except for it being High mode only. The benefit of the first one is different modes for different lighting needs....including nearly doubling battery life and still getting 400 lumens. Also, I KNOW the specs on this one, so there's no gamble. On the second one, I feel like there's less crap to fail.

I had settled on the XML, but these are brighter and have 1/3 the current draw. Am I missing something? Maybe smaller LED's simply do better??

Also, I have a 3D Maglite to start this project with. I'll be getting that this weekend to start looking at for ideas.

 

[350xfire]

Okay...so I've gone back and forth a couple of times, but I'm trying to decide between this and this. There aren't as many specs on the second one, but it seems identical to the first one, except for it being High mode only. The benefit of the first one is different modes for different lighting needs....including nearly doubling battery life and still getting 400 lumens. Also, I KNOW the specs on this one, so there's no gamble. On the second one, I feel like there's less crap to fail.

I had settled on the XML, but these are brighter and have 1/3 the current draw. Am I missing something? Maybe smaller LED's simply do better??

Also, I have a 3D Maglite to start this project with. I'll be getting that this weekend to start looking at for ideas.

Here is my advice:
Stop wasting your time and take my advice... lol!
The XML and the 5 LED module are close to the same brightness. The 5-LED module is going to take mode modding to get into the head. I have used both so I can tell you what it takes... The 5-LED reflector is deeper- requiring machining of the maglite internally to fit- and also floodier in beam pattern. Not bad but a bit more flood. I assume this is your first build. Jeep it simple go with XML module.
Also, my opinion on multi-mode is that it is, mostly a waste. I have never told myself "I wish I could have less light"... But that's just me... May be if I was running a 6,000 lumen monster against my buddy's 1000 lumen light.

And no, they are not 1/3 current draw. They draw about the same 1.5 to 1.7 amps at 7.4 volts.