Planning a DIY Canister Light build, need some help

[darkavenger]

Hello all!

This all started out as a simple mod to an old light, however I'm being sucked in.

So I'm ready to jump into a canister light build. Almost. I've made progress toward a plan anyway. I've selected some parts to use:

Modded Maglite body cut down in size
3 x XM-L T5 neutral white (4500k) LEDs
3 LED reflector
3 Mode Boost/Buck 3 LED XML driver 30W 3A (3 LEDs in serial)
Heatsink (copper over aluminum?)
Lens - Boro Glass 5mm thick
Cable/Connectors
Switch (piezo)
Battery Housing

What I still need to figure out...

Battery Pack
Recharge system

I've been toying with the idea of filling the LED housing/Maglite body with mineral oil to reduce the effects of pressure and help the light's thermal properties. If everything is installed properly, this housing could be sealed fairly permanently with the oil inside. Any thoughts?

 

[jspeybro]

HI,

if you're looking for places to get certain parts, there are 2 sticky threads in this forum that have plenty of information on where to get parts.
tlslights.com is a good start for parts involving a maglite mod.

Johan

 

[Packhorse]

I would forget the oil. Its messy and not needed. I have build lots of mag based can lights and never had any fail due to pressure. Several have been below 50 meters and a couple to 80 meters.

30 watts of light will generate quite a bit of heat. The heat sink will need to make good contact with the LED's and the body. A disc shape should be fine as long as it makes good contact with the flat part in side the head.

a 3000K temp will give you a really warm beam. But it will also be a lot less intense that if you used higher temp LED's. IMHO you are better off using 5500-6500K LED's If you are doing video or photography then warm LED's may be more preferable but then thats debatable.

Using 3 XML's will give you quite a wide beam. no where wide enough for video mind you.

 

[darkavenger]

Alright, it's a work in progress. 3000k does look a bit yellow now... somewhere in the 4000-5000k range looks nice though. I'd like a warmer light. I have a 400 lumen SL4 that does a nice job but is a bit too bright in the hotspot for my taste. I'm looking more for warm and floody, but still focused to a point. Maybe something like a 40-60 degree beam? In the ballpark should be good enough for me.

Are there any 'local' stores that might carry cables/connectors?

For the battery pack, I'm thinking a larger PVC tube would work well, any thoughts? I'm not sure quite how to go about this, can the pack be sealed semi-permanently? Or is there another option?

Some simple lens questions... I'd like an acrylic lens, any thickness recommendations? What diameter should I use for a maglite? Any special considerations? I'd like to ditch the o-ring and glue the lens to the case, which adhesive should I use?

Also, I'm going to cut the maglite down to a manageable size, I've seen JB weld recommended. Is this what's typically used? Will it work to attach the piezo switch as well?

 

[darkavenger]

Ok, I think I've figured out most of this. The last thing left is the battery pack. I'm using a buck/boost circuit capable of using 2-4 Lithium cells to drive 3 xml LEDs in series. I figure the best voltage to run at is going to be 3 Lithium cells in series, correct? This would be the more efficient route vs 2 cell step up or 4 cell step down.

Is there a tutorial of sorts for a lithium pack? I've seen the recommended unprotected cells with custom monitor circuit. Is there a recommendation to how many cells to use or overall capacity? Casing type? How to recharge?

 

[DIWdiver]

The regulator's efficiency may be better in boost or buck mode, depends on the design. Does the data sheet give you any efficiency data?

 

[darkavenger]

It's the KD Boost/Buck 3-18V 3 Modes 24mm 30W Circuit Board. I can't figure out the exact specs.

I have another question about heat management. Both my LEDs and Driver need good thermal management. I was thinking about doing something a bit more elaborate than just a heat sink. Say I machine a small heat sink about 1" thick and drill a few holes through the sides. The holes would allow water to cool the heat sink. Here is a quick drawing of what I mean:

Copper to disperse the heat more evenly to the heat sink. The red is where holes could be drilled.

 

[DIWdiver]

The holes are overkill. All you have to do is get the heat to the surface of the light, and the water will take it away faster than you can generate it.

 

[darkavenger]

I'm going with a tenergy battery pack + charger.

Parts are on order!

One more lens question, it's still worrying me. I'm not going with the boro lens yet, I'm going to order a 3 LED optical lens and a few glass lenses to stack together. Will this work?

 

[DIWdiver]

If you stack lenses, it's best to glue them together. If you get any water between the lenses, it will create irritating patterns, and over time perhaps mineral deposits or staining like mildew. Only if you seal the outermost lens will you be able to prevent this, and that's probably the hardest one to seal.


Another problem is that if there's any debris or non-uniformity in the lenses, they will create stress points, which make them dramatically weaker.

And lastly, two lenses stacked are twice as strong as one, but a lens twice as thick is 4 times as strong.

 

[DIWdiver]

You're not going to get good specs on a KD or DX regulator, unless they're done by a CPF member, which is pretty unlikely. You'd probably never buy one if you did. Those super-cheap ones tend to have poor regulation, poor efficiency, and worse than poor build quality. Experienced folks often buy several to ensure they get one that works.

But they do often work and they're REALLY cheap, so they're popular, even among those who know what they are getting.

I'd say pretty likely the only way to determine the most efficient battery configuration is to test it. This can be done with just the batteries, the LEDs, and a very cheap Digital Multi Meter, though a low-value resistor (0.1 ohm) is also helpful. If you need advice, ask and ye shall receive.

 

[darkavenger]

Yes, I'm sure the circuit won't be ideal, however I'm on a budget and being my first build I don't want invest too much... yet anyway. The circuit seems one of the easiest parts to replaced and very well might be. I read somewhere that some of those drivers are about 70% efficient vs 90% with a quality driver. I guess I'll just have to test it to find out though.

I've changed my canister design around to be latch free. I think I might go with the boro lens after all.

I also need to order my switch. Would a standard piezo work for this application? If so, I don't need any additional circuitry? I have a couple piezo elements laying around.

 

[Packhorse]

Piezos are momentary switches. Some prolonged momentary.

If you have a driver that will accept a momentary switch then fine. But most require the power being supplied to them to be switched. So you need a latching/Toggle switch rated at the current you expect the circuit to consume.
In this case you will require a toggle circuit to be activated by the piezo.

 

[darkavenger]

Ok, prolonged it is for ease of use and ability to upgrade circuitry.

Ok, one more idea. Can a cable gland be installed into the side of a pvc pipe easily? Does this require reinforcing? Does the gland need a flat surface to mount onto?

 

[darkavenger]

Part 1: Battery Pack

Cut and resoldered

Heat shrink

Final touches

and not enough clearance. Heat shrink off...

New heat shrink over the canister(cosmetic)

 

[350xfire]

I'm going with a tenergy battery pack + charger.

Parts are on order!

One more lens question, it's still worrying me. I'm not going with the boro lens yet, I'm going to order a 3 LED optical lens and a few glass lenses to stack together. Will this work?

Stacking lenses is a bad idea!

 

[Klem]

Switch
If you have piezo buzzers lying around and want the satisfaction of turning them into switches then read this thread by 'DIW';
http://www.candlepowerforums.com/vb/showthread.php?296774-DIY-piezo-switch&highlight=piezo+buzzer

For the effort of doing this my vote is go with a commercial piezo. If you think they are expensive then... 'Buy once, cry once'... compared to other switches piezo's are low profile, easy to use underwater, and bomb-proof.

Lens
5mm Perspex works for me in a Maglite (5*52mm). Buy it from your local glazier and hole-saw it out. It's as cheap as glass. I step the edge to fill forward of the bezel, so as to allow enough thread on the bezel to take on the head. 2-part Epoxy putty (Selleys Aqua-Knead-it) works to hold it and seal it to the bezel.

I'll add to the discussion on multiple lenses...don't do it. I agree with the others on this. Appreciate that as pressure builds the outside lens cracks first, then the next gets the same pressure and so on. The failure will be a quick cascade of individual lenses. If there's no glue binding them together at the macroscopic level then they may as well be a mile apart. There will be some support of the outer lenses by the inner lenses but not enough to offset typical dive pressures. You can glue them together with clear glue however the lumens loss and the faffing around to do this is the same as the effort to source a proper single thickness lens.

 

[darkavenger]

Thanks for the input! I've already ordered a boro lens and prolonged piezo.

Any thoughts on the pressure plug design? It removes the latches for the equation. Do I need to take extra precautions against flooding? Any thoughts about adding the glands to the side of the pipe rather than top?

 

[Klem]

Thanks for the input! I've already ordered a boro lens and prolonged piezo.

Any thoughts on the pressure plug design? It removes the latches for the equation. Do I need to take extra precautions against flooding? Any thoughts about adding the glands to the side of the pipe rather than top?

By 'pressure plug' and no 'latches' I take it you mean water pressure is going to keep two parts of the torch together, like this;

It works, however you need to be careful that it doesn't come apart by accident given it is only being held together by the friction of the O ring/s. When you push the plug in you compress air inside the canister which will then be pushing the join open. The more internal surface area you displace the more pressure you build inside. This torch never leaked but it was only ever one decent knock away from flooding. It has been upgraded by one which holds the plug/seal firmly by a floating threaded bezel.

Here's the same torch with a mechanism for holding the plug firmly in placel;

B

Being able to mechanically seal the torch up is peace of mind. The free-floating white locking bezel is far more clean-lined and safe than a couple of latches which you have to drill into the sides of a pretty severe 50mm diam curve. The Maglite head is now permanently glued/sealed to the yellow body and the access port and O rings down the other end where the piezo is.

Heat/Wattage
30watts will be no problems in a Maglite. No need to have cooling tubes as long as your thermal path the the outside is unbroken and the XM-L's sit on a decent base. FYI the yellow torch is 42W (4*XM-L) and has no problems staying cool enough underwater. Just don't expect to use it above water on 'High' power for any more than a few seconds.

Gland coming in from the side, curved surface.
No problems if this is what you want to do. Just build up a flat surface with epoxy putty or fill the void with plenty of sealant. That said I am wondering why you want to do this. I can see a big risk of failure when you bang or compress that gland against something underwater and the sealant cracks from the lack of flat surface to take the pressure.

The red piezo on the silver Maglite above is a similar idea with a flat surface glued sideways into a curved surface. JB Weld, Devcon or similar two-part epoxy.

One more thought...If you are going to use a cable/canister/wand design try to make your torch as clean as possible. Work with the shape, not against it. A cable gland coming in from the side is asking for trouble, too many edges to catch onto and lever the torch into breaking. A gland coming in from the end, screwed into a flat surface is much better.

 

[darkavenger]

Ok, good ideas! Thank you so much. I was thinking about putting the cable glands on the bottom(sealed), however they will be very hard to access there. I do have more than enough wire to run the length of the canister to the access point for a charging connection. Also, I sealing the access point with a pressure test stopper:

The plug is inserted then tightened down. It seems to make a very nice seal. I'm using 1 1/2" ABS piping.