I am looking to build two canister lights for my wife and I. I have already built two such units using overdriven halogens (10 degree 35 watt MR16 at 17 volts) in mag light heads. But time marches on and I want to go the LED route. I am looking for units with a tight bright beam and low weight. My current canisters use 14 NiMh C cells. I have been lurking quite a while trying to sort things out but the more I learn it seems the more iI need to know. It is like drinking form a firehose. I have a lathe but neither it nor I are up to complex machining nor do I have a lot of stock available so I would prefer a maglight head solution. I would appreciate info on things like aspheric v reflector, single v multiple die and specific LEDs to use, regulator (I am assuming a buck?), suitable battery types, and anything else I may have left out. Just a few pointers to get me started in the right direction
help designing a canister light
- Thread starter bowzer
- Start date
Help Support Candle Power Flashlight Forum
DIWdiver
Flashlight Enthusiast
Wow, that's a lot of questions in a brief post!
For tight beam, especially if you want little spill, I doubt you're going to get it in a mag head with multiple emitters. And if you want light equal to or exceeding that overdriven 35W halogen (I estimate about 1000 lm), you'd want a triple XPG board or an SST-90. There are some triple XML board/optics combos available, but you might not find one with a beam you like. Packhorse has been saying he likes the 44mm aspheric from DX, and I think he's posted beamshots of it with the SST-90. A single XML is around 1000 lm, but by the time you get through your optics, you'd probably have less light than your halogen, albeit at vastly less power. XPG's are around 500 lm each, and SST-90 is probably 2000, though they claim 2200 at full power.
Basically, a tight beam means big optics. And a big LED means big optics. The physics just don't let you cheat. For a reflector, collecting lots of light into your central beam (vs letting it spill out the sides without hitting the reflector) means the reflector needs to be deep. A reflector with depth and width approximately equal lets out a good bit of spill. Your halogen is probably in this range.
For an aspheric, any light that doesn't hit the back of the lens is either absorbed by the body or reflected out as spill, so capturing lots of light means keeping it close to the LED. That means you need a short 'focal length'. I know that's not exactly the right term, but it's close enough. Short-focus aspherics are thick, bug-eye looking things. Thin aspherics have long focal length. They will give a fine beam, but lose a lot of light.
Your battery pack is probably fine, depending on the driver you end up with. Or you could reduce the weight and/or increase the runtime by switching to lithium. The 18650 is by far the most popular. I personally like the 26650. That's C cell diameter, but longer. In case you don't know, the first two digits are the diameter in millimeters, and the next two are the length, and the zero on the end means it's cylindrical. Sizes are approximate. Cells vary slightly in diameter and considerably in length. Choose a cell and get reliable numbers for that cell before counting on an exact size (cells of the same model and brand will be pretty consistent). I won't try to discuss chemistries, that's a huge subject that's been debated ad nauseum elsewhere.
You will certainly want a buck driver, since your battery voltage is higher than the LED voltage. I suppose you are used to single mode (on/off), and if you're happy with that, Der Wichtel makes a driver suitable for the SST-90. I make one that can have high/low/off modes, and drive an SST-90, but it's a linear driver, which means you have to keep the battery voltage close to the LED voltage or the efficiency sucks. It turns out that 4 NiMH cells and one LED works out great. Eight and two or nine and three also works. And of course you can parallel as many cells or LEDs as you want, so you could run a triple XML board on a 9S2P pack. Unfortunately there isn't a combination of LiIon cells and LEDs that works very well. If you want more modes, your best bet is going to be a triple board or single XML. I've not seen or heard of a multi-mode buck driver that can run 9A to the SST-90, though Taskled has some that are 6.6A IIRC. Taskled makes great drivers.
If you go LiIon, you'll want to think seriously about a 'hobby' charger. These are designed to charge a variety of multi-cell packs quickly and to treat each cell individually, as is necessary for proper care of LiIon cells. Alternatively, a good balancing protection board can be built into the pack, but it gives you less control and less information about the health of individual cells.
For tight beam, especially if you want little spill, I doubt you're going to get it in a mag head with multiple emitters. And if you want light equal to or exceeding that overdriven 35W halogen (I estimate about 1000 lm), you'd want a triple XPG board or an SST-90. There are some triple XML board/optics combos available, but you might not find one with a beam you like. Packhorse has been saying he likes the 44mm aspheric from DX, and I think he's posted beamshots of it with the SST-90. A single XML is around 1000 lm, but by the time you get through your optics, you'd probably have less light than your halogen, albeit at vastly less power. XPG's are around 500 lm each, and SST-90 is probably 2000, though they claim 2200 at full power.
Basically, a tight beam means big optics. And a big LED means big optics. The physics just don't let you cheat. For a reflector, collecting lots of light into your central beam (vs letting it spill out the sides without hitting the reflector) means the reflector needs to be deep. A reflector with depth and width approximately equal lets out a good bit of spill. Your halogen is probably in this range.
For an aspheric, any light that doesn't hit the back of the lens is either absorbed by the body or reflected out as spill, so capturing lots of light means keeping it close to the LED. That means you need a short 'focal length'. I know that's not exactly the right term, but it's close enough. Short-focus aspherics are thick, bug-eye looking things. Thin aspherics have long focal length. They will give a fine beam, but lose a lot of light.
Your battery pack is probably fine, depending on the driver you end up with. Or you could reduce the weight and/or increase the runtime by switching to lithium. The 18650 is by far the most popular. I personally like the 26650. That's C cell diameter, but longer. In case you don't know, the first two digits are the diameter in millimeters, and the next two are the length, and the zero on the end means it's cylindrical. Sizes are approximate. Cells vary slightly in diameter and considerably in length. Choose a cell and get reliable numbers for that cell before counting on an exact size (cells of the same model and brand will be pretty consistent). I won't try to discuss chemistries, that's a huge subject that's been debated ad nauseum elsewhere.
You will certainly want a buck driver, since your battery voltage is higher than the LED voltage. I suppose you are used to single mode (on/off), and if you're happy with that, Der Wichtel makes a driver suitable for the SST-90. I make one that can have high/low/off modes, and drive an SST-90, but it's a linear driver, which means you have to keep the battery voltage close to the LED voltage or the efficiency sucks. It turns out that 4 NiMH cells and one LED works out great. Eight and two or nine and three also works. And of course you can parallel as many cells or LEDs as you want, so you could run a triple XML board on a 9S2P pack. Unfortunately there isn't a combination of LiIon cells and LEDs that works very well. If you want more modes, your best bet is going to be a triple board or single XML. I've not seen or heard of a multi-mode buck driver that can run 9A to the SST-90, though Taskled has some that are 6.6A IIRC. Taskled makes great drivers.
If you go LiIon, you'll want to think seriously about a 'hobby' charger. These are designed to charge a variety of multi-cell packs quickly and to treat each cell individually, as is necessary for proper care of LiIon cells. Alternatively, a good balancing protection board can be built into the pack, but it gives you less control and less information about the health of individual cells.
Wow, thanks for the input. As we primarily dive cold freshwater a tight beam is preferred as the water isn't that clear. So it sounds like I will be looking into an aspheric. From what I see I believe that a flat window should be placed in front (air gap) so as to maintain beam focus. I was planning on making new battery packs and going Lithium. From my reading the SST-90 seems to be a very wide set of dies on a single substrate. From what I have read about aspherics I assume that they can focus a beam from such a wide input source? A later project will be a light for my GoPro so the SST-90 has some appeal. How many lithium cells would be required with your driver and an SST-90 and will it take a Piezo?
dave
dave
DIWdiver
Flashlight Enthusiast
Yes, a flat window is needed in front of the aspheric when in the water.
The SST-90 is a single die. The DX 44mm aspheric that Packhorse used can focus it very well. There was some talk that that lens was no longer available, then available again. You may want to make sure it's what you want.
My driver isn't a good match for Lithium and SST-90. You'd need 2 cells in series, and your efficiency would absolutely suck. Much better off using 2-3 cells in series and a switching (buck) driver.
No, my driver isn't compatible with Piezo switches. I don't think Der Wichtel's is either. Some (all?) of Taskled's are if you use 'prolonged' piezos.
For 9A, I think you're looking at Der Wichtel, KaiDomain (questionable quality, but probably piezo-compatible), or Ampere! (datasheet in German).
Here is the list I searched to get these results: http://www.videofoundry.co.nz/ianman/laboratory/research/driverlist.php
The SST-90 is a single die. The DX 44mm aspheric that Packhorse used can focus it very well. There was some talk that that lens was no longer available, then available again. You may want to make sure it's what you want.
My driver isn't a good match for Lithium and SST-90. You'd need 2 cells in series, and your efficiency would absolutely suck. Much better off using 2-3 cells in series and a switching (buck) driver.
No, my driver isn't compatible with Piezo switches. I don't think Der Wichtel's is either. Some (all?) of Taskled's are if you use 'prolonged' piezos.
For 9A, I think you're looking at Der Wichtel, KaiDomain (questionable quality, but probably piezo-compatible), or Ampere! (datasheet in German).
Here is the list I searched to get these results: http://www.videofoundry.co.nz/ianman/laboratory/research/driverlist.php
Last edited:
Thanks.
Yeah, found Packhorses mag/sst-90 thread seems the new lenses have a longer FL.
That list you have in the link is quite useful.
I noted the poor match between LIs and the forward voltage of the SST-90 at 3 amps. And from comments you made in other posts figured it might not be a good match. I have been looking at the DiveGeni driver as I used them in current canister lights with no problems for a number of years.
Looked into Hobby chargers per your recommendation and will probably get one. I note some are battery driven and some use mains voltage. Would be nice to get one that does both is not too expensive as I do some diving while camping.
Well at last is seems that my mental picture is coming together. Thanks for the guidance.
dave
Yeah, found Packhorses mag/sst-90 thread seems the new lenses have a longer FL.
That list you have in the link is quite useful.
I noted the poor match between LIs and the forward voltage of the SST-90 at 3 amps. And from comments you made in other posts figured it might not be a good match. I have been looking at the DiveGeni driver as I used them in current canister lights with no problems for a number of years.
Looked into Hobby chargers per your recommendation and will probably get one. I note some are battery driven and some use mains voltage. Would be nice to get one that does both is not too expensive as I do some diving while camping.
Well at last is seems that my mental picture is coming together. Thanks for the guidance.
dave
DIWdiver
Flashlight Enthusiast
The battery powered chargers can be run off mains power with a 12V power supply, which is often offered as an accessory. Be careful you get one with a high enough current output to feed the charger.
The Dive Light Genie is a PWM driver, and not a current regulator. While that makes it great for incans, it will destroy an LED instantly. You MUST use the LED Genie.
The Dive Light Genie is a PWM driver, and not a current regulator. While that makes it great for incans, it will destroy an LED instantly. You MUST use the LED Genie.
Thanks
understood about the LED Geni, I should have been more specific in my statement. LED life would probably be measured in 1000ths of a second with my current setup.
on another aspect, have you heard of anyone making a sliding heatsink/LED mount. I figure a delrin/nylon rod thru a gland in the tail end of the lighthead. A notch in the heat sink and screw thru the wall sliding in the notch would prevent rotation. Slide it in and out for focus. Would require regular attention to the gland/o-ring and probably recoating of heat sink paste to the walls of the tube but it would allow focus adjustment and be fairly easy to make with simple tools. I figure a simple push pull, all the way out, all the way in; spot/wide. Probably over complicating a first LED build by quite a bit, it seems obvious but I haven't found anything on it in the forums.
understood about the LED Geni, I should have been more specific in my statement. LED life would probably be measured in 1000ths of a second with my current setup.
on another aspect, have you heard of anyone making a sliding heatsink/LED mount. I figure a delrin/nylon rod thru a gland in the tail end of the lighthead. A notch in the heat sink and screw thru the wall sliding in the notch would prevent rotation. Slide it in and out for focus. Would require regular attention to the gland/o-ring and probably recoating of heat sink paste to the walls of the tube but it would allow focus adjustment and be fairly easy to make with simple tools. I figure a simple push pull, all the way out, all the way in; spot/wide. Probably over complicating a first LED build by quite a bit, it seems obvious but I haven't found anything on it in the forums.
DIWdiver
Flashlight Enthusiast
The problem with a simple push-pull is that the water pressure will push it all the way in as soon as you get to 10-20 ft. The other problem is that good thermal transfer requires obscenely tight fits. Not impossible, just very tight. And with any mechanism like this, the thermal compound will not want to go back into the hole once it's out, so you only get one or two slides before most of the compound is piled up outside the joint.
IMHO, it is for these reasons that I have never seen a sliding heatsink.
Have you looked at the MAG focus mechanism?
Someone (commercially) made a light with a moving lens that's similar, but I'm not sure it was a dive light.
I made a light in which the LED is mounted to the body, and the head just holds the outer (flat) lens and the aspheric, and is on screw threads. It worked really well until I took it diving, and the threads got jammed. The o-rings are on the outside of the threads, so that should have kept sand out, but maybe sand is between the head and the body, outside the o-rings. Or maybe moving the threads under pressure wrecked them. I haven't disected it yet to find out.
IMHO, it is for these reasons that I have never seen a sliding heatsink.
Have you looked at the MAG focus mechanism?
Someone (commercially) made a light with a moving lens that's similar, but I'm not sure it was a dive light.
I made a light in which the LED is mounted to the body, and the head just holds the outer (flat) lens and the aspheric, and is on screw threads. It worked really well until I took it diving, and the threads got jammed. The o-rings are on the outside of the threads, so that should have kept sand out, but maybe sand is between the head and the body, outside the o-rings. Or maybe moving the threads under pressure wrecked them. I haven't disected it yet to find out.
Was thinking that a small diameter rod would exert minimal force in descent but you may be correct. Detents, etc would not be good passing a gland. Avoided the screw head as my machinist skills and equipment would be marginal also as 10mm movement or so would likely be required it is a lot of twisting. Likewise moving the lens there is less bearing surface per diameter so exact fit and possible lubrication would be required. I was considering the length of the heatsink slug would be better at keeping the system aligned. Ah well far less expensive dissussing the system here than spending a lot of time/money repeating the discovery of others. I should just keep it simple at first.
cummings66
Newly Enlightened
If I made a light like that, I'd thread the end of my lights I build, then cap it so it's water tight, use the threaded portion to move the lens up and down to focus the beam. I've thought about trying that, but to be honest, the lights I make work beautiful in the water with just a reflector. The focused lens would give a spot under water but for the diving most of my buddies do we prefer a slight spread to the beam.
I've been thinking of building one of my next lights with this idea of moving the lens for testing, just haven't had the time yet to do that.
Mine are machined from a chunk of aluminium so I can do most anything I want to with it. I think the optics moving only would be doable, leave the LED locked. I'll see about testing that next time I order some parts. Should be easy enough to do.
I've been thinking of building one of my next lights with this idea of moving the lens for testing, just haven't had the time yet to do that.
Mine are machined from a chunk of aluminium so I can do most anything I want to with it. I think the optics moving only would be doable, leave the LED locked. I'll see about testing that next time I order some parts. Should be easy enough to do.
Not being much of a machinist but I would assume that a normal thread in any material of appropriate thickness would be fairly fine. From what I have learned about aspherics I would assume that about 10mm of adjustment would be required. That would be a lot of turning. That is unless some special thread design were used. Not so sure I like to do a lot of twisting on an o-ring underwater. I was considering the push pull as I could easily tighten down a bit on the gland like a stuffing box. Anyway I will follow the Kiss principal for my first units.
For tight beam I would use a single XML or buy a DX module with multiple LEDs like the 5 and 6 LED modules they sell. Surprisingly enough those multi-LED modules are very tight. For even more lumens, ditch the stock driver and run up the current with something like a Taskled H6CC.
The DX modules like 2 Series (7.4v) LiIon packs. With the Taskled, you can run up to 22 volts and be good.
The DX modules like 2 Series (7.4v) LiIon packs. With the Taskled, you can run up to 22 volts and be good.
Last edited:
