Based on the discussion in the LED thread, I sent JTR a message. Here is how it has gone so far:
ddrueding said:If you are interested in building something similar, let me know. I'll consider funding it (once we have a budget) and let you play for a while before taking it from you.
jtr1962 said:Hi Dave,
I'll think about it. Heck, if it's cheap enough I might build myself one while I'm building yours. Let me see what's available as far as emitters go. I have a gut feeling going with high-bin Crees would be more viable but it might end up costing a lot more. I've actually wanted to build an LED light of some type putting out ~100K lumens.
Note that this is a pretty busy month for my consulting project, but if this project doesn't take a huge amount of time I might squeeze it in. It looks like the builder used off-the-shelf Chinese regulators to power the emitters. Those are readily available on eBay, and pretty inexpensive ( from under $1 to maybe $25, depending upon how much power they can handle ). Battery could be any one of a number of sources, like conventional li-ion, LiFePO4, NiMH, etc. Given the high current requirements, probably LiFePO4 would make the most sense. It might also be interesting to have an option of power it indoors from a 12V or 24V power supply. You could build your own indoor sun room! And of course knowing me I'd probably make it dimmable. If it runs for 10 minutes putting out 100K lumens then it'll run for something like 20 hours putting out 1,000 lumens. That might come in handy during a blackout.
ddrueding said:I have a bunch of 3-cell and 4-cell Lipo packs from my drone stuff. 5Ah and 8Ah batteries that should be enough for some serious use. I suspect most of my applications would be running on battery or tied into an automotive system, so optimizing for 12v+/- would make the most sense. 100k+ lumens would be a nice target.
jtr1962 said:Hi Dave,
Just starting some preliminary research. For the emitters, there are any number of 100 watt eBay LEDs ranging in price from about $8 to over $30. Most of the lower priced ones aren't terribly efficient, and frankly they seem like garbage to me. The best one seems to be this. They claim 8,000 to 10,000 lumens. Of course, this being eBay, I'd take that with a grain of salt. On Mouser I looked at a bunch of high lumen COB (chip-on-board) LEDs. The Crees seemed to come in at 200 lumens per dollar or more. The best bang for the buck in terms of lumens per dollar are these or these but they're on order. ~17,800 lumens for $45.73, and efficiency of ~125 lm/W. The second option, which is available, is this. These are 11,430 lumens @121 lm/W for $41.60. Not as good bang for the buck as the other two, but they are available, and the 35V operating voltage is a lot easier to work with than the 54V of the other two. Since Mouser charges me sales tax, these would be $45.30 each plus shipping. Note however there is quantity discount for 10. That brings the cost to only $40.03 with tax. If I made myself a light and we order 5 spares, we would only be paying $34.96 each with sales tax.
I found what I think might be a great regulator for only $4.31 each. You would need one per LED, so ten total.
I haven't yet bothered researching a heatsink, but I can probably cobble together something for under $50. Using the more efficient LEDs, my estimates are we would only need to dissipate about 500 to 550 watts of waste heat.
The rough total then for materials should likely come in at $500, give or take. That assumes you'll be using some of your spare batteries as the power source. I'm not sure yet how much the labor charge might be, but I don't think this is a very labor intensive project. My guess might be $200 tops. There isn't much design involved in this project. The hardest part might be figuring out how to adjust the regulators, and also adding on some circuitry so you can vary the intensity.
I don't know if you'll also want optics to focus the output. Those would of course cost more. However, I think they're something you can easily add yourself once you have the finished light in hand.
All in all this is a somewhat expensive project but probably not as costly as I had thought. A few years ago something like this wouldn't have even been feasible. I also like the fact this is a modular design. If making a light with ten emitters is too costly, the project can easily be scaled down to five or three or even just one.
ddrueding said:Thanks for putting in the effort. This seems quite reasonable for what it is. Adding even more modularity is appealing to me; I've had really good luck with heatsinks from this manufacturer that are available on Amazon.
If 2" of this 10" wide x 3" tall profile is enough to handle 3 emitters and regulators, I'd be tempted to build them as modules and be able to hang them from a rail. 4 modules (12 LEDs total) could hang from some aluminum square bar and be daisy-chain-able using XT60 or similar connectors (those are what my batteries already have).
As always, I know just enough to ask stupid questions. All ideas above are just that and very flexible. The available "Plan B" LEDs seem fine. I'll send you a check and we can get started?
jtr1962 said:Hi Dave,
By Plan B LEDS, I assume you mean these?
Making these in groups of three, each with a connection, will enable me to test each module with my 12V, 60 A power supply. That simplifies things a lot given that I won't need a 12V source capable of putting out in excess of 100 amps.
This might be an alternate idea for the heat sink. One heat sink like this with a 120mm fan could deal with the heat just fine. It's worth noting given the limits of run time due to battery capacity that the heat sink's thermal mass will play an important part. The battery may well be depleted before the heat sink temperature rises to levels which might be of concern even without a fan. I'll have to experiment with this. The heat sinks you picked should work fine for 3 emitters and regulators if we want to go that route instead. I guess you would mount them from the rail such that the fins were parallel in order to use one fan to cool all the modules.
$1K will easily cover labor and materials for this project. I can order the stuff I know we're sure of for now (the emitters and regulators) before we decide on exactly what heat sink arrangement and connectors we'll use. This will let me test the concept using some heat sinks I have. I'll probably built myself one for now with maybe 4 emitters, perhaps more depending upon how much money I want to spend.
I guess that's it for the time being. Yes, let's definitely get started and blow what the other guy did out of the water!
ddrueding said:Yup, those are the emitters that I was talking about. Also optics; how narrow a beam is possible with normal close-fit optics? I'll also need to look at some kind of shroud to prevent people who are in line with the unit from being hit with any of the light directly.
I'll be talking with a buddy who does search and rescue from aircraft later, how would this compare to the normal searchlights they use?
jtr1962 said:Hi Dave,
I'm thinking with an aspheric optic like this we could get a very narrow beam. Note that the beam angle with this optic depends upon how far from the emitter it is. They say you can get 5°.
This might be comparable to an aircraft searchlight. The specs for this one are a 1600 watt xenon short arc lamp and a 4° to 20° beam. IIRC, a xenon arc lamp is about 50 lm/W, so we're talking maybe 80K lumens here. If we focus the 120K+ lumens from 12 emitters into a fairly tight beam we may have something very similar.
A shroud is probably a good idea if you're using optics here. The focused beam will be much, much brighter than full sunlight. It can easily cause eye damage (that's why I'm probably not using optics on mine).