top-quark.blog

LED Navlight (boring technical)

One noticeable shortcoming of most weaponlight setups is the lack of a more sane brightness setting for general navigation indoors. If you have white walls, it's a nearly guaranteed self-blinding scenario, or a loss of night vision, at minimum. One of the few exceptions is the Surefire M9xx series. These, however, are expensive. I'm not in the mood to spend that much, so I made my own solution.

There were a handful of influencing factors behind my design.

Firstly, the technique of strobing a tactical light. One of the most common criticisms of tactical lights by non-experienced users is that an opponent will be able to spot your location. This is really only is a problem if you continuously hold down the light switch. The solution is to "strobe" your primary; the technique is usually to press a flashlight's switch intermittently. A company called NightOps recently began selling a strobing tactical light called the Gladius, which I believe intermittently supplies power to a Luxeon LED. It's expensive, though (surprise!)

My current setup consists of a Surefire G2 or Pentagon X2 light; both have the same endcap threads, and both will use the same pressure pad I bought for my G2. Unfortunately, the pressure pad's internals began to fail, so I began an overhaul. The switch is hard to (dis)assemble, though, so I took advantage of the situation to discover that grafting a secondary light in through the pressure pad would be the easiest approach. This has some shortcomings, as I'll discuss later, but this is just a prototype.

Design Criteria:

1. LED element - I have some pretty obnoxiously bright keychain LED lights that are overdriven constantly by two button-cells; overall voltage totals 6.5 volts. That some survive is something like natural selection. The LEDs that survive the gross overvoltage are very tough, and should survive the 6 volts I feed them with my two CR123As in the flashlights. Later on, I'll be upgrading to some Luxeon high-output LEDs.
2. Strobing - Figuring out what sort of strobe frequency to use was a bit challenging. I wanted a frequency that would be considered uncomfortable to look at directly, and would be difficult to pinpoint the point of origination from. Generating some animated GIFs in ImageReady lead me to believe that 30 Hz was pretty ideal (other tests: 5, 10, 20, 30, 40 Hz). This is pretty subjective methodology, but so too is the objective effect, so I think I'm coming out OK.
3. 6V Design, with power supply through grafts on pressure switch wires - The original plan was to wire in the secondary light in parallel with the Xenon bulb, but this wasn't practical, from an easy-modification standpoint, as well as a wire layout perspective. The approach I went with was grafting in the navlight in parallel with the pressure switch. Unfortunately, what this means is that the navlight is run in series with the bulb, which I suspected might hurt efficiency a good bit. How much, though? Incandescent bulbs are not particularly ohmic, so I just experimented, learning that the effect is not so terrible. I learned later that my navlight draws 23 mA at 6V on its own, and 30mA total in series with the Xenon bulb. At 30-some percent loss, it seems like a pretty crappy compromise, but it's really not so bad, when you consider that the Xenon primary draws a full amp.
4. Low weight
5. High durability

The balance of my design criteria dictated some sort of relatively simple layout, with an IC and epoxy potting. Pretty much out of laziness, I settled on a design using the astable 555 reference circuit. Elements Ra and Rb calculated out with rough values of 22 kOhm and 220 kOhm respectively, with C set as 0.1 uF.

Implementation:

1. I put the circuit together on some perfboard I had sitting around. Good performance.
2. With a working light in hand, the next step was operational testing on my rifle, which demands a higher degree of durability. Epoxy potting is the answer for things like this. I started by constructing a pour cup from a very high tech source - the lid of a macadamia nut can.
3. I don't have a vaccuum pump here, so I get bubbly pours. It's ok, though, since this isn't all that crucial.
4. The end product was pretty decent, after a good bit of heat curing. Squared up with some sandpaper, it's almost presentable

Testing indicates that the result is really quite good.

I have some qualms, though. Firstly, that it's a little too fast (20Hz is probably a better choice, in practice), and the light is a little too bright. This leads me to believe that I may have taken the wrong approach. A strobing primary, with a constant dull navlight may be the better choice, or perhaps even just a strobing primary. Without this prototype, though, I wouldn't have known. Call it a win.

I have a Luxeon 5W to play with now, but I haven't figured out an elegant and easy wasy to handle the thermal output. It's basically the same thing, except that the 555 doesn't have enough current handling abilit to handle the 5W the Luxeon needs, so the output uses transistor switching, in the form of a Darlington pair. Hopefully I'll have an update soon.

Entry touched up slightly on 8/13/06, to bring HTML into line, and to correct some layout issues