First mod: Luxeon Lantern! (long post with pics)

Candle Power Forums

Help Support Candle Power:

KevinL

Flashlight Enthusiast
Joined
Jun 10, 2004
Messages
5,866
City & State/Province
At World's End
Wooohooo... my first mod is done, at long last.

Special thanks to Doug Owen who went out of his way to illustrate to a clueless n00b that 1. a LED must have a power source with higher voltage than its Vf 2. a LED doesn't like too high voltages 3. Use a resistor or ohm's law will own you.

I initially planned to use a food jar to create a lantern. The thing must have a power switch on the TOP, so I can tap the switch and the lantern will come on. No clumsy side switches for me. Quickly on, quickly off. Passive circuit, driven with 15 ohm resistance for 120mA. After all, just a lantern. Power source would be 4 x AA since 2 x AA barely produces any light at all. Caters for both alkaline and NiMH. Usually NiMH, since I have a set of 4 of them - 1300mAH, leftovers from years past, wimpy enough as it is, but strong enough to drive a Lux at decent brightness.

Of course, no plan survives first contact. The 1W Q4H from the Shoppe arrived as planned and I stuck it on a nice green chipset heatsink to test it. I started with solid wire because I had some extra Category 5E network cable sitting around (one of the hazards of the job.. /ubbthreads/images/graemlins/grin.gif )- a couple of 1ft lengths too short even for patch cables. Slicing open the outer PVC jacket yields 4 pairs of 24AWG wire. I won't be using it again in a hurry. I quickly learned that solid wire is a pain in the rear because it tends to retain it shape, and it breaks very easily.

On my quest to find decent stranded wire, I discovered an ATX computer power supply that could be scrapped to yield large amounts of 22 and 18 AWG wire. 18AWG is pretty damn big. The power supply also kindly provided a 8.2 ohm resistor, which, judging by its size, is a 2W resistor (though realistically it will never have to dissipate more than 300mA), and that brings current up to 220mA, which is pretty good. The resistor was desoldered and now has a new home.

Along the way, I had a bright idea that I haven't yet seen on CPF. The power supply yielded plenty of Molex 4-pin connectors, and the smaller floppy drive power connectors. I considered using that but the male end was so difficult to solder and hard enough to find that I gave up anyway and went back to the bigger and more forgiving 4-pin connectors. As I was prepping them and removing two of the unneeded wires, one of the connectors left two nice parallel cuts on my left hand.. well, you know what they say - the project's got to get its blood sacrifice for all to go well, and if you have to donate, so be it /ubbthreads/images/graemlins/tongue.gif

The idea behind a Molex connector is to support MULTIPLE battery packs. The Luxeon is connected to the male end of the connector, and the battery packs and attendant resistors/power driver circuitry are attached to a female end. Essentially, this is the same as the "dumb" LED heads on some lights, with all the control electronics attached to the battery packs. I'm working on building a lithium battery pack for this one to take 2x123s after my Surefire lights are done with them. This pack will be direct drive, because after the SF lights eat up most of the energy, 2x123 will give out around 4.8-5V and only drive the Lux at something like 300-400mA, which is well within spec especially with my kind of heatsink. Already confirmed this with two spent cells and my DMM.

The salsa jar also proved to be troublesome. It has a metal lid, which is nice to drill holes in, and a transparent glass surface. However, the Luxeon can be blinding if looked at directly, so that's not such a great idea. But the main problem is eating all the salsa at one go so I can use the jar. According to the side of the jar, there are about ten servings. *NO!!*

While on another raid at the supermarket looking for a power strip in the electrical section, I found a bunch of Eveready $3 lanterns. They take 4 D cells, run a KPR113 4.8V 0.75A bulb according to the writing on the lamp and the packaging. OK, for 3 bucks I can't possibly lose, so I bought one and took it home. It's plastic, it's not waterproof (splash resistant at best), but it doesn't quite matter. The salsa jar is out of a job.

I cut the central bulb post off with a hacksaw, this kind of plastic is very easy to saw through. The heatsink was fastened to the white bulb platform with two screws (hidden underneath the platform), I wanted to use the two holes as strain relief for the wires. This works as well with heatsinks lacking the mounting holes. The Lux is not epoxied on, just a drop of Thermalloy ThermalCote grease - epoxy will follow tomorrow.

Other than that, it was just drilling the hole in the translucent diffuser (which solved my LED-blinding-people problem) WITHOUT the proper tools to accomodate the switch (it's quite a painful process, I need a real drill), soldering everything up, doing continuity checks with the DMM, discovering I'd soldered the Lux BACKWARDS (wrong polarity), burning myself on the soldering iron a couple of times (thankfully nothing serious enough to even warrant attention - fast reflexes help), testing, re-testing, stripping wire like there's no tomorrow, taking pictures.. etc.

Pics? Sure.

Heart of the lantern with stock bulb:
lantern1.jpg

Battery compartment and diffuser:
lantern2.jpg

New guts, new engine, Molex connectors and batt pack:
lantern3.jpg

The lantern looks exactly the same as stock except for the switch, and the fact that it's MUCH whiter:
lantern4.jpg


More to follow when I get the full picture gallery up. There are a lot more but these are just the essentials.

The lantern runs 220mA, doesn't run hot at all (but of course!!), can be held upside down if I want to, there is a small handle that enables that. Is it brighter than stock? No. Is it close to stock brightness? Yes, at less than 1/3rd the current and with support for multiple battery packs, including a lithium 123 eater to finish off the cells. Am I richer for the experience? Certainly, and my soldering technique steadily improves towards the end (commensurate with number of burns sustained /ubbthreads/images/graemlins/wink.gif ). Better make my mistakes now than with the TWOL that just got delivered....

And last but not least, THANK YOU CPF!! There is no stopping now, my Hotlips + TWOL arrived today, there is no turning back.. whoever thought building lights could be so rewarding.
 
Thanks /ubbthreads/images/graemlins/smile.gif Future plans, if I'm up to it, include a 3W Lux Lantern.

This enhanced lantern will feature a low beam, high beam, and MAX beam mode, but only a single battery option. Two switches. One switch will be cabled to a 5.6 ohm resistor. When this switch is turned on, the lamp will light at 375mA for low beam. The second switch will be cabled to a 3.3 ohm resistor, for 635mA of light - high beam.

When both switches are flipped, there will be two parallel current paths to the LED, and the two resistors in parallel will yield ~2.07 ohms of resistance, and close to 1 amp of current for maximum light, a full 80 lumens. Some lamp indeed.
 
You could use a tri-light lamps switch like I used
in my 3-cubed mag mod. That gives you off, resistor A,
resistor B, or resistors A & B in parallel. From
a single rotary switch.

greg
 
Now for the real objective of the Lux Lantern - as a 123A battery eater!

I crafted a new battery pack from a 2AA holder, after you quarter it (sliced three ways.. yes, three because I remove some of the center part), you can put it back together the way you want it in order to fit two 123 cells. Put it on a plastic base and use cyanoacrylate superglue to hold the parts in place. If necessary, bend the little springs to fit. Voila - series 123 battpack. Apparently, that Terralux case is a really really good fit!

I then soldered the leads of the battpack to another Molex connector, which gives me additional battery options should I need it.

Believe it or not it's direct drive. Yup, DD 3.2Vf Lux1 on 2x123, it can be done provided a few conditions are met. First, put the cells in a Surefire 6P and run them till the light goes out.. then you put them in your lantern. /ubbthreads/images/graemlins/grin.gif since I'm the only one using the lantern I don't feel the need to put a warning label on the battpack that says don't ever use fresh 123s or you'll blow the Lux into low earth orbit.

When the cells are fresh off the 6P, one clocks in at 2.55V, 2 flash amps, the other one is 2.33V, 0.5 flash amps (unusual battery behavior is noted in another thread in the batteries section). Both together in series drive the Lux at 220mA with no resistor, as measured by my DMM. The reason for no resistor is that I want this to really suck the batteries to death.

Runtime tests - it's still a pretty darn bright lantern. By coincidence, 220mA is my measured current with the normal 4xAA NiMH battpack, so no noticeable decrease in brightness. It takes 9 hours for the output to drop to a level comparable to moon mode on most Luxeons, and at this time the cells are running 1.43V on the stronger one, 0.20 flash amps, and 1.13V on the weaker one, 0.10 flash amps. By this time the light makes a GREAT nightlight. Not too dim, not too bright, and for dark-adapted eyes, perfect. So I go to sleep with it on.

Wake up.. 8 hours later, the light is still on, and about the same brightness. Test halted as I had to leave the home. The tests continue tonight. Yes, it involves giving the lithiums a break, but I figure that's the way real nightlights run (at night). Besides, it's a battery eater, not a mission critical light (that's what SFs with fresh cells are for /ubbthreads/images/graemlins/smile.gif ).
 
As an Amazon Associate we earn from qualifying purchases. Product prices and availability are accurate as of the date/time indicated and are subject to change.
As you notice, the springs in the battpack are very loose. This was a design consideration in order not to have the superglue forced apart by the springs at maximum tension, however the cells are also slightly loose. Don't shake the lantern too much. I would redo it all over again, but ironically, the superglue is proving to REALLY be superglue this time round, so I'll just let it slide.

Followup:
The Extreme Battery Runtime test is over, and the light has done surprisingly well. To summarize:

Normal brightness ("dead" cells unloaded from SF 6P light):
Cell 1 starts at: 2.55V, 2 flash amps
Cell 2 starts at: 2.33V, 0.5 flash amps

Runtime: 9+ hours till light dims to "moon mode" (about as bright as a KL1 in moon mode)

Round 2 in nightlight mode.
Cell 1 starts at : 1.43V, 0.20 flash amps
Cell 2 starts at : 1.13V, 0.10 flash amps

Light runs for FIFTEEN and a half hours before the brightness drops to where it is no brighter than the indicator LEDs on my keyboard. To run it beyond this point would be possible but utterly pointless since the light level has dropped to the point where it's no longer useful even in total darkness except to tell you "you left it on" (and leaving it on is the point, anyway /ubbthreads/images/graemlins/grin.gif ). I cancelled the test and took final measurements:

Cell 1 ends at : 1.08V, 0.12 flash amps
Cell 2 ends at : 0.99V, ZERO flash amps - or at least the amperage is so low that my inexpensive $9.95 DMM is incapable of picking it up at all! (Anyone want to make me a present of a Fluke? /ubbthreads/images/graemlins/grin.gif )

Man, direct drive can really bleed a cell quite literally to death. At first when I saw the DMM "flatline" (0.00 on the display), I thought I either didn't plug the lead into the amp measurement port, or something else was wrong. Seeing 0.12 on the next cell convinced me that yes indeed the cell is well and truly as dead as a doornail. I am surprised the Lux ran so long at such low voltages, Vf for even a Q4H bin should still be above 3V nominal.
 

Latest posts

Back
Top