MJ-852 mod with XM-L2

After the failure of my full-custom attempt at an XM-L handheld with 1x 26650, I decided to try my hand at a mod. I looked around for a nice body, intending to replace the driver and LED. I settled on the MagicShine MJ-852, despite MagicShine's poor reputation for quality. I figured I could clean it up, replace the O-rings, and I was planning on replacing the driver anyway. Here's a picture of it as it arrived.


I did a pretty thorough evaluation of it here: http://www.candlepowerforums.com/vb/showthread.php?316449-Anyone-tried-this-light


I was hoping to use ahorton's 27mm aspheric lens, but there's just no way to fit it in the light. You'd have to bore out the head until the wall was dangerously thin. So I bought a couple aspheric Fresnel lenses from jspeybro. At 25mm dia and 2mm thick, it's a much better fit. The Fresnel lens is a compromise, not as tight a beam and not as bright, but it works quite well. I did a brief comparison of the lenses here:
http://www.candlepowerforums.com/vb/showthread.php?369326-ahorton-vs-jspeybro-lenses


It turns out I had to chuck the entire guts and rebuild the pill. The driver is two boards connected by three 1/8" diameter brass rods. The front board has the linear very-low-dropout regulator, and the back board has three hall sensors to drive the regulator in low-med-high (0.3A, 1A, 3A). I was quite impressed with the results when I put it all together and got it working at 8 PM the night I left for a week of diving. When I took beamshots today, I discovered the output is much lower than it should be. When I measured the tailcap current, it's only 2A. After the diving season ends in a few weeks, I'll tear it apart and figure out what's going on. Until then it will do quite nicely. It's my new favorite light. Here's a couple pics of the new pill.



and the lens and lens spacer



As you can see I machined a new aluminum pill. The LED is an XM-L2 T6 3C (5000K) on a 16 mm copper star from Illumination Supply. It's glued in with Artic Silver epoxy. The driver board is screwed to the back of the pill, and there's a thermal pad between. Max power dissipation of the driver is around 2.5W, so that should be sufficient. The wires were soldered to the LED before installation, and solder directly to the driver board, so they are only about 3/8" long, 24AWG solid copper.

I deliberately defocussed the beam slightly to remove the artifacts in the image of the die. Here's a beamshot. The light is 10 feet from the target, and you can see the inch and cm markings on the scale. At 20" it's about a 10 degree beam (some would call it 5 deg). My cheapo lux meter shows about 650 lux max.




Hope you guys like it! I like it enough I think I'll make a spare. And my dive buddy on my last dive wants one too. It was every bit as bright with a wider beam compared to his 10W HID canister unit!

Nice build and great to see some people using these lenses as well!
I hope you like the lens. I didn't defocus it in my lights, and it's a nice tight beam. I can imagine that some people like it a bit wider, but since in my case it is my backup light, I prefer something that is better at signaling.
Still plenty lenses available btw

Johan

Yes, nice build DIWdiver!

I'm playing with the lenses from jspeybro myself, in my case the 50mm version.
Will post comparisons later this week as with everything it's a mixed blessing.

I figured out why the current was low. The resistance of the traces to and from the current sense resistor added enough extra resistance to throw off the current regulator. A couple of wires added over the traces have brought the current up to 2.85A. New board layout should take care of the problem completely. Unfortunately, it will be weeks before I know for sure.

Two of these took a chamber ride to 300 ft tonight. No sign of any leaks.

One was the very first mod, after having leaked on it's 15th (or so) dive. It seems I didn't tighten the front bezel sufficiently (shame on me for careless maintenance). The other was a new mod for a friend. The new one was complete sans battery (testing before delivery) and the old one was just the shell.

The ride consisted of an excursion to around 60 ft, return to surface, 200 ft, return to surface, 300 ft for 10 minutes, 20 ft for 20 minutes, 10 ft for 20 minutes. The shallow depth is because I believe some leaks will seal under high pressure, and only leak under low pressure. In my mind, testing at multiple depths is required to ensure proper design and implementation.

After removing the lights, I wanted to test the chamber, to see how deep I could test. I only tested the lights to 300 ft because I was afraid the front glass lens might fail, and I didn't want to ruin a good light or two testing to a condition they'd never be exposed to. With nothing but water in the chamber, I raised the pressure to 486.6 PSI (according to my expensive sensor and decent DVM), which equates to 1069 FSW. I could have gone deeper, but I didn't really see any point. Plus, everything was designed for 500 PSI.

My chamber may be a PITA to use, but it's very safe and has proved it can be used to test at any reasonable depth!

Did you make or buy the driver?

Made it.

Very nicely done!