Triple Nichia219 Goodness in a mini-mag

Flashlight Modding and DIY Parts
1

Gentlemen,

I was asked to make a mini-mag by a member here for him similar to the mini-match V2.0 I had posted awhile back. It was to be bone stock on the outside, but with high CRI goodness on the inside. Besides the emitters, this build differs from the original in another way, and that is the pill construction. Call this one V2.1

Due to the twisting mechanism of the MM, and how these normally are assembled it us usually not practical to use a double sided driver. On the last on I got around this with by using a small strip of copper epoxied over the chips for the negative contact. It was functional, but not elegant. For this iteration, something a bit more refined and structurally sound was tried.

From brass barstock, I milled a cup for the driver to sit in which would not only allow for a much sturdier ring for the negative contact, but also not add too much overall thickness. This cup then will hold the 3mm thick copper heatsink. Here one can see how the rim sits above the backside driver chips:

Here is the top view, with driver installed and soldered to the brass along with the thermal pad in place:

Here is the pill along with the copper heatsink in it's pre-polished phase:

The copper disk is an extremely tight press fit into the brass. In testing last night, this design would heat up the mag head just as fast as the previous build, so heat transfer is working well. Not the most scientific test, but my thermocouples and meter were left at work.

Next, 3 nichia219 emitters were reflowed onto a parallel star that was lapped to the pill, and affixed with arctic silver while clamped. The two leads were soldered and turned out like this:

I've since switched from using a drill bit to using the lathe to bore out the heads. I find that the finish is a lot smoother, straighter, and I can keep the tolerances much tighter. Here is the bottom shot of the bored head with the new pill in place:

I really like how this turned out. Having the solid brass ring for ground contact should last a long, long time for switching.

The last step was to prep the body. This was rather straightforward in that the head end anodizing was removed, and the stock black plastic insert was modified to allow positive contact. I milled a (very) small copper insert, drilled a hole in the plastic spacer, and assembled it.

Note that since the top of the stock black spacer isn't used (except as a stand in the above photo), it is removable. Something to keep in mind during battery changes (don't want to lose it...). After final finishing, the end result of all this looks "almost" stock

Final thoughts:

The nichias219's to me have been a love/hate kind of deal (love the cri/hate the low output). Having 3 of them in a svelte package really alleviates the lack of lumens, and I found myself carrying this light around for most of the evening just to watch colors pop in dark rooms. The final output numbers ended up as follows:

High: 600L (dropping to 550L after awhile)

Med: 190L

Low: 26.1L

I had planned on getting some beamshots last night, but it turned out to be a drizzly / foggy evening - not really conducive for quality pics... I hope the new owner is happy with this little mini. Now to start building another (for me ).

Thanks for reading,

-Match

2

Beautiful work as usual, Match! I will be building one similar in micro form that a member has commissioned me to build. I'm sure another one will follow for myself. First I'll be making one for myself using triple XP-G emitters as my test subject.

3

All I can say is, WOW! Beautiful work. Thanks for sharing.

4

Is that the ld-25 driver? If so, which one 2a or 2.6 or is it the ld-29 running two 14500’s? I’m interested because I want to do a similar light engine in a trustfire t2 and I’m struggling to understand the vf deal on series/parallel emitters, I figured parallel would suit better as the specs on the series star I found suggested a vf of 8.4v which made things a bit of a none starter.

Nice job on the light engine looks really sweet, like a factory set up. 8)

Edit, its not the ld-25, that’s a one sided board on the bottom, so is it ld-29? :bigsmile:

5

Thanks for the nice comments, guys.

Gords1001, the driver is a 3amp buck from Kaidomain (link here). There's a few sources for this board, my favorite was a guitar shop in Washington (don't ask) until their supply ran out. I've had good luck with them. Despite many sites claiming 17mm, the actual diameter is 19mm (18.92mm), so it really is a squeeze to get it in there.

6

Having machinery gives so many more options...

Very nice build. Since the board is parallel, it allows for the higher amp output, what do you figure each led is seeing for amperage? Is there a way to know how much amperage each led is seeing? and do you think parallel will shorten the lifespan on the leds?

Very nice Match. It's hard to deal with these small lights.

7

Man I like what you did there .

8

That’s strange, I’d not looked properly before but there’s this ld-29a

http://www.intl-outdoor.com/ld29-28a-12-cell-circuit-board-p-416.html

And I have this waiting to go in a Maglite xm-l build

I should pay more attention but I’m sure this is called an ld-29 board on the invoice.

Ah well, I’m glad you like it, there’s a good chance I’ll like it too. :bigsmile:

I certainly like the ld-25 board I put in my edc, the slight drop to 2.6a is nice in a smaller host and I’m not convinced you’d notice the lumen drop without an integrating sphere tbh.

http://www.intl-outdoor.com/ld25-26a-35modes-circuit-board-p-414.html

9

Nicely done, again.

10

Great read. I’ve used that board with 7 nimh’s for 3 x xre and it works great. Still have some.

11

Very tasty match. I always wondered how to use the thermal pads supplied with some drivers. The description is usually to smash it over the driver. It still amazes me how big things look in pictures where in reality everything is usually quiet small.

12

Everyone is out of stock, except for ebay at $13 each. Guess demand is higher than supply.

13

I have used one of these which I think is the same with no problems. You may want to ask if they have stock first if you are ordering though.

http://www.lightmalls.com/super-output-ssc-p7-led-driver-board-low-high-middle-5-5v-15v-2

14

Theoretically each led should see ~1A. Knowing full well that there are most likely small vf differences, the actual numbers might be 1.12, 0.86, 0.98....who know? Due to all being mounted onto the same pcb, the only way would have been to test them beforehand under load, but being on the same pcb also helps to ensure they're all heated evenly - so the margin of difference should be minimal. Does it shorten the lifespan? Possibly, but then again I've yet to lose one led that wasn't due to negligent and/or blatant abuse :)

After building that 6xXML light over a year ago (and hardly using it) I've really lost interest in large lights...that, and I like the challenge that minimal space provides.

15

Very nice work :slight_smile: