Noctigon 20mm copper triple XPG/XPE/XTE coming soon!

Yes lens would add another light blocker, same as the grease will. The only thing the grease might do is change how the light refracts.

I don’t plan to use glass, just an o-ring for a seal.

Too bad the lenses are not more easily available. I bought mine with all the rest from IS with Nichias. But gosh the shipping prices outside US.

Yeah true, I use copper for the spacer since I have it and it’s thickest from the bars I have. I stack 4 and solder them. Provides mass to absorb the heat.

I could stack five all the way to 15.6mm (1mm less than the height of original reflector), that’s the max I can do. With a 1.55m MCPCB.
With Noctigon 1.75mm (if it really has that thickness) it would be 15.4mm.
All with protected Sanyo UR18650FM 69.5mm long battery.
I would get 14.4mm with 5 layers of copper, but the 4 layers are already giving it a weight.

And only 1 layer of 2.5-3mm is all that is needed to make it tighten with an O-ring in front of the Carclo 10507 lens.

Could possibly shorten the head by 12.0-12.5mm, almost half an inch.
But don’t have anything to remove the thread in the inside for o-rings :confused:

I’m not sure if shipping would be cheaper or not, but ledsupply.com also carries the Carclo 3up optics:

http://www.ledsupply.com/carclo-optics-3up.php

Thanks, I have it. Ledsupply has only $35 international shipping.
IS is fine as long as you fill the whole package, $6.8 is not cheap but not so bad. The cheapest you can get I think when shipping from US. It’s just regular airmail.

Well I was prepared to be disappointed by the TIR optics, but I like it! (the frosted 10508, not so much) The plain 10507 has a big hotspot with soft edges, but very little spill.

The frosted version is only good for lighting up an entire room in one go without having to worry about where the light is pointed - I'm sure somebody out there will say "wow, that's exactly what I've been looking for!' - but not me.

Ceiling-bounce test, 3XP Convoy S3 @2.8A vs. XML2 Convoy S3 @3.5A, the 3XP is MUCH brighter. And given the way triples in parallel affect the effective Vf, it should be able to maintain the rated current down to a lower input voltage than the single XML2 which will hit the Vf ceiling sooner.

p.s. I REALLY REALLY HATE the design of the S3, no removable head - everything has to be loaded through the front, then the pill tightened against the bezel by reaching all the way through the body from the tailcap end. Convoy S2 or TR-801 from now on for me, thanks...

I think the problem is the tir’s are meant for household (cant think of the right term) lighting, not torches….we just pinch the tech and hope it works, the frosted lense sounds perfect for one of my intended uses for these boards, namely making some crappy mr16 dropins far more acceptable, I’m a bit away from this as I’ve hit my paypal limit for this year lol, I’m busy jumping through the hoops necessary to get my paypal verified.

I do intend to fit one of these in my blf a8 with a stacked qlite, I think with some iS high cri xp-g2’s it’ll make a very pleasing upgrade.

Damn, glad I never bought an S3!
Glad I have this and made FastTech list it too.
Or you can get it from fancyflashlights but the tailcap silicone boot is too high, the FT version should be better. The surface is really great as the whole light is anti slippery.

What I’ve seen from all TIRs, is that they have big hot spot and no spill. At least the Carclo’s I have from IS and tiny 12mm TIR.

How do triples affect effective Vf?
XP-G2 should have low Vf where as XM-L2 got a little screwed up and the Vf is higher than XM-L or XP-G(G2).

How can the triple be brighter? Sure it may be more effective and the XP-G2 has lower Vf but still.

I want to see pics and measurements if possible :slight_smile:

xpg-2’s Vf is not only lower when compared to an XML-2 at the same current but because the current is divided by the 3 LEDs, it is EVEN lower (3A through triplr XPG has a lower Vf than 1 XPG at 3A). At 3A, each LED is getting on 1A and the Vf at 1A is very good for a linear driver light.

What I’m most surprised by is the fact that comfy noticed the 3xpg-2 at 2.8A is brighter than an xml-2 at 3.5A. I was doing some rough estimates from Match’s data and it didn’t seem like I would benefit much from a triple xpg-2 (besides the lower Vf) unless I was pushing greater than 3A, so I decided not to build a triple xpg-2. However, from what comfy is saying, I might just go ahead and build a triple xpg-2 after I finish another nichia219.

Direct Drive running ~7A on an AW IMR18350 Fireplace is 18’ away, mantle is 81.5” long, the light fixture on the ceiling is 8’

We know lower Vf (or higher voltage) will increase the current draw. My thinking is that since multiples in parallel draw a higher current from the same source in DD as compared to a single in DD, that if you regulate the current down, that should extend the runtime before the cell voltage vs. Vf starts to reduce the current.

For instance checking with a really low voltage, a single CR123, the 3XP draws 540mA. A single XML2 light with an identical driver only draws 420mA from the same CR123.

Convoy S2 is cheaper than the X-Power, and the pattern of knurling & grooves on the head would make it easier to hide the cut line when the head's shortened to fit the 3XP. And the TR-801 has a really slick pill design, I think it's better from a heat standpoint than any of the other little tube lights. TR-801 has about an extra 1/4" or more in the head/bezel that can be cut off when converting to the 3XP and ditching the glass lens (and cutting off the ridiculous 'strike bezel' BS on the front).

XR-E by default, driver unknown, sure as host why not.
Although the MCPCB seems to be hiding inside a lot, so putting in a 20mm board 3XP board, maybe a no go.
The stock XR-E reflector is most likely useless.
So much work to make it work.
If it tailstands, but that’s something to always check or get lower boots.

S2 doesn’t seem to have the thread inside a lot, so you may not be able to cut it as much.

At least it may be useful to others I hope, what host they can pick and the pros cons.

What could I use to remove the inner thread? Aluminum/dural head it is. Dremel, drill I have, can get, drill bits but not those you have to eat from side and some grinding ceramic bits for dremel. Need 5mm removed.

It's not necessary to shorten the head, you can always use a round plug on top of/inside the original pill as a spacer so the pill/driver sits in more or less the original location. The spacer required for a TR-801 would of course be much shorter than that for the S2, since the 801 starts out with such a short reflector, much like the S3. Using a spacer in the S2 would mean you're carrying around a much longer flashlight than what's needed to make the insides work, so shortening the head can be thought of as just an optimization, not a requirement.

If you're looking for a light that can be converted with only pliers and a screwdriver I'm afraid you'll end up disappointed. This is not a p60 dropin.

I have the Xiaozhi/D4/X-Power, that’s where it’s going, I’m just thinking of using only 6mm of copper spacer instead of my 11.5mm. Sure I can change the driver spring if necessary and it will work fine but why if I can cut it shorter by 9mm.

I could put there even bigger spacer of 15mm but it gets heavy, I mean really front heavy due to the chunk of copper in there. 6-9mm of copper spacer is ok, that’s fine, 12mm gets heavy, 15mm even more.

I can cut it and file it flat, just dunno if any of the drill/dremel bits will be able to eat aluminum to remove the inner thread to make it flat for double o-rings.

Man thats some awesome work comfychair. You make it look so easy. I don’t have the proper tools to make those cuts and keep everything clean. Otherwise I would tackle this build and make a shorty triple like you mentioned. Looks like a very nice EDC light.

Have you considered cutting and threading for 18350 use? Maybe then cut a second body to use as an extension like an l2m. 18350 for super compact triple or 18650 when needed.

It's just a cheap import drill press ($129-139 at Lowes/HD, maybe even Walmart), a vise, and leftover hardware parts to make some fixturing so I can make more or less concentric cuts with small Dremel end mills. There's all kinds of slack in the spindle, I just have to make some approximations and try to compensate for it. Make small cuts and stop a lot to take measurements. Most of it is done by eye. For instance o-rings can be squashed a lot or a little and still seal, so the ID of the bore in the head isn't super-critical. It just has to work.

Other than the drill press, it's just basic stuff like an old Dremel (with a broken switch) and cordless drill, sandpaper, and hand files (diamond files are amazing, get some). The diamond cutoff wheels are a 10-pack from Fasttech, I think only .025" thick, make awesome cuts on AL and brass/copper. Trying to make these cuts by hand wouldn't be possible, it's the drill press and a solid surface to move the parts around that makes me look good.

I've had the idea of a cut-down 18350 version for a while but I can't do any rethreading, so it would have to be sectioned and then machined for a press fit back together, and that only works with just the right body design where both the join can be hidden AND there's enough wall thickness to accept an inner step on one piece and outer step on the other. I'm still looking...

This is my second time attempting this, last time I had the ENTIRE POST done, hit preview, and it said I was not recognized and not allowed to make new posts, needed to log in first... crossed my fingers and clicked the 'back' button, but no, the whole thing was gone... well...

----------

AGAIN, the TR-801 pill...

Cut out the entire original shelf, took a little off the top to get a nice flat surface around the perimeter, then cut a counterbore for a new shelf to be pressed in:

HERE IS A CLOSE-UP OF HOW THE COUNTERBORE IS MADE (AGAIN!!!11):

.125" aluminum...

Will it post this time? Only one way to find out, right?

Well done Comfy! Bet those finger tips get plenty hot! :slight_smile:

Noooo... if it gets hot, the tooling is worn out. Keeping things sharp plus only making really shallow cuts .003-.005 at a time is the only way to do this. Otherwise it's just incredibly frustrating (and lots of parts get ruined). Yes it takes forever making such small cuts but it's a lot cheaper than a lathe.

Look how tiny it is!

:love: