What did you mod today?

Hey Steve, coming from one of the best Tool Makers I have seen in a long time, that’s a heck of a compliment! Thank you very much! :beer:

I’m wanting to get back into it, tired of building CNC’s and Special machines, I really love turning handles, so for now small, simple and quick mech/mods keep me some what sane. :smiley:

Somewhat? :person_facepalming:

Ok, so that looks like you got your practice in, when is my grater X6 20700 tube coming? :wink:

Nice Kawi, really nice, glad to see you doing retrowork… :smiley:

Oh, and for the record, you also need to make us some 21700 tubes for the GT mini, it’ll do 1711 lumens on the 30T straight out of the box. :smiley: :smiley: :smiley: (measured, not speculation)

I got a spare piece of GRATER just in case I find something with a larger inside diameter. The M1 head I.D. measures 23mm so something around 25-26mm should be doable. I have a GT incoming, when…… who the heck knows??

Did you use some sort of formula with the knurling or just go for it?

Looks real neat kawi :+1: :+1:

Made a Nitecore TIP cover design and 3D printed it. It has raised edge around buttons to prevent accidental activation in pocket. It covers the micro USB port from dust. And it has a place for two magnets. Actually one enough to hold it in place.




TIP cover
STLSTLSTLSTL
I hope the link works.

No going for it didn’t work, those single post swivel head knurling tools wouldn’t get it for me either, here is a vid I found a couple years ago. I have the same set-up and tool post. It still took a couple tries before all was good! :person_facepalming:

Great, thank you :slight_smile:

18650 extension tube for that dreadful Astrolux E01. It’s a E02 now. Almost.


pic

Cutting square threads isn’t that much of a nightmare actually.


pic

Since I wanted to make this light usable, I cooked up an an attiny85 driver with [1x7135][1x7136@2.5A]*
(* instead of a FET since I wanted to have _some _regulation on turbo; plus no room for stacking 7135s)
and put Anduril (in FW3A config) on it.
Some tweaking required , now it’s fantastic. 1A is plenty for my EDC tasks and that should stay in regulation down to ~3.2V.
Lowest usable moon draws 2.4mA, quiescent current is 15µA. Not bad.

It still has the stock XP-L @4000k in it. I got a few LH351D’s and Luxeon V’s on the way to see how they compare ( (both 4000K, CRI ~70 vs 90+…. choices…. :expressionless: … any ideas?)

No knurling yet (as I don’t have the proper tools) but I kind of like it this simple. Maybe I’ll pattern etch it, maybe it gets a rubber cover.
We’ll see how this holds up anyway, it’s only AW-2007 aluminium which isn’t very corrosion resistant.

@kikkoman what tool did you use for the square threads? …and why is the EO1 dreadful?

Nice mod anyway :beer:

A hand-dremeled carbide insert (started out as a 11ER 60°), squared the end & sides, narrowed it down to ~.65mm, ground large side & end relief angles and put a tiny radius at the tip.
Could have started with a HSS blank too, I just figured grinding this would be quicker.
From there on it was just manual threading, 100rpm, 0.1mm depth of cut. That AW-2007 machines nicely.

Why dreadful? The driver, mainly. Terrible mode spacing (one mode low, the next three between 350 and ~1000 lm IIRC). No LVP either.
The UI wasn’t that bad (nice and simple) , good form factor too but the modes killed it.

Thanks! :beer: :beer: :beer:

Nice tip. I’ll try that one day. :+1: :beer:

Today I’ve tried Optisolis quad in AT40: Amutorch AT40

I got the pair of Eagle Eye X6 hosts in the mail today, already had the quad board from Noctigon loaded with Samsungs and waiting, so I made a nice thick 36x12mm copper spacer/heat sink to fit the 33mm Noctigon board and Ledil Angie into the head of the X6. Worked out really nice, need to build a driver so when a new thermal compound get’s here Monday I’ll be ready to assemble it. :slight_smile:

Edit: I went ahead and built a 17mm driver with Bistro and assembled the X6 Quad, after polishing up the top surface of the copper sink. Didn’t use any thermal paste or glue in the sink as I am waiting on Arctic Alumina Thermal Adhesive and a new synthetic diamond thermal paste that will be here Monday. A quick initial test with the light dry assembled showed some 3400 lumens on an LG HE4.

Edit II: The epoxy came early, went ahead and re-flowed the 33mm MCPCB to the copper sink then glued the assembly into the X6 head with the Arctic Alumina Thermal Adhesive and assembled it all….

Bistro was working fine on an LG HE4, but with a Sony VTC5A in place it glitches some of the time. And when it’s not glitching? How about 4616.1 lumens out the front? :smiley:

Edit II: Ok, so I applied the Bistro fix and recharged the VTC5A, now it works solidly and is really amazing…. 5171.55 lumens in Turbo at start. Pretty nice little light! :smiley:

Also got my GT mini yesterday, modified it today. It now pushes 1683.6 lumens out the front on a Sanyo GA cell. :smiley:

Samsung LH351D 5000K 80CRI emitter, domed, on the stock DTP 20mm copper MCPCB. Flashed Anduril to the driver, replaced the FET with an SIR800DP and put 18ga wires on it. 18ga spring bypass on the driver. Removed the pcb in the tail cap and fitted a copper disc with the large spring that came with the light, 20 ga bypass inside the spring.

Yes, I lost some of the tightness of the beam but the color and intensity is so much nicer I can live with the trade-off.

Some revisions to one of my Sipik 58s.

Bezel and optic same style as I had previous in this light, but now redone for a cleaner look.


This is the stock lens retaining ring with the interior widened with a cylindrical grinding bit. The LED-Lenser style optic is from a $10 zoom light from OSH.

The previous bezel and optic I had in this late used the same setup, but looked much messier because I ground the inside edge of the bezel where it meets the optic back. And I’d messed up with the glue and some had run down the backside of the optic creating blindspots in the spot beam. This latest attempt looks much more professional.

Minor revisions to the interior of the tailcap.
I replaced the momentary pushbutton I had in there with a bigger one with a more noticeable click. I also replaced the post in the tailcap with a wire-bypassed spring:

Previously, I had a spring at the head of the light and a post made from 2 copper disks at the back. That setup worked, but I noticed the light dimming or flickering after a minute or two of use. the problem was a bad ground connection. Adding the second spring fixed that problem by more firmly pressing the guts of the light together.

NOTE: This particular Sipik 58 clone is very heavily modified. Modifications include:

  • All anodizing removed and bare aluminum polished up.
  • LED mounted on a laminated copper post on top of a copper Noctigon.
  • Stock 18mm aspheric lens replaced with LED-Lenser style zooming TIR optic.
  • Driver replaced with e-switch FET driver from Mountain Electronics (moppydrv firmware. If I were doing the driver today I’d go with a FET+1 driver with D4 firmware).
  • Interior of battery compartment widened to make room for a rolled up plastic sheet and brass sheet to serve as electrically isolated e-switch contact.
  • Tailcap guts rebuilt with copper structure and momentary pushbutton switch.
  • contact rings added at both ends of the battery compartment for e-switch.
  • tailcap switch boot replaced with light-green GITD boot.
  • LED replaced with XPL HI 5D

What’s going on here?
Is that a DR. Jones RGB driver on the test stand?
Yes, yes it is!

More to come, off to ITCC finals!!!

And here it is (my ITCC event got pushed till this afternoon)!

After a stupid routing mistake cost me my $10 driver boards, and not willing to try, and wait for, boards again I switched direction on my triple Rook.

Dr Jones RGB driver! It’s his standard 17mm board with modded rgbw FW (to be 3 channel, using RGB for white). Each channel is 2x 7135.
I really like how well the flood optic works with the RGB and runtime should be ~1h on turbo “white”, though it’s more of a toy instead of a useful light so I may stack another 2 chips per channel later and sacrifice that runtime.