Pics from yesterday and today with (quite a lot of) text:
So, the copper tube was too narrow to get the battery in, or rather, to get it out again.
I originally planned to hammer the tube over the internal assembly (battery, charger PCB, switch and driver), but this didn’t seem wise, so after looking for anything to ream the tube out, this was the only thing i could use:
Subtle tool…
This SDS drill is just over 20mm across the chisel point, and i used some household scrubbing fluid along with it.
This did the job, but the outside of the tube got a little wider too, and not evenly, so…
…the outside had to be sanded back to (near) perfection.
Drilled the switch button hole too.
Also drastically enlarged the centre hole in the shelf, because…
…this is how close the driver board will be to the shelf, so i needed to make some room for the slightly thick wires (22 AWG)…
This doesn’t look beneficial for the thermal path from LED’s to the shelf, but i figured it would still be good enough.
The LEDs are mounted on a DTP copper LED board, and although it’s a thin LED board, i think it will still conduct the heat pretty well to the shelf.
…but i did enlarge that hole more than necessary…
O well…
Switch button hole.
The tube is not 1 mm too long, as it turns out…
Not yet sure how to finish the tail.
Size comparison with 2 compact 18650 lights:
(Yes, the DQG is empty…)
_
_
And then, i noticed some problems in my plan…
It was the lack of a plan, i guess…
The USB port sat pretty deep, so it could only be charged with a long micro USB plug, and then it still didn’t go in far enough to snap in place…
The charging indicator LEDs didn’t line up nicely either with the side (at 90° of the USB port / switch.
Conclusion: the charger board should move a little towards the wall of the tube to solve these things.
But i used epoxy… No way i can move it…
Additionally, trying some switch boots, the switch was sitting a bit deeper than i would have liked.
And another problem with the switch itself occurred !?
Probably too much heat when soldering, but it had some problems every 8 clicks (full rotation of the inner parts every 8 clicks, or, i think it’s 8, could be 12 too…)
It could be traces of super glue too that crept in too…
Injecting WD40 with a syringe did help, and possibly it will improve further with use.
SOLUTION:
I have plenty of those charging boards (ordered 10 of them at the time).
I have another 20700.
I have a bunch of small Omtens.
I have another LD25 driver.
…so i make another assembly… :person_facepalming:
That’s a lot better.
It cost me a PCB trace though (that’s why the 1st attempt ended up so deep), so i had to lead the + to the indicator LEDs with a wire (yellow).
The switch has to be glued on top of the the charger IC again, but both go more towards the outside.
Need something on the back of the switch to glue it to the charger IC.
A bit of cotton from a cotton tip swab drenched with super glue will do.
Driver on top etcetera, and compared to the first assembly:
Goals met for now.
Bit of a waste of parts on the first assembly though…
No idea what to do with it or the parts it consists of…
All epoxied…
O well, who cares.
CONTINUED HERE