So I recently decided after losing my O’pen, that I “needed” a AAA light with a clip to replace it. The Lumintop Tool was the chosen host due to having a clicky and a nice matte black Type III anodizing, same as the O’pen before it.
First, a flurry of ordering and waiting for bits to arrive, by far the longest process…
To start, the host was opened up, the driver and XP-G2 were removed. The tailspring was shortened due to the extra height of the ATTiny.
Next up, a combination of wight’s 12mm FET+7135 10440 driver board the MTN-10DD kit, an ATTiny13 and a 7135 from RMM’s shop becomes an impossibly small FET+7135 driver. Note the red line on the driver: an extra jumper wire needed to be soldered for the 7135 to work.
Some filing/sanding is required to make the driver fit into the pill and lock into the cutouts to prevent rotation. A slot was cut into the pill for future access.
The BAT+ strap being quite thick and difficult to work with was shaped, pressed flat, soldered and sanded down for an extra ~0.3mm as the battery otherwise protrudes and prevents contact with the tailcap.
Next up was reflowing that tasty 4000K Nichia 219C from Cutter with my super advanced reflow setup That thing that looks a lot like a hot air tool is just random piece of metal used to hold down the star.
The process taken is fairly simple. Heat up and wet star with excess solder, scrape the solder flat using the edge of the chisel tip, do the same with the LED. Coat star with a little bit of flux. Place star on heat resistant surface, position LED, crank up the power on the soldering station, heat the star until the LED reflows.
To help dissipate heat, some copper wire (It’s amazingly hard to find copper sheet locally) was packed into the pill and then filled with solder to act as extra thermal mass to reduce heat sag in turbo.
The pill is assembled and due to not having a ready supply of Kapton tape and the real possibility of a short, an insulating frame is cut out of some plastic packaging.
Last but not least, Toykeeper’s (very well documented) BLF-A6 firmware is programmed in after some tweaking of the modes and battery values to suit. One nice thing about the ATTiny being on the battery side is that the SOIC clip can go straight into the head for programming.
For reference, the values for battery voltages for my chip are (Using a lab supply and DMM to check):
4.20 - 194
4.00 - 185
3.80 - 176
3.500 - 162
3.000 - 138
2.800 - 129
2.700 - 125
It appears that the voltage to value relationship is linear, just taking the values for 4.2v and 2.7v and plotting the rest should yield adequate results for the battery warning.
Less than 2 years ago I would never have envisaged anything more than buying a different LED or driver from one of the Asian online outfits, certainly there were no decent AAA sized options available. In short, BLF has been amazing!
Moon mode works with PWM 1 and uses 2.73mA, which should be sublumen.
Turbo uses 2.70A (thanks to the low vF of the 219C) which should give about 840 LED lumens according to djozz’s testing.
(i.e. almost a 1000:1 ratio of brightness can be called at will)
Finally, I can put this thread to rest.