I donāt find much time to sneak mods in but today I found a few minutes to do this quick & dirty emitter swap on a Thorfire KL02. Even essentially serving as a spare battery holder I found itās CW tint too harsh. I swapped in a Nichia 119A (3,500k I think?)
Given my time crunch along w/ the fact that my heatgun & dremel cutoff discs are all loaned out I didnāt fashion a good heat path for it. I had initially thought to put in a dual emitter in the light. However I ended up just going into my parts bin & grabbing an emitter pulled from somewhere. As the stock emitter has itās anode & cathode similar to a luxeon I or III -> it took a little finagling to fit a Nichia 119A in it.
Given that this light will just serve as a battery carrier for me Iām not all that certain that I will spend anymore time on this light.
Hereās a god awful pic (I see that now. Sorry. But take my word for it. Itās tint is very pleasing to the eye.)
I was going to change the driver and the emitter in a Thorfire C8S, but ended up with only swapping the emitter to a XP-L HI because they changed the driver to 20mm in the version I got :disappointed:
EDIT :
I was going to put a BLF X5/X6 in it , but the driver size put an end to that plan.
Here is the new plan : 20 mm boost driver from KD and a XHP50.2 :
Today I taught a flashlight how to see. It turns out that LEDs can be used to sense light if you run them in reverse. So, all I had to do was solder on an āoptic nerveā between its brain and its eye (MCU pin3 to LED-), tell it how to distinguish bright and dark signals (auto-calibrate), and then teach it something like Morse code.
The upshot of this is that you can configure it by pointing it at a phone or computer screen and playing an animated GIF or something for it to watch. (eventually, havenāt written the app yet)
ā¦ and the compiled code for it is shorter than the text of this post.
Iām not really sure yet if itās actually a useful feature or just a silly gimmick, but it was fun to make.
BLF should start protecting the IP we are generating here and file patents! Unbelievable amounts of progress we are achieving here, seeing flashlights, apps or screens to program lights, ā¦ before you know you can talk to flashlights directly.
I am sure many of the technical solutions discussed and developed in various threads on here are 1) new and 2) not patented yet. The amount of innovation and curiosity here is mind-blowing.
i.e. ādata exchange for contactless programming of flashlights via a light-emitting diodeā
I didnāt do a patent search on these things but I do have access to various content providers at work.
This is very very cool, I saw this few months back in some youtube video, good chance it was on a shot show, they made software for pc where you set all preferences (modes) etc and then point the flashlight at dedidated part of the screen, it blinks out (black/white) āsettingsā and thats all you have to do to get custom modes.
I just hope this could be done with so widespread Atmel Tiny13A mcu.
Well technically it will be either brighter or more effitient with a spring bypass but the difference probably wouldnāt be noticeable. That said, I bypass springs in all lights that run 2amps or higher.
Patents? Hah. I suspect the patents for this have already expired. Itās not new.
Thereās a Lux-RC driver which does it, though IIRC it uses a dedicated light sensor instead of the LED. Itās used in the super-expensive BOSS light. And for some reason, it has a ridiculous limitation of only 4 modes. At that point, why bother? I know someone who has one, and he wishes it was more like his Convoy triple.
Whatās actually new or different is:
Itās open-source.
It can work in any host without changes. Just make sure the driver has an āoptic nerveā.
It runs on low-end MCUs like the tiny25 (sorry, doesnāt fit on tiny13).
I only have an early prototype so far, a proof of concept. All it does at this point is calibrate itself then read a list of modes to configure a mode group. Itās pretty janky. Donāt expect anything stable for a while.
Iām using an old BLF X5 prototype for testing (bistro tiny25 driver), totally stock except for the firmware and a small wire I added. Hopefully there will eventually be drivers which include this connection by default instead of requiring an extra wire. Otherwise though, itās a totally normal build. Cheap stock host, normal MCPCB, only two wires going between the driver and LED (s). It likely wouldnāt work with multiple cells in series, and it shouldnāt work on UV or infrared lights, but those are really about the only restrictions.
One other thing to noteā¦ While all this complex onboard-programming stuff is fun and fancy, itās not necessarily an upgrade compared to something simple and straightforward. I think Crescendo (ramping UI for clickies) is probably better overall, precisely because itās so simple. And Narsilās ramping UI is better still. Why bother with a dozen buttons and levers when you can do everything with a single knob?