I’ve always wanted a very high dynamic range e-switch light, with the lowest possible moonlight mode as well as a very high turbo. I really enjoy low moonlight modes, as much for dark-room eye candy as for useful light. There is something amazing about a light than can range from so dim I can stare at the emitters to bright enough to light up an open field.
There are a lot of very good lights available with very high turbo modes, but low moonlights are harder to find. My two lowest moonlight modes are on my L3 Illumination L11c and a Convoy S3 triple with a DrJones H17F 3-ch 1/7/FET driver from mtn-electronics. However, neither is an e-switch light, which I’ve really come to prefer for ramping.
So, are there any firmware tricks or hardware mods that could be done to give a commonly available e-switch light (e.g. Emisar D4S or BLF Q8) a lower moonlight mode? Reading Anduril’s cfg-emisar-d4s.h implies that the driver has two channels, 3x7135s and FET. The lowest ramp value seems to set up PWM for a 1/256 duty cycle. I see two obvious ways to reduce average power to the emitters: use fewer 7135s, or use a lower PWM duty cycle. The latter seems tricky since the attiny 85 doesn’t have 10-bit or 16-bit PWM, and 1/256 is already the best you can do with 8-bit PWM. The former could work, but would also reduce the corresponding regulated high mode (which might be okay).
Are there any 3-channel (1/N/FET) e-switch drivers available? I know Lexel was working on one for the Q8. What about the D4 or D4S? Also, does Anduril support 3 channels?
Not what you’re asking…but you may find QTC interesting.
No driver, but you can ramp from ultra-low to quite bright.
My CRX-modded XHP50 light has the lowest mode much dimmer than a tritium tube and does 1600 lm at the max. Though the ramping is quite unstable at ultra-low modes, it jumps a bit up and down before settling.
The Q8 already has a very decent range, and most importantly it does not break the bank just for this specific and uncommon requirement of yours, ie i find it pointless to wanna do 0.001 lumens and then 60k lumens at the other end.
Change the firefly mode to 1 or 2 (max being 7), in fact in 1 might not even light up the XP-Ls unless the batts are fully charged. 2 is extremely dim. Be it the BLF ver (less throwy, slightly more lumens) or the Sofirn ver (slightly less lumens, > 100k cd)
BTW, the Q8 in firefly mode = 1 has a low mode so low that you gotta hold the darn soda can just 0.5cm away from your eyes and use your other hand to block the gaps so that ambient light does not enter in a dark room, so that you can see that the XP-Ls are just merely glowing. That’s how low it is.
It would be quite bright if you’d use a 3rd gen night vision to see if the LEDs are glowing, of course.
I don’t know if it’s technically possible to get any lower than that.
Thanks, I do find that super interesting. Modern materials science is amazing. I will order some of those QTC pills to play around with.
Yeah, like I said above this requirement is mostly for fun. I love a good, practical, and reasonably priced flashlight like the Q8 for practical lighting tasks. It has been especially useful for me when working on cars - small enough to put wherever I want, bright enough to bounce-light a large area, and enough runtime to get the job done. However, most of my interest in this hobby is more playful and theoretical. I realize that I will likely never use a 10^–10 lumen flashlight for anything practical, but it’s still an interesting limit to push.
Interesting, my Q8 is much brighter on moonlight. It produces a visible spot on the wall in a dark room when lit. I do not have any way to accurately measure lumen output, but I’d say it’s 10-100x as bright as my 2-year-old cheap green tritium zipper-pull vial. I have flashed it with Anduril and configured it for the minimum possible output in ramp settings. Anduril might use 1/150 instead of 1/256 for its minimum duty cycle though. I am still trying to work through the register configuration to figure out how the ATtiny85 timers are set up. Maybe I should just pop my Q8 open again and hook up the scope to measure the PWM waveform directly.
The AMC7135 datasheet doesn’t talk about timing at all, but I would imagine that there is a minimum on-time required for the regulator to begin conducting current. Even if I could set up a 1/1024 or 1/65536 duty-cycle with a high-bit-depth PWM, I am not sure if I would be able to get the emitter to light with a lower average current or if I would instead just hit a point below which the LED doesn’t even turn on. At that point, a truly adjustable current source would be required. The existing moonlight mode is already on the order of 2mA. Perhaps I should put together a test-bed with an OpAmp-based adjustable current source (this and this look interesting) and see what happens if I feed µA or nA to the emitter.
Also, just curious, what watch is that? I am on WatchUSeek pretty often too.
Hmm…that’s interesting. Mine when set in the config mode for moonlight level, setting #1 produces this ultra ultra dim mode even when run with batts @ 4.2V.
But it’s of no real practical use, like i said you’d need like a 2nd gen night vision google to at least see it.
The photo was taken at ISO 6400, f1.4, 1/10 seconds.
Moonlight setting #2 is more practical, i guess it could be useful in some settings like astronomy, esp with a red filter in front.
Yes at that level, it’s just able to project a very very dim hotspot onto the wall, and is just nice for very dark adapted eyes. You can do things with this level, but it’s pretty difficult to try and read words. So I find no practical reason to go any dimmer.
I’m not really a watch enthusiast. That’s a Vostok Europe Anchar, been wearing it since like 2006 or something.
